top-skimming import from sf.net

git-svn-id: https://tesseract-ocr.googlecode.com/svn/trunk/trunk@2 d0cd1f9f-072b-0410-8dd7-cf729c803f20

.cvsignore

@@ -0,0 +1,6 @@
+BUILD
+OWNERS
+Makefile
+README.google
+runautoconf
+config_auto.h

AUTHORS

@@ -0,0 +1,8 @@
+Ray Smith (lead developer) <theraysmith@users.sourceforge.net>
+Phil Cheatle
+Simon Crouch
+Dan Johnson
+Mark Seaman
+Sheelagh Huddleston
+Chris Newton
+... and several others.

COPYING

@@ -0,0 +1,23 @@
+This package contains the Tesseract Open Source OCR Engine.
+Orignally developed at Hewlett Packard Laboratories Bristol and
+at Hewlett Packard Co, Greeley Colorado, all the code
+in this distribution is now licensed under the Apache License:
+
+** Licensed under the Apache License, Version 2.0 (the "License");
+** you may not use this file except in compliance with the License.
+** You may obtain a copy of the License at
+** http://www.apache.org/licenses/LICENSE-2.0
+** Unless required by applicable law or agreed to in writing, software
+** distributed under the License is distributed on an "AS IS" BASIS,
+** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+** See the License for the specific language governing permissions and
+** limitations under the License.
+
+
+Other Dependencies and Licenses:
+================================
+The Aspirin/MIGRAINES system is no longer used.
+
+Tesseract can also make use of the libtiff library. (www.libtiff.org)
+Without libtiff, Tesseract can only read uncompressed and G3 compressed
+TIFF files.

ChangeLog

@@ -0,0 +1,20 @@
+June  2006 - V1.0 of open source Tesseract checked-in.
+Sep 7 2006 - V1.01.
+          Added mfcpch.cpp and getopt.cpp for VC++.
+          Fixed problem with greyscale images and no libtiff.
+          Stopped debug window from being used for the usage output.
+          Fixed load of inttemp for big-endian architectures.
+          Fixed some Mac compilation issues.
+Oct 4 2006 - V1.02
+          Removed dependency on Aspirin.
+          Fixed a few missing Apache license headers.
+          Removed $log.
+Feb 2 2007 - V1.03
+          Added mftraining and cntraining.
+          Added baseapi with adaptive thresholding for grey and color.
+          Fixed many memory leaks.
+          Fixed several bugs including lack of use of adaptive classifier.
+          Added ifdefs to eliminate graphics code and add embedded platform support.
+          Incorporated several patches, including 64-bit builds, Mac builds.
+          Minor accuracy improvements.
+

INSTALL

@@ -0,0 +1,229 @@
+Copyright 1994, 1995, 1996, 1999, 2000, 2001, 2002 Free Software
+Foundation, Inc.
+
+   This file is free documentation; the Free Software Foundation gives
+unlimited permission to copy, distribute and modify it.
+
+Basic Installation
+==================
+
+   These are generic installation instructions.
+
+   The `configure' shell script attempts to guess correct values for
+various system-dependent variables used during compilation.  It uses
+those values to create a `Makefile' in each directory of the package.
+It may also create one or more `.h' files containing system-dependent
+definitions.  Finally, it creates a shell script `config.status' that
+you can run in the future to recreate the current configuration, and a
+file `config.log' containing compiler output (useful mainly for
+debugging `configure').
+
+   It can also use an optional file (typically called `config.cache'
+and enabled with `--cache-file=config.cache' or simply `-C') that saves
+the results of its tests to speed up reconfiguring.  (Caching is
+disabled by default to prevent problems with accidental use of stale
+cache files.)
+
+   If you need to do unusual things to compile the package, please try
+to figure out how `configure' could check whether to do them, and mail
+diffs or instructions to the address given in the `README' so they can
+be considered for the next release.  If you are using the cache, and at
+some point `config.cache' contains results you don't want to keep, you
+may remove or edit it.
+
+   The file `configure.ac' (or `configure.in') is used to create
+`configure' by a program called `autoconf'.  You only need
+`configure.ac' if you want to change it or regenerate `configure' using
+a newer version of `autoconf'.
+
+The simplest way to compile this package is:
+
+  1. `cd' to the directory containing the package's source code and type
+     `./configure' to configure the package for your system.  If you're
+     using `csh' on an old version of System V, you might need to type
+     `sh ./configure' instead to prevent `csh' from trying to execute
+     `configure' itself.
+
+     Running `configure' takes awhile.  While running, it prints some
+     messages telling which features it is checking for.
+
+  2. Type `make' to compile the package.
+
+  3. Optionally, type `make check' to run any self-tests that come with
+     the package.
+
+  4. Type `make install' to install the programs and any data files and
+     documentation.
+
+  5. You can remove the program binaries and object files from the
+     source code directory by typing `make clean'.  To also remove the
+     files that `configure' created (so you can compile the package for
+     a different kind of computer), type `make distclean'.  There is
+     also a `make maintainer-clean' target, but that is intended mainly
+     for the package's developers.  If you use it, you may have to get
+     all sorts of other programs in order to regenerate files that came
+     with the distribution.
+
+Compilers and Options
+=====================
+
+   Some systems require unusual options for compilation or linking that
+the `configure' script does not know about.  Run `./configure --help'
+for details on some of the pertinent environment variables.
+
+   You can give `configure' initial values for configuration parameters
+by setting variables in the command line or in the environment.  Here
+is an example:
+
+     ./configure CC=c89 CFLAGS=-O2 LIBS=-lposix
+
+   *Note Defining Variables::, for more details.
+
+Compiling For Multiple Architectures
+====================================
+
+   You can compile the package for more than one kind of computer at the
+same time, by placing the object files for each architecture in their
+own directory.  To do this, you must use a version of `make' that
+supports the `VPATH' variable, such as GNU `make'.  `cd' to the
+directory where you want the object files and executables to go and run
+the `configure' script.  `configure' automatically checks for the
+source code in the directory that `configure' is in and in `..'.
+
+   If you have to use a `make' that does not support the `VPATH'
+variable, you have to compile the package for one architecture at a
+time in the source code directory.  After you have installed the
+package for one architecture, use `make distclean' before reconfiguring
+for another architecture.
+
+Installation Names
+==================
+
+   By default, `make install' will install the package's files in
+`/usr/local/bin', `/usr/local/man', etc.  You can specify an
+installation prefix other than `/usr/local' by giving `configure' the
+option `--prefix=PATH'.
+
+   You can specify separate installation prefixes for
+architecture-specific files and architecture-independent files.  If you
+give `configure' the option `--exec-prefix=PATH', the package will use
+PATH as the prefix for installing programs and libraries.
+Documentation and other data files will still use the regular prefix.
+
+   In addition, if you use an unusual directory layout you can give
+options like `--bindir=PATH' to specify different values for particular
+kinds of files.  Run `configure --help' for a list of the directories
+you can set and what kinds of files go in them.
+
+   If the package supports it, you can cause programs to be installed
+with an extra prefix or suffix on their names by giving `configure' the
+option `--program-prefix=PREFIX' or `--program-suffix=SUFFIX'.
+
+Optional Features
+=================
+
+   Some packages pay attention to `--enable-FEATURE' options to
+`configure', where FEATURE indicates an optional part of the package.
+They may also pay attention to `--with-PACKAGE' options, where PACKAGE
+is something like `gnu-as' or `x' (for the X Window System).  The
+`README' should mention any `--enable-' and `--with-' options that the
+package recognizes.
+
+   For packages that use the X Window System, `configure' can usually
+find the X include and library files automatically, but if it doesn't,
+you can use the `configure' options `--x-includes=DIR' and
+`--x-libraries=DIR' to specify their locations.
+
+Specifying the System Type
+==========================
+
+   There may be some features `configure' cannot figure out
+automatically, but needs to determine by the type of machine the package
+will run on.  Usually, assuming the package is built to be run on the
+_same_ architectures, `configure' can figure that out, but if it prints
+a message saying it cannot guess the machine type, give it the
+`--build=TYPE' option.  TYPE can either be a short name for the system
+type, such as `sun4', or a canonical name which has the form:
+
+     CPU-COMPANY-SYSTEM
+
+where SYSTEM can have one of these forms:
+
+     OS KERNEL-OS
+
+   See the file `config.sub' for the possible values of each field.  If
+`config.sub' isn't included in this package, then this package doesn't
+need to know the machine type.
+
+   If you are _building_ compiler tools for cross-compiling, you should
+use the `--target=TYPE' option to select the type of system they will
+produce code for.
+
+   If you want to _use_ a cross compiler, that generates code for a
+platform different from the build platform, you should specify the
+"host" platform (i.e., that on which the generated programs will
+eventually be run) with `--host=TYPE'.
+
+Sharing Defaults
+================
+
+   If you want to set default values for `configure' scripts to share,
+you can create a site shell script called `config.site' that gives
+default values for variables like `CC', `cache_file', and `prefix'.
+`configure' looks for `PREFIX/share/config.site' if it exists, then
+`PREFIX/etc/config.site' if it exists.  Or, you can set the
+`CONFIG_SITE' environment variable to the location of the site script.
+A warning: not all `configure' scripts look for a site script.
+
+Defining Variables
+==================
+
+   Variables not defined in a site shell script can be set in the
+environment passed to `configure'.  However, some packages may run
+configure again during the build, and the customized values of these
+variables may be lost.  In order to avoid this problem, you should set
+them in the `configure' command line, using `VAR=value'.  For example:
+
+     ./configure CC=/usr/local2/bin/gcc
+
+will cause the specified gcc to be used as the C compiler (unless it is
+overridden in the site shell script).
+
+`configure' Invocation
+======================
+
+   `configure' recognizes the following options to control how it
+operates.
+
+`--help'
+`-h'
+     Print a summary of the options to `configure', and exit.
+
+`--version'
+`-V'
+     Print the version of Autoconf used to generate the `configure'
+     script, and exit.
+
+`--cache-file=FILE'
+     Enable the cache: use and save the results of the tests in FILE,
+     traditionally `config.cache'.  FILE defaults to `/dev/null' to
+     disable caching.
+
+`--config-cache'
+`-C'
+     Alias for `--cache-file=config.cache'.
+
+`--quiet'
+`--silent'
+`-q'
+     Do not print messages saying which checks are being made.  To
+     suppress all normal output, redirect it to `/dev/null' (any error
+     messages will still be shown).
+
+`--srcdir=DIR'
+     Look for the package's source code in directory DIR.  Usually
+     `configure' can determine that directory automatically.
+
+`configure' also accepts some other, not widely useful, options.  Run
+`configure --help' for more details.
+

Makefile.am

@@ -0,0 +1,15 @@
+# TODO(luc) Add 'doc' to this list when ready
+SUBDIRS = ccstruct ccutil classify cutil dict display image textord viewer wordrec ccmain training
+
+EXTRA_DIST = tessdata phototest.tif tesseract.dsp tesseract.dsw
+#EXTRA_DIST = doc/html doc/@PACKAGE_NAME@_@PACKAGE_VERSION@.pdf doc/@PACKAGE_NAME@_@PACKAGE_VERSION@.ps.gz
+
+dist-hook:
+# Need to remove CVS directories from directories
+# added using EXTRA_DIST. $(distdir)/tessdata would in
+# theory suffice.
+	rm -rf `find $(distdir) -name CVS`
+# Also remove extra files not needed in a distribution
+	rm -rf `find $(distdir) -name configure.ac`
+	rm -rf `find $(distdir) -name acinclude.m4`
+	rm -rf `find $(distdir) -name aclocal.m4`

Makefile.in

@@ -0,0 +1,628 @@
+# Makefile.in generated by automake 1.9.6 from Makefile.am.
+# @configure_input@
+
+# Copyright (C) 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002,
+# 2003, 2004, 2005  Free Software Foundation, Inc.
+# This Makefile.in is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY, to the extent permitted by law; without
+# even the implied warranty of MERCHANTABILITY or FITNESS FOR A
+# PARTICULAR PURPOSE.
+
+@SET_MAKE@
+srcdir = @srcdir@
+top_srcdir = @top_srcdir@
+VPATH = @srcdir@
+pkgdatadir = $(datadir)/@PACKAGE@
+pkglibdir = $(libdir)/@PACKAGE@
+pkgincludedir = $(includedir)/@PACKAGE@
+top_builddir = .
+am__cd = CDPATH="$${ZSH_VERSION+.}$(PATH_SEPARATOR)" && cd
+INSTALL = @INSTALL@
+install_sh_DATA = $(install_sh) -c -m 644
+install_sh_PROGRAM = $(install_sh) -c
+install_sh_SCRIPT = $(install_sh) -c
+INSTALL_HEADER = $(INSTALL_DATA)
+transform = $(program_transform_name)
+NORMAL_INSTALL = :
+PRE_INSTALL = :
+POST_INSTALL = :
+NORMAL_UNINSTALL = :
+PRE_UNINSTALL = :
+POST_UNINSTALL = :
+build_triplet = @build@
+host_triplet = @host@
+DIST_COMMON = README $(am__configure_deps) $(srcdir)/Makefile.am \
+	$(srcdir)/Makefile.in $(top_srcdir)/config/config.h.in \
+	$(top_srcdir)/configure AUTHORS COPYING ChangeLog INSTALL NEWS \
+	config/config.guess config/config.sub config/depcomp \
+	config/install-sh config/missing config/mkinstalldirs
+subdir = .
+ACLOCAL_M4 = $(top_srcdir)/aclocal.m4
+am__aclocal_m4_deps = $(top_srcdir)/acinclude.m4 \
+	$(top_srcdir)/config/ac_define_versionlevel.m4 \
+	$(top_srcdir)/config/acinclude_custom.m4 \
+	$(top_srcdir)/configure.ac
+am__configure_deps = $(am__aclocal_m4_deps) $(CONFIGURE_DEPENDENCIES) \
+	$(ACLOCAL_M4)
+am__CONFIG_DISTCLEAN_FILES = config.status config.cache config.log \
+ configure.lineno configure.status.lineno
+mkinstalldirs = $(SHELL) $(top_srcdir)/config/mkinstalldirs
+CONFIG_HEADER = config_auto.h
+CONFIG_CLEAN_FILES =
+SOURCES =
+DIST_SOURCES =
+RECURSIVE_TARGETS = all-recursive check-recursive dvi-recursive \
+	html-recursive info-recursive install-data-recursive \
+	install-exec-recursive install-info-recursive \
+	install-recursive installcheck-recursive installdirs-recursive \
+	pdf-recursive ps-recursive uninstall-info-recursive \
+	uninstall-recursive
+ETAGS = etags
+CTAGS = ctags
+DIST_SUBDIRS = $(SUBDIRS)
+DISTFILES = $(DIST_COMMON) $(DIST_SOURCES) $(TEXINFOS) $(EXTRA_DIST)
+distdir = $(PACKAGE)-$(VERSION)
+top_distdir = $(distdir)
+am__remove_distdir = \
+  { test ! -d $(distdir) \
+    || { find $(distdir) -type d ! -perm -200 -exec chmod u+w {} ';' \
+         && rm -fr $(distdir); }; }
+DIST_ARCHIVES = $(distdir).tar.gz
+GZIP_ENV = --best
+distuninstallcheck_listfiles = find . -type f -print
+distcleancheck_listfiles = find . -type f -print
+ACLOCAL = @ACLOCAL@
+AMDEP_FALSE = @AMDEP_FALSE@
+AMDEP_TRUE = @AMDEP_TRUE@
+AMTAR = @AMTAR@
+AUTOCONF = @AUTOCONF@
+AUTOHEADER = @AUTOHEADER@
+AUTOMAKE = @AUTOMAKE@
+AWK = @AWK@
+CC = @CC@
+CCDEPMODE = @CCDEPMODE@
+CFLAGS = @CFLAGS@
+CPPFLAGS = @CPPFLAGS@
+CXX = @CXX@
+CXXCPP = @CXXCPP@
+CXXDEPMODE = @CXXDEPMODE@
+CXXFLAGS = @CXXFLAGS@
+CXXRPOFLAGS = @CXXRPOFLAGS@
+CYGPATH_W = @CYGPATH_W@
+DEFS = @DEFS@
+DEPDIR = @DEPDIR@
+ECHO_C = @ECHO_C@
+ECHO_N = @ECHO_N@
+ECHO_T = @ECHO_T@
+EGREP = @EGREP@
+EXEEXT = @EXEEXT@
+GNUWIN32_DIR = @GNUWIN32_DIR@
+HAVE_GNUWIN32_FALSE = @HAVE_GNUWIN32_FALSE@
+HAVE_GNUWIN32_TRUE = @HAVE_GNUWIN32_TRUE@
+HAVE_LIBTIFF_FALSE = @HAVE_LIBTIFF_FALSE@
+HAVE_LIBTIFF_TRUE = @HAVE_LIBTIFF_TRUE@
+INSTALL_DATA = @INSTALL_DATA@
+INSTALL_PROGRAM = @INSTALL_PROGRAM@
+INSTALL_SCRIPT = @INSTALL_SCRIPT@
+INSTALL_STRIP_PROGRAM = @INSTALL_STRIP_PROGRAM@
+LDFLAGS = @LDFLAGS@
+LIBOBJS = @LIBOBJS@
+LIBS = @LIBS@
+LIBTIFF_CFLAGS = @LIBTIFF_CFLAGS@
+LIBTIFF_LIBS = @LIBTIFF_LIBS@
+LTLIBOBJS = @LTLIBOBJS@
+MAINT = @MAINT@
+MAINTAINER_MODE_FALSE = @MAINTAINER_MODE_FALSE@
+MAINTAINER_MODE_TRUE = @MAINTAINER_MODE_TRUE@
+MAKEINFO = @MAKEINFO@
+OBJEXT = @OBJEXT@
+OPTS = @OPTS@
+PACKAGE = @PACKAGE@
+PACKAGE_BUGREPORT = @PACKAGE_BUGREPORT@
+PACKAGE_DATE = @PACKAGE_DATE@
+PACKAGE_NAME = @PACKAGE_NAME@
+PACKAGE_STRING = @PACKAGE_STRING@
+PACKAGE_TARNAME = @PACKAGE_TARNAME@
+PACKAGE_VERSION = @PACKAGE_VERSION@
+PACKAGE_YEAR = @PACKAGE_YEAR@
+PATH_SEPARATOR = @PATH_SEPARATOR@
+RANLIB = @RANLIB@
+RPO_NO = @RPO_NO@
+RPO_YES = @RPO_YES@
+SET_MAKE = @SET_MAKE@
+SHELL = @SHELL@
+STRIP = @STRIP@
+USING_CL_FALSE = @USING_CL_FALSE@
+USING_CL_TRUE = @USING_CL_TRUE@
+VERSION = @VERSION@
+ac_ct_CC = @ac_ct_CC@
+ac_ct_CXX = @ac_ct_CXX@
+ac_ct_RANLIB = @ac_ct_RANLIB@
+ac_ct_STRIP = @ac_ct_STRIP@
+am__fastdepCC_FALSE = @am__fastdepCC_FALSE@
+am__fastdepCC_TRUE = @am__fastdepCC_TRUE@
+am__fastdepCXX_FALSE = @am__fastdepCXX_FALSE@
+am__fastdepCXX_TRUE = @am__fastdepCXX_TRUE@
+am__include = @am__include@
+am__leading_dot = @am__leading_dot@
+am__quote = @am__quote@
+am__tar = @am__tar@
+am__untar = @am__untar@
+bindir = @bindir@
+build = @build@
+build_alias = @build_alias@
+build_cpu = @build_cpu@
+build_os = @build_os@
+build_vendor = @build_vendor@
+datadir = @datadir@
+exec_prefix = @exec_prefix@
+host = @host@
+host_alias = @host_alias@
+host_cpu = @host_cpu@
+host_os = @host_os@
+host_vendor = @host_vendor@
+includedir = @includedir@
+infodir = @infodir@
+install_sh = @install_sh@
+libdir = @libdir@
+libexecdir = @libexecdir@
+localstatedir = @localstatedir@
+mandir = @mandir@
+mkdir_p = @mkdir_p@
+oldincludedir = @oldincludedir@
+prefix = @prefix@
+program_transform_name = @program_transform_name@
+sbindir = @sbindir@
+sharedstatedir = @sharedstatedir@
+sysconfdir = @sysconfdir@
+target_alias = @target_alias@
+
+# TODO(luc) Add 'doc' to this list when ready
+SUBDIRS = ccstruct ccutil classify cutil dict display image textord viewer wordrec ccmain training
+EXTRA_DIST = tessdata phototest.tif tesseract.dsp tesseract.dsw
+all: config_auto.h
+	$(MAKE) $(AM_MAKEFLAGS) all-recursive
+
+.SUFFIXES:
+am--refresh:
+	@:
+$(srcdir)/Makefile.in: @MAINTAINER_MODE_TRUE@ $(srcdir)/Makefile.am  $(am__configure_deps)
+	@for dep in $?; do \
+	  case '$(am__configure_deps)' in \
+	    *$$dep*) \
+	      echo ' cd $(srcdir) && $(AUTOMAKE) --gnu '; \
+	      cd $(srcdir) && $(AUTOMAKE) --gnu  \
+		&& exit 0; \
+	      exit 1;; \
+	  esac; \
+	done; \
+	echo ' cd $(top_srcdir) && $(AUTOMAKE) --gnu  Makefile'; \
+	cd $(top_srcdir) && \
+	  $(AUTOMAKE) --gnu  Makefile
+.PRECIOUS: Makefile
+Makefile: $(srcdir)/Makefile.in $(top_builddir)/config.status
+	@case '$?' in \
+	  *config.status*) \
+	    echo ' $(SHELL) ./config.status'; \
+	    $(SHELL) ./config.status;; \
+	  *) \
+	    echo ' cd $(top_builddir) && $(SHELL) ./config.status $@ $(am__depfiles_maybe)'; \
+	    cd $(top_builddir) && $(SHELL) ./config.status $@ $(am__depfiles_maybe);; \
+	esac;
+
+$(top_builddir)/config.status: $(top_srcdir)/configure $(CONFIG_STATUS_DEPENDENCIES)
+	$(SHELL) ./config.status --recheck
+
+$(top_srcdir)/configure: @MAINTAINER_MODE_TRUE@ $(am__configure_deps)
+	cd $(srcdir) && $(AUTOCONF)
+$(ACLOCAL_M4): @MAINTAINER_MODE_TRUE@ $(am__aclocal_m4_deps)
+	cd $(srcdir) && $(ACLOCAL) $(ACLOCAL_AMFLAGS)
+
+config_auto.h: stamp-h1
+	@if test ! -f $@; then \
+	  rm -f stamp-h1; \
+	  $(MAKE) stamp-h1; \
+	else :; fi
+
+stamp-h1: $(top_srcdir)/config/config.h.in $(top_builddir)/config.status
+	@rm -f stamp-h1
+	cd $(top_builddir) && $(SHELL) ./config.status config_auto.h
+$(top_srcdir)/config/config.h.in: @MAINTAINER_MODE_TRUE@ $(am__configure_deps) 
+	cd $(top_srcdir) && $(AUTOHEADER)
+	rm -f stamp-h1
+	touch $@
+
+distclean-hdr:
+	-rm -f config_auto.h stamp-h1
+uninstall-info-am:
+
+# This directory's subdirectories are mostly independent; you can cd
+# into them and run `make' without going through this Makefile.
+# To change the values of `make' variables: instead of editing Makefiles,
+# (1) if the variable is set in `config.status', edit `config.status'
+#     (which will cause the Makefiles to be regenerated when you run `make');
+# (2) otherwise, pass the desired values on the `make' command line.
+$(RECURSIVE_TARGETS):
+	@failcom='exit 1'; \
+	for f in x $$MAKEFLAGS; do \
+	  case $$f in \
+	    *=* | --[!k]*);; \
+	    *k*) failcom='fail=yes';; \
+	  esac; \
+	done; \
+	dot_seen=no; \
+	target=`echo $@ | sed s/-recursive//`; \
+	list='$(SUBDIRS)'; for subdir in $$list; do \
+	  echo "Making $$target in $$subdir"; \
+	  if test "$$subdir" = "."; then \
+	    dot_seen=yes; \
+	    local_target="$$target-am"; \
+	  else \
+	    local_target="$$target"; \
+	  fi; \
+	  (cd $$subdir && $(MAKE) $(AM_MAKEFLAGS) $$local_target) \
+	  || eval $$failcom; \
+	done; \
+	if test "$$dot_seen" = "no"; then \
+	  $(MAKE) $(AM_MAKEFLAGS) "$$target-am" || exit 1; \
+	fi; test -z "$$fail"
+
+mostlyclean-recursive clean-recursive distclean-recursive \
+maintainer-clean-recursive:
+	@failcom='exit 1'; \
+	for f in x $$MAKEFLAGS; do \
+	  case $$f in \
+	    *=* | --[!k]*);; \
+	    *k*) failcom='fail=yes';; \
+	  esac; \
+	done; \
+	dot_seen=no; \
+	case "$@" in \
+	  distclean-* | maintainer-clean-*) list='$(DIST_SUBDIRS)' ;; \
+	  *) list='$(SUBDIRS)' ;; \
+	esac; \
+	rev=''; for subdir in $$list; do \
+	  if test "$$subdir" = "."; then :; else \
+	    rev="$$subdir $$rev"; \
+	  fi; \
+	done; \
+	rev="$$rev ."; \
+	target=`echo $@ | sed s/-recursive//`; \
+	for subdir in $$rev; do \
+	  echo "Making $$target in $$subdir"; \
+	  if test "$$subdir" = "."; then \
+	    local_target="$$target-am"; \
+	  else \
+	    local_target="$$target"; \
+	  fi; \
+	  (cd $$subdir && $(MAKE) $(AM_MAKEFLAGS) $$local_target) \
+	  || eval $$failcom; \
+	done && test -z "$$fail"
+tags-recursive:
+	list='$(SUBDIRS)'; for subdir in $$list; do \
+	  test "$$subdir" = . || (cd $$subdir && $(MAKE) $(AM_MAKEFLAGS) tags); \
+	done
+ctags-recursive:
+	list='$(SUBDIRS)'; for subdir in $$list; do \
+	  test "$$subdir" = . || (cd $$subdir && $(MAKE) $(AM_MAKEFLAGS) ctags); \
+	done
+
+ID: $(HEADERS) $(SOURCES) $(LISP) $(TAGS_FILES)
+	list='$(SOURCES) $(HEADERS) $(LISP) $(TAGS_FILES)'; \
+	unique=`for i in $$list; do \
+	    if test -f "$$i"; then echo $$i; else echo $(srcdir)/$$i; fi; \
+	  done | \
+	  $(AWK) '    { files[$$0] = 1; } \
+	       END { for (i in files) print i; }'`; \
+	mkid -fID $$unique
+tags: TAGS
+
+TAGS: tags-recursive $(HEADERS) $(SOURCES)  $(TAGS_DEPENDENCIES) \
+		$(TAGS_FILES) $(LISP)
+	tags=; \
+	here=`pwd`; \
+	if ($(ETAGS) --etags-include --version) >/dev/null 2>&1; then \
+	  include_option=--etags-include; \
+	  empty_fix=.; \
+	else \
+	  include_option=--include; \
+	  empty_fix=; \
+	fi; \
+	list='$(SUBDIRS)'; for subdir in $$list; do \
+	  if test "$$subdir" = .; then :; else \
+	    test ! -f $$subdir/TAGS || \
+	      tags="$$tags $$include_option=$$here/$$subdir/TAGS"; \
+	  fi; \
+	done; \
+	list='$(SOURCES) $(HEADERS)  $(LISP) $(TAGS_FILES)'; \
+	unique=`for i in $$list; do \
+	    if test -f "$$i"; then echo $$i; else echo $(srcdir)/$$i; fi; \
+	  done | \
+	  $(AWK) '    { files[$$0] = 1; } \
+	       END { for (i in files) print i; }'`; \
+	if test -z "$(ETAGS_ARGS)$$tags$$unique"; then :; else \
+	  test -n "$$unique" || unique=$$empty_fix; \
+	  $(ETAGS) $(ETAGSFLAGS) $(AM_ETAGSFLAGS) $(ETAGS_ARGS) \
+	    $$tags $$unique; \
+	fi
+ctags: CTAGS
+CTAGS: ctags-recursive $(HEADERS) $(SOURCES)  $(TAGS_DEPENDENCIES) \
+		$(TAGS_FILES) $(LISP)
+	tags=; \
+	here=`pwd`; \
+	list='$(SOURCES) $(HEADERS)  $(LISP) $(TAGS_FILES)'; \
+	unique=`for i in $$list; do \
+	    if test -f "$$i"; then echo $$i; else echo $(srcdir)/$$i; fi; \
+	  done | \
+	  $(AWK) '    { files[$$0] = 1; } \
+	       END { for (i in files) print i; }'`; \
+	test -z "$(CTAGS_ARGS)$$tags$$unique" \
+	  || $(CTAGS) $(CTAGSFLAGS) $(AM_CTAGSFLAGS) $(CTAGS_ARGS) \
+	     $$tags $$unique
+
+GTAGS:
+	here=`$(am__cd) $(top_builddir) && pwd` \
+	  && cd $(top_srcdir) \
+	  && gtags -i $(GTAGS_ARGS) $$here
+
+distclean-tags:
+	-rm -f TAGS ID GTAGS GRTAGS GSYMS GPATH tags
+
+distdir: $(DISTFILES)
+	$(am__remove_distdir)
+	mkdir $(distdir)
+	$(mkdir_p) $(distdir)/config
+	@srcdirstrip=`echo "$(srcdir)" | sed 's|.|.|g'`; \
+	topsrcdirstrip=`echo "$(top_srcdir)" | sed 's|.|.|g'`; \
+	list='$(DISTFILES)'; for file in $$list; do \
+	  case $$file in \
+	    $(srcdir)/*) file=`echo "$$file" | sed "s|^$$srcdirstrip/||"`;; \
+	    $(top_srcdir)/*) file=`echo "$$file" | sed "s|^$$topsrcdirstrip/|$(top_builddir)/|"`;; \
+	  esac; \
+	  if test -f $$file || test -d $$file; then d=.; else d=$(srcdir); fi; \
+	  dir=`echo "$$file" | sed -e 's,/[^/]*$$,,'`; \
+	  if test "$$dir" != "$$file" && test "$$dir" != "."; then \
+	    dir="/$$dir"; \
+	    $(mkdir_p) "$(distdir)$$dir"; \
+	  else \
+	    dir=''; \
+	  fi; \
+	  if test -d $$d/$$file; then \
+	    if test -d $(srcdir)/$$file && test $$d != $(srcdir); then \
+	      cp -pR $(srcdir)/$$file $(distdir)$$dir || exit 1; \
+	    fi; \
+	    cp -pR $$d/$$file $(distdir)$$dir || exit 1; \
+	  else \
+	    test -f $(distdir)/$$file \
+	    || cp -p $$d/$$file $(distdir)/$$file \
+	    || exit 1; \
+	  fi; \
+	done
+	list='$(DIST_SUBDIRS)'; for subdir in $$list; do \
+	  if test "$$subdir" = .; then :; else \
+	    test -d "$(distdir)/$$subdir" \
+	    || $(mkdir_p) "$(distdir)/$$subdir" \
+	    || exit 1; \
+	    distdir=`$(am__cd) $(distdir) && pwd`; \
+	    top_distdir=`$(am__cd) $(top_distdir) && pwd`; \
+	    (cd $$subdir && \
+	      $(MAKE) $(AM_MAKEFLAGS) \
+	        top_distdir="$$top_distdir" \
+	        distdir="$$distdir/$$subdir" \
+	        distdir) \
+	      || exit 1; \
+	  fi; \
+	done
+	$(MAKE) $(AM_MAKEFLAGS) \
+	  top_distdir="$(top_distdir)" distdir="$(distdir)" \
+	  dist-hook
+	-find $(distdir) -type d ! -perm -777 -exec chmod a+rwx {} \; -o \
+	  ! -type d ! -perm -444 -links 1 -exec chmod a+r {} \; -o \
+	  ! -type d ! -perm -400 -exec chmod a+r {} \; -o \
+	  ! -type d ! -perm -444 -exec $(SHELL) $(install_sh) -c -m a+r {} {} \; \
+	|| chmod -R a+r $(distdir)
+dist-gzip: distdir
+	tardir=$(distdir) && $(am__tar) | GZIP=$(GZIP_ENV) gzip -c >$(distdir).tar.gz
+	$(am__remove_distdir)
+
+dist-bzip2: distdir
+	tardir=$(distdir) && $(am__tar) | bzip2 -9 -c >$(distdir).tar.bz2
+	$(am__remove_distdir)
+
+dist-tarZ: distdir
+	tardir=$(distdir) && $(am__tar) | compress -c >$(distdir).tar.Z
+	$(am__remove_distdir)
+
+dist-shar: distdir
+	shar $(distdir) | GZIP=$(GZIP_ENV) gzip -c >$(distdir).shar.gz
+	$(am__remove_distdir)
+
+dist-zip: distdir
+	-rm -f $(distdir).zip
+	zip -rq $(distdir).zip $(distdir)
+	$(am__remove_distdir)
+
+dist dist-all: distdir
+	tardir=$(distdir) && $(am__tar) | GZIP=$(GZIP_ENV) gzip -c >$(distdir).tar.gz
+	$(am__remove_distdir)
+
+# This target untars the dist file and tries a VPATH configuration.  Then
+# it guarantees that the distribution is self-contained by making another
+# tarfile.
+distcheck: dist
+	case '$(DIST_ARCHIVES)' in \
+	*.tar.gz*) \
+	  GZIP=$(GZIP_ENV) gunzip -c $(distdir).tar.gz | $(am__untar) ;;\
+	*.tar.bz2*) \
+	  bunzip2 -c $(distdir).tar.bz2 | $(am__untar) ;;\
+	*.tar.Z*) \
+	  uncompress -c $(distdir).tar.Z | $(am__untar) ;;\
+	*.shar.gz*) \
+	  GZIP=$(GZIP_ENV) gunzip -c $(distdir).shar.gz | unshar ;;\
+	*.zip*) \
+	  unzip $(distdir).zip ;;\
+	esac
+	chmod -R a-w $(distdir); chmod a+w $(distdir)
+	mkdir $(distdir)/_build
+	mkdir $(distdir)/_inst
+	chmod a-w $(distdir)
+	dc_install_base=`$(am__cd) $(distdir)/_inst && pwd | sed -e 's,^[^:\\/]:[\\/],/,'` \
+	  && dc_destdir="$${TMPDIR-/tmp}/am-dc-$$$$/" \
+	  && cd $(distdir)/_build \
+	  && ../configure --srcdir=.. --prefix="$$dc_install_base" \
+	    $(DISTCHECK_CONFIGURE_FLAGS) \
+	  && $(MAKE) $(AM_MAKEFLAGS) \
+	  && $(MAKE) $(AM_MAKEFLAGS) dvi \
+	  && $(MAKE) $(AM_MAKEFLAGS) check \
+	  && $(MAKE) $(AM_MAKEFLAGS) install \
+	  && $(MAKE) $(AM_MAKEFLAGS) installcheck \
+	  && $(MAKE) $(AM_MAKEFLAGS) uninstall \
+	  && $(MAKE) $(AM_MAKEFLAGS) distuninstallcheck_dir="$$dc_install_base" \
+	        distuninstallcheck \
+	  && chmod -R a-w "$$dc_install_base" \
+	  && ({ \
+	       (cd ../.. && umask 077 && mkdir "$$dc_destdir") \
+	       && $(MAKE) $(AM_MAKEFLAGS) DESTDIR="$$dc_destdir" install \
+	       && $(MAKE) $(AM_MAKEFLAGS) DESTDIR="$$dc_destdir" uninstall \
+	       && $(MAKE) $(AM_MAKEFLAGS) DESTDIR="$$dc_destdir" \
+	            distuninstallcheck_dir="$$dc_destdir" distuninstallcheck; \
+	      } || { rm -rf "$$dc_destdir"; exit 1; }) \
+	  && rm -rf "$$dc_destdir" \
+	  && $(MAKE) $(AM_MAKEFLAGS) dist \
+	  && rm -rf $(DIST_ARCHIVES) \
+	  && $(MAKE) $(AM_MAKEFLAGS) distcleancheck
+	$(am__remove_distdir)
+	@(echo "$(distdir) archives ready for distribution: "; \
+	  list='$(DIST_ARCHIVES)'; for i in $$list; do echo $$i; done) | \
+	  sed -e '1{h;s/./=/g;p;x;}' -e '$${p;x;}'
+distuninstallcheck:
+	@cd $(distuninstallcheck_dir) \
+	&& test `$(distuninstallcheck_listfiles) | wc -l` -le 1 \
+	   || { echo "ERROR: files left after uninstall:" ; \
+	        if test -n "$(DESTDIR)"; then \
+	          echo "  (check DESTDIR support)"; \
+	        fi ; \
+	        $(distuninstallcheck_listfiles) ; \
+	        exit 1; } >&2
+distcleancheck: distclean
+	@if test '$(srcdir)' = . ; then \
+	  echo "ERROR: distcleancheck can only run from a VPATH build" ; \
+	  exit 1 ; \
+	fi
+	@test `$(distcleancheck_listfiles) | wc -l` -eq 0 \
+	  || { echo "ERROR: files left in build directory after distclean:" ; \
+	       $(distcleancheck_listfiles) ; \
+	       exit 1; } >&2
+check-am: all-am
+check: check-recursive
+all-am: Makefile config_auto.h
+installdirs: installdirs-recursive
+installdirs-am:
+install: install-recursive
+install-exec: install-exec-recursive
+install-data: install-data-recursive
+uninstall: uninstall-recursive
+
+install-am: all-am
+	@$(MAKE) $(AM_MAKEFLAGS) install-exec-am install-data-am
+
+installcheck: installcheck-recursive
+install-strip:
+	$(MAKE) $(AM_MAKEFLAGS) INSTALL_PROGRAM="$(INSTALL_STRIP_PROGRAM)" \
+	  install_sh_PROGRAM="$(INSTALL_STRIP_PROGRAM)" INSTALL_STRIP_FLAG=-s \
+	  `test -z '$(STRIP)' || \
+	    echo "INSTALL_PROGRAM_ENV=STRIPPROG='$(STRIP)'"` install
+mostlyclean-generic:
+
+clean-generic:
+
+distclean-generic:
+	-test -z "$(CONFIG_CLEAN_FILES)" || rm -f $(CONFIG_CLEAN_FILES)
+
+maintainer-clean-generic:
+	@echo "This command is intended for maintainers to use"
+	@echo "it deletes files that may require special tools to rebuild."
+clean: clean-recursive
+
+clean-am: clean-generic mostlyclean-am
+
+distclean: distclean-recursive
+	-rm -f $(am__CONFIG_DISTCLEAN_FILES)
+	-rm -f Makefile
+distclean-am: clean-am distclean-generic distclean-hdr distclean-tags
+
+dvi: dvi-recursive
+
+dvi-am:
+
+html: html-recursive
+
+info: info-recursive
+
+info-am:
+
+install-data-am:
+
+install-exec-am:
+
+install-info: install-info-recursive
+
+install-man:
+
+installcheck-am:
+
+maintainer-clean: maintainer-clean-recursive
+	-rm -f $(am__CONFIG_DISTCLEAN_FILES)
+	-rm -rf $(top_srcdir)/autom4te.cache
+	-rm -f Makefile
+maintainer-clean-am: distclean-am maintainer-clean-generic
+
+mostlyclean: mostlyclean-recursive
+
+mostlyclean-am: mostlyclean-generic
+
+pdf: pdf-recursive
+
+pdf-am:
+
+ps: ps-recursive
+
+ps-am:
+
+uninstall-am: uninstall-info-am
+
+uninstall-info: uninstall-info-recursive
+
+.PHONY: $(RECURSIVE_TARGETS) CTAGS GTAGS all all-am am--refresh check \
+	check-am clean clean-generic clean-recursive ctags \
+	ctags-recursive dist dist-all dist-bzip2 dist-gzip dist-hook \
+	dist-shar dist-tarZ dist-zip distcheck distclean \
+	distclean-generic distclean-hdr distclean-recursive \
+	distclean-tags distcleancheck distdir distuninstallcheck dvi \
+	dvi-am html html-am info info-am install install-am \
+	install-data install-data-am install-exec install-exec-am \
+	install-info install-info-am install-man install-strip \
+	installcheck installcheck-am installdirs installdirs-am \
+	maintainer-clean maintainer-clean-generic \
+	maintainer-clean-recursive mostlyclean mostlyclean-generic \
+	mostlyclean-recursive pdf pdf-am ps ps-am tags tags-recursive \
+	uninstall uninstall-am uninstall-info-am
+
+#EXTRA_DIST = doc/html doc/@PACKAGE_NAME@_@PACKAGE_VERSION@.pdf doc/@PACKAGE_NAME@_@PACKAGE_VERSION@.ps.gz
+
+dist-hook:
+# Need to remove CVS directories from directories
+# added using EXTRA_DIST. $(distdir)/tessdata would in
+# theory suffice.
+	rm -rf `find $(distdir) -name CVS`
+# Also remove extra files not needed in a distribution
+	rm -rf `find $(distdir) -name configure.ac`
+	rm -rf `find $(distdir) -name acinclude.m4`
+	rm -rf `find $(distdir) -name aclocal.m4`
+# Tell versions [3.59,3.63) of GNU make to not export all variables.
+# Otherwise a system limit (for SysV at least) may be exceeded.
+.NOEXPORT:

NEWS

@@ -0,0 +1 @@
+Stub file. To be populated at a later stage.

README

@@ -0,0 +1,85 @@
+Introduction
+============
+This package contains the Tesseract Open Source OCR Engine.
+Orignally developed at Hewlett Packard Laboratories Bristol and
+at Hewlett Packard Co, Greeley Colorado, all the code
+in this distribution is now licensed under the Apache License:
+
+** Licensed under the Apache License, Version 2.0 (the "License");
+** you may not use this file except in compliance with the License.
+** You may obtain a copy of the License at
+** http://www.apache.org/licenses/LICENSE-2.0
+** Unless required by applicable law or agreed to in writing, software
+** distributed under the License is distributed on an "AS IS" BASIS,
+** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+** See the License for the specific language governing permissions and
+** limitations under the License.
+
+
+Other Dependencies and Licenses:
+================================
+The Aspirin/MIGRAINES system is no longer required.
+
+Tesseract can also make use of the libtiff library. (www.libtiff.org)
+Without libtiff, Tesseract can only read uncompressed and G3 compressed
+TIFF files.
+
+
+History:
+========
+The engine was developed at Hewlett Packard Laboratories Bristol and
+at Hewlett Packard Co, Greeley Colorado between 1985 and 1994, with some
+more changes made in 1996 to port to Windows, and some C++izing in 1998.
+A lot of the code was written in C, and then some more was written in C++.
+Since then all the code has been converted to at least compile with a C++
+compiler. Currently it builds under Linux with gcc2.95 and under Windows
+with VC++6. The C++ code makes heavy use of a list system using macros.
+This predates stl, was portable before stl, and is more efficent than stl
+lists, but has the big negative that if you do get a segmentation violation,
+it is hard to debug. Another "feature" of the C/C++ split is that the C++
+data structures get converted to C data structures to call the low-level C
+code. This is ugly, and the C++izing of the C code is a step towards
+eliminating the conversion, but it has not happened yet.
+
+
+Directory Structure (ordered by dependency):
+============================================
+ccmain     Top-level code. The main program resides in tesseractmain.cpp.
+display    An "editor" to view and operate on the internal structures.
+           (Requires a working viewer - batteries not included.)
+wordrec    The word-level recognizer.
+textord    The module that organizes(orders) text into lines and words.
+classify   The low-level character classifiers.
+ccstruct   Classes to hold information about a page as it is being processed.
+viewer     The client side of a client server viewing system.
+           Unfortunately, at this time, the server side is not available.
+image      Image class and processing functions.
+dict       Language model code.
+cutil      Code for file I/O, lists, heaps etc, from the old C code.
+ccutil     Somewhat newer code for lists, memory allocation etc from the
+           old C++ code.
+
+
+About the Engine
+================
+This code is a raw OCR engine. It has NO PAGE LAYOUT ANALYSIS, NO OUTPUT
+FORMATTING, and NO UI. It can only process an image of a single column
+and create text from it. It can detect fixed pitch vs proportional text.
+Having said that, in 1995, this engine was in the top 3 in terms of character
+accuracy, and it compiles and runs on both Linux and Windows. Another current
+limitation is that it only recognizes English and its character set is only
+US-ASCII. Training code IS included in the open source release however, and
+will be included in a future release.
+
+
+Using the Engine
+================
+The usage of both Windows and Linux versions is the same.
+The executable must reside in the same directory as the tessdata directory
+The command line is:
+tesseract <image.tif> <output> batch
+The image file requires an .tif extension for its type to be recognized
+correctly. If a file exists with the .tif extension replaced by .uzn, then it
+will be interpreted as a UNLV-style zone file. (See www.isri.unlv.edu for
+details of the zone files.)
+

ReleaseNotes

@@ -0,0 +1,78 @@
+Tesseract release notes Feb 2, 2007 - V1.03.
+Added mftraining and cntraining. Using an image with a box file, tesseract
+generates .tr output files. cntraining runs on the .tr files to make
+normproto that lives in tessdata. mftraining runs on the .tr files to
+make inttemp and pffmtable in tessdata. These are the main data files
+that tesseract uses to recognize characters. At present, the code to make
+dictionary files is not yet available, nor are any sample box files or
+rebuilt inttemp or documentation to create any of these. Recognition is
+still limited to the ASCII set, but when this problem is fixed, documentation
+will follow.
+
+Added a new API with adaptive thresholding for grey and color images.
+See ccmain/baseapi.h/cpp for details. The main program has been converted
+to use the API as an example. See main() in ccmain/tesseractmain.cpp for
+details. The API is designed to make it easy to add subclasses with ability
+to output the bounding boxes etc from the internal structures. The adaptive
+thresholding improves accuracy (most of the time) on non-binary images.
+
+Many memory leaks have been fixed. There are no known leaks left from using
+the API correctly.
+
+The adaptive classifier was not operating correctly. This bug, and several
+others have been fixed, including poor chopping, an indefinite (if not quite
+infinite) loop in the number parser, and a couple of crash bugs. Thanks to
+all that have contributed bugs and bug fixes.
+
+It is now possible to build without any of the graphics support to save code
+size using #define GRAPHICS_DISABLED. There is also a new EMBEDDED define
+for use on operating systems with limited library support.
+
+64-bit and Mac OSX buildability is now included in the mainline source tree.
+Thanks to all that have contributed patches and comments to help with that.
+1.03 is also endian-independent, apart from the tiff i/o, so if you use
+libtiff, the code should run on all platforms, even if you get/create new
+data files of a different endinanness.
+
+Some of the bug fixes improve accuracy, and so do some of the changes to
+DangAmbigs and user-words.
+
+Tesseract release notes, Oct 4 2006 - V1.02.
+Removed dependency on aspirin. *All* code is now licensed under Apache2.0.
+
+Tesseract release notes, Sep 7 2006 - V1.01.
+
+Fixes for this release:
+Added mfcpch.cpp and getopt.cpp for VC++.
+Fixed problem with greyscale images and no libtiff.
+Stopped debug window from being used for the usage output.
+Fixed load of inttemp for big-endian architectures.
+Fixed some Mac compilation issues.
+
+This version should read uncompressed 8 bit grey and 24 bit color tiffs
+without having to have libtiff. It does a dumb threshold though, so don't
+expect good results from poor contrast or images of natural scenes etc.
+
+If you just run tesseract with no command line args you should now get a
+sensible usage message on linux, with or without X-windows.
+
+If you can get it to compile on a PPC Mac, it may now run correctly,
+although not all the build issues are fixed yet.
+
+Building Tesseract:
+Windows:
+Unpack the tar.gz archive
+Open tesseract.dsw in DevStudio (preferably version 6, higher versions will be more difficult)
+Set Win32 - Release as the active configuration.
+Build.
+Copy tesseract.exe from bin.rel up one directory level.
+Run tesseract phototest.tif phototest
+This will create phototest.txt.
+
+Linux:
+Unpack the tar.gz archive
+./configure
+make
+Copy tesseract from ccmain up one directory level (or create a symbolic link)
+Run tesseract phototest.tif phototest
+This will create phototest.txt.

acinclude.m4

@@ -0,0 +1,10 @@
+# Master include for AC macros. This directory structure allows
+# for more flexibility with respect to CVS modules.
+#
+# Author: Luc Vincent
+
+### m4_include(config/ac_compile_check_sizeof.m4)dnl
+#m4_include(config/ac_create_stdint_h.m4)dnl
+#m4_include(config/ax_create_stdint_h.m4)dnl
+m4_include(config/ac_define_versionlevel.m4)dnl
+m4_include(config/acinclude_custom.m4)dnl

aclocal.m4

@@ -0,0 +1,920 @@
+# generated automatically by aclocal 1.9.6 -*- Autoconf -*-
+
+# Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
+# 2005  Free Software Foundation, Inc.
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY, to the extent permitted by law; without
+# even the implied warranty of MERCHANTABILITY or FITNESS FOR A
+# PARTICULAR PURPOSE.
+
+# Copyright (C) 2002, 2003, 2005  Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# AM_AUTOMAKE_VERSION(VERSION)
+# ----------------------------
+# Automake X.Y traces this macro to ensure aclocal.m4 has been
+# generated from the m4 files accompanying Automake X.Y.
+AC_DEFUN([AM_AUTOMAKE_VERSION], [am__api_version="1.9"])
+
+# AM_SET_CURRENT_AUTOMAKE_VERSION
+# -------------------------------
+# Call AM_AUTOMAKE_VERSION so it can be traced.
+# This function is AC_REQUIREd by AC_INIT_AUTOMAKE.
+AC_DEFUN([AM_SET_CURRENT_AUTOMAKE_VERSION],
+	 [AM_AUTOMAKE_VERSION([1.9.6])])
+
+# AM_AUX_DIR_EXPAND                                         -*- Autoconf -*-
+
+# Copyright (C) 2001, 2003, 2005  Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# For projects using AC_CONFIG_AUX_DIR([foo]), Autoconf sets
+# $ac_aux_dir to `$srcdir/foo'.  In other projects, it is set to
+# `$srcdir', `$srcdir/..', or `$srcdir/../..'.
+#
+# Of course, Automake must honor this variable whenever it calls a
+# tool from the auxiliary directory.  The problem is that $srcdir (and
+# therefore $ac_aux_dir as well) can be either absolute or relative,
+# depending on how configure is run.  This is pretty annoying, since
+# it makes $ac_aux_dir quite unusable in subdirectories: in the top
+# source directory, any form will work fine, but in subdirectories a
+# relative path needs to be adjusted first.
+#
+# $ac_aux_dir/missing
+#    fails when called from a subdirectory if $ac_aux_dir is relative
+# $top_srcdir/$ac_aux_dir/missing
+#    fails if $ac_aux_dir is absolute,
+#    fails when called from a subdirectory in a VPATH build with
+#          a relative $ac_aux_dir
+#
+# The reason of the latter failure is that $top_srcdir and $ac_aux_dir
+# are both prefixed by $srcdir.  In an in-source build this is usually
+# harmless because $srcdir is `.', but things will broke when you
+# start a VPATH build or use an absolute $srcdir.
+#
+# So we could use something similar to $top_srcdir/$ac_aux_dir/missing,
+# iff we strip the leading $srcdir from $ac_aux_dir.  That would be:
+#   am_aux_dir='\$(top_srcdir)/'`expr "$ac_aux_dir" : "$srcdir//*\(.*\)"`
+# and then we would define $MISSING as
+#   MISSING="\${SHELL} $am_aux_dir/missing"
+# This will work as long as MISSING is not called from configure, because
+# unfortunately $(top_srcdir) has no meaning in configure.
+# However there are other variables, like CC, which are often used in
+# configure, and could therefore not use this "fixed" $ac_aux_dir.
+#
+# Another solution, used here, is to always expand $ac_aux_dir to an
+# absolute PATH.  The drawback is that using absolute paths prevent a
+# configured tree to be moved without reconfiguration.
+
+AC_DEFUN([AM_AUX_DIR_EXPAND],
+[dnl Rely on autoconf to set up CDPATH properly.
+AC_PREREQ([2.50])dnl
+# expand $ac_aux_dir to an absolute path
+am_aux_dir=`cd $ac_aux_dir && pwd`
+])
+
+# AM_CONDITIONAL                                            -*- Autoconf -*-
+
+# Copyright (C) 1997, 2000, 2001, 2003, 2004, 2005
+# Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# serial 7
+
+# AM_CONDITIONAL(NAME, SHELL-CONDITION)
+# -------------------------------------
+# Define a conditional.
+AC_DEFUN([AM_CONDITIONAL],
+[AC_PREREQ(2.52)dnl
+ ifelse([$1], [TRUE],  [AC_FATAL([$0: invalid condition: $1])],
+	[$1], [FALSE], [AC_FATAL([$0: invalid condition: $1])])dnl
+AC_SUBST([$1_TRUE])
+AC_SUBST([$1_FALSE])
+if $2; then
+  $1_TRUE=
+  $1_FALSE='#'
+else
+  $1_TRUE='#'
+  $1_FALSE=
+fi
+AC_CONFIG_COMMANDS_PRE(
+[if test -z "${$1_TRUE}" && test -z "${$1_FALSE}"; then
+  AC_MSG_ERROR([[conditional "$1" was never defined.
+Usually this means the macro was only invoked conditionally.]])
+fi])])
+
+
+# Copyright (C) 1999, 2000, 2001, 2002, 2003, 2004, 2005
+# Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# serial 8
+
+# There are a few dirty hacks below to avoid letting `AC_PROG_CC' be
+# written in clear, in which case automake, when reading aclocal.m4,
+# will think it sees a *use*, and therefore will trigger all it's
+# C support machinery.  Also note that it means that autoscan, seeing
+# CC etc. in the Makefile, will ask for an AC_PROG_CC use...
+
+
+# _AM_DEPENDENCIES(NAME)
+# ----------------------
+# See how the compiler implements dependency checking.
+# NAME is "CC", "CXX", "GCJ", or "OBJC".
+# We try a few techniques and use that to set a single cache variable.
+#
+# We don't AC_REQUIRE the corresponding AC_PROG_CC since the latter was
+# modified to invoke _AM_DEPENDENCIES(CC); we would have a circular
+# dependency, and given that the user is not expected to run this macro,
+# just rely on AC_PROG_CC.
+AC_DEFUN([_AM_DEPENDENCIES],
+[AC_REQUIRE([AM_SET_DEPDIR])dnl
+AC_REQUIRE([AM_OUTPUT_DEPENDENCY_COMMANDS])dnl
+AC_REQUIRE([AM_MAKE_INCLUDE])dnl
+AC_REQUIRE([AM_DEP_TRACK])dnl
+
+ifelse([$1], CC,   [depcc="$CC"   am_compiler_list=],
+       [$1], CXX,  [depcc="$CXX"  am_compiler_list=],
+       [$1], OBJC, [depcc="$OBJC" am_compiler_list='gcc3 gcc'],
+       [$1], GCJ,  [depcc="$GCJ"  am_compiler_list='gcc3 gcc'],
+                   [depcc="$$1"   am_compiler_list=])
+
+AC_CACHE_CHECK([dependency style of $depcc],
+               [am_cv_$1_dependencies_compiler_type],
+[if test -z "$AMDEP_TRUE" && test -f "$am_depcomp"; then
+  # We make a subdir and do the tests there.  Otherwise we can end up
+  # making bogus files that we don't know about and never remove.  For
+  # instance it was reported that on HP-UX the gcc test will end up
+  # making a dummy file named `D' -- because `-MD' means `put the output
+  # in D'.
+  mkdir conftest.dir
+  # Copy depcomp to subdir because otherwise we won't find it if we're
+  # using a relative directory.
+  cp "$am_depcomp" conftest.dir
+  cd conftest.dir
+  # We will build objects and dependencies in a subdirectory because
+  # it helps to detect inapplicable dependency modes.  For instance
+  # both Tru64's cc and ICC support -MD to output dependencies as a
+  # side effect of compilation, but ICC will put the dependencies in
+  # the current directory while Tru64 will put them in the object
+  # directory.
+  mkdir sub
+
+  am_cv_$1_dependencies_compiler_type=none
+  if test "$am_compiler_list" = ""; then
+     am_compiler_list=`sed -n ['s/^#*\([a-zA-Z0-9]*\))$/\1/p'] < ./depcomp`
+  fi
+  for depmode in $am_compiler_list; do
+    # Setup a source with many dependencies, because some compilers
+    # like to wrap large dependency lists on column 80 (with \), and
+    # we should not choose a depcomp mode which is confused by this.
+    #
+    # We need to recreate these files for each test, as the compiler may
+    # overwrite some of them when testing with obscure command lines.
+    # This happens at least with the AIX C compiler.
+    : > sub/conftest.c
+    for i in 1 2 3 4 5 6; do
+      echo '#include "conftst'$i'.h"' >> sub/conftest.c
+      # Using `: > sub/conftst$i.h' creates only sub/conftst1.h with
+      # Solaris 8's {/usr,}/bin/sh.
+      touch sub/conftst$i.h
+    done
+    echo "${am__include} ${am__quote}sub/conftest.Po${am__quote}" > confmf
+
+    case $depmode in
+    nosideeffect)
+      # after this tag, mechanisms are not by side-effect, so they'll
+      # only be used when explicitly requested
+      if test "x$enable_dependency_tracking" = xyes; then
+	continue
+      else
+	break
+      fi
+      ;;
+    none) break ;;
+    esac
+    # We check with `-c' and `-o' for the sake of the "dashmstdout"
+    # mode.  It turns out that the SunPro C++ compiler does not properly
+    # handle `-M -o', and we need to detect this.
+    if depmode=$depmode \
+       source=sub/conftest.c object=sub/conftest.${OBJEXT-o} \
+       depfile=sub/conftest.Po tmpdepfile=sub/conftest.TPo \
+       $SHELL ./depcomp $depcc -c -o sub/conftest.${OBJEXT-o} sub/conftest.c \
+         >/dev/null 2>conftest.err &&
+       grep sub/conftst6.h sub/conftest.Po > /dev/null 2>&1 &&
+       grep sub/conftest.${OBJEXT-o} sub/conftest.Po > /dev/null 2>&1 &&
+       ${MAKE-make} -s -f confmf > /dev/null 2>&1; then
+      # icc doesn't choke on unknown options, it will just issue warnings
+      # or remarks (even with -Werror).  So we grep stderr for any message
+      # that says an option was ignored or not supported.
+      # When given -MP, icc 7.0 and 7.1 complain thusly:
+      #   icc: Command line warning: ignoring option '-M'; no argument required
+      # The diagnosis changed in icc 8.0:
+      #   icc: Command line remark: option '-MP' not supported
+      if (grep 'ignoring option' conftest.err ||
+          grep 'not supported' conftest.err) >/dev/null 2>&1; then :; else
+        am_cv_$1_dependencies_compiler_type=$depmode
+        break
+      fi
+    fi
+  done
+
+  cd ..
+  rm -rf conftest.dir
+else
+  am_cv_$1_dependencies_compiler_type=none
+fi
+])
+AC_SUBST([$1DEPMODE], [depmode=$am_cv_$1_dependencies_compiler_type])
+AM_CONDITIONAL([am__fastdep$1], [
+  test "x$enable_dependency_tracking" != xno \
+  && test "$am_cv_$1_dependencies_compiler_type" = gcc3])
+])
+
+
+# AM_SET_DEPDIR
+# -------------
+# Choose a directory name for dependency files.
+# This macro is AC_REQUIREd in _AM_DEPENDENCIES
+AC_DEFUN([AM_SET_DEPDIR],
+[AC_REQUIRE([AM_SET_LEADING_DOT])dnl
+AC_SUBST([DEPDIR], ["${am__leading_dot}deps"])dnl
+])
+
+
+# AM_DEP_TRACK
+# ------------
+AC_DEFUN([AM_DEP_TRACK],
+[AC_ARG_ENABLE(dependency-tracking,
+[  --disable-dependency-tracking  speeds up one-time build
+  --enable-dependency-tracking   do not reject slow dependency extractors])
+if test "x$enable_dependency_tracking" != xno; then
+  am_depcomp="$ac_aux_dir/depcomp"
+  AMDEPBACKSLASH='\'
+fi
+AM_CONDITIONAL([AMDEP], [test "x$enable_dependency_tracking" != xno])
+AC_SUBST([AMDEPBACKSLASH])
+])
+
+# Generate code to set up dependency tracking.              -*- Autoconf -*-
+
+# Copyright (C) 1999, 2000, 2001, 2002, 2003, 2004, 2005
+# Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+#serial 3
+
+# _AM_OUTPUT_DEPENDENCY_COMMANDS
+# ------------------------------
+AC_DEFUN([_AM_OUTPUT_DEPENDENCY_COMMANDS],
+[for mf in $CONFIG_FILES; do
+  # Strip MF so we end up with the name of the file.
+  mf=`echo "$mf" | sed -e 's/:.*$//'`
+  # Check whether this is an Automake generated Makefile or not.
+  # We used to match only the files named `Makefile.in', but
+  # some people rename them; so instead we look at the file content.
+  # Grep'ing the first line is not enough: some people post-process
+  # each Makefile.in and add a new line on top of each file to say so.
+  # So let's grep whole file.
+  if grep '^#.*generated by automake' $mf > /dev/null 2>&1; then
+    dirpart=`AS_DIRNAME("$mf")`
+  else
+    continue
+  fi
+  # Extract the definition of DEPDIR, am__include, and am__quote
+  # from the Makefile without running `make'.
+  DEPDIR=`sed -n 's/^DEPDIR = //p' < "$mf"`
+  test -z "$DEPDIR" && continue
+  am__include=`sed -n 's/^am__include = //p' < "$mf"`
+  test -z "am__include" && continue
+  am__quote=`sed -n 's/^am__quote = //p' < "$mf"`
+  # When using ansi2knr, U may be empty or an underscore; expand it
+  U=`sed -n 's/^U = //p' < "$mf"`
+  # Find all dependency output files, they are included files with
+  # $(DEPDIR) in their names.  We invoke sed twice because it is the
+  # simplest approach to changing $(DEPDIR) to its actual value in the
+  # expansion.
+  for file in `sed -n "
+    s/^$am__include $am__quote\(.*(DEPDIR).*\)$am__quote"'$/\1/p' <"$mf" | \
+       sed -e 's/\$(DEPDIR)/'"$DEPDIR"'/g' -e 's/\$U/'"$U"'/g'`; do
+    # Make sure the directory exists.
+    test -f "$dirpart/$file" && continue
+    fdir=`AS_DIRNAME(["$file"])`
+    AS_MKDIR_P([$dirpart/$fdir])
+    # echo "creating $dirpart/$file"
+    echo '# dummy' > "$dirpart/$file"
+  done
+done
+])# _AM_OUTPUT_DEPENDENCY_COMMANDS
+
+
+# AM_OUTPUT_DEPENDENCY_COMMANDS
+# -----------------------------
+# This macro should only be invoked once -- use via AC_REQUIRE.
+#
+# This code is only required when automatic dependency tracking
+# is enabled.  FIXME.  This creates each `.P' file that we will
+# need in order to bootstrap the dependency handling code.
+AC_DEFUN([AM_OUTPUT_DEPENDENCY_COMMANDS],
+[AC_CONFIG_COMMANDS([depfiles],
+     [test x"$AMDEP_TRUE" != x"" || _AM_OUTPUT_DEPENDENCY_COMMANDS],
+     [AMDEP_TRUE="$AMDEP_TRUE" ac_aux_dir="$ac_aux_dir"])
+])
+
+# Copyright (C) 1996, 1997, 2000, 2001, 2003, 2005
+# Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# serial 8
+
+# AM_CONFIG_HEADER is obsolete.  It has been replaced by AC_CONFIG_HEADERS.
+AU_DEFUN([AM_CONFIG_HEADER], [AC_CONFIG_HEADERS($@)])
+
+# Do all the work for Automake.                             -*- Autoconf -*-
+
+# Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005
+# Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# serial 12
+
+# This macro actually does too much.  Some checks are only needed if
+# your package does certain things.  But this isn't really a big deal.
+
+# AM_INIT_AUTOMAKE(PACKAGE, VERSION, [NO-DEFINE])
+# AM_INIT_AUTOMAKE([OPTIONS])
+# -----------------------------------------------
+# The call with PACKAGE and VERSION arguments is the old style
+# call (pre autoconf-2.50), which is being phased out.  PACKAGE
+# and VERSION should now be passed to AC_INIT and removed from
+# the call to AM_INIT_AUTOMAKE.
+# We support both call styles for the transition.  After
+# the next Automake release, Autoconf can make the AC_INIT
+# arguments mandatory, and then we can depend on a new Autoconf
+# release and drop the old call support.
+AC_DEFUN([AM_INIT_AUTOMAKE],
+[AC_PREREQ([2.58])dnl
+dnl Autoconf wants to disallow AM_ names.  We explicitly allow
+dnl the ones we care about.
+m4_pattern_allow([^AM_[A-Z]+FLAGS$])dnl
+AC_REQUIRE([AM_SET_CURRENT_AUTOMAKE_VERSION])dnl
+AC_REQUIRE([AC_PROG_INSTALL])dnl
+# test to see if srcdir already configured
+if test "`cd $srcdir && pwd`" != "`pwd`" &&
+   test -f $srcdir/config.status; then
+  AC_MSG_ERROR([source directory already configured; run "make distclean" there first])
+fi
+
+# test whether we have cygpath
+if test -z "$CYGPATH_W"; then
+  if (cygpath --version) >/dev/null 2>/dev/null; then
+    CYGPATH_W='cygpath -w'
+  else
+    CYGPATH_W=echo
+  fi
+fi
+AC_SUBST([CYGPATH_W])
+
+# Define the identity of the package.
+dnl Distinguish between old-style and new-style calls.
+m4_ifval([$2],
+[m4_ifval([$3], [_AM_SET_OPTION([no-define])])dnl
+ AC_SUBST([PACKAGE], [$1])dnl
+ AC_SUBST([VERSION], [$2])],
+[_AM_SET_OPTIONS([$1])dnl
+ AC_SUBST([PACKAGE], ['AC_PACKAGE_TARNAME'])dnl
+ AC_SUBST([VERSION], ['AC_PACKAGE_VERSION'])])dnl
+
+_AM_IF_OPTION([no-define],,
+[AC_DEFINE_UNQUOTED(PACKAGE, "$PACKAGE", [Name of package])
+ AC_DEFINE_UNQUOTED(VERSION, "$VERSION", [Version number of package])])dnl
+
+# Some tools Automake needs.
+AC_REQUIRE([AM_SANITY_CHECK])dnl
+AC_REQUIRE([AC_ARG_PROGRAM])dnl
+AM_MISSING_PROG(ACLOCAL, aclocal-${am__api_version})
+AM_MISSING_PROG(AUTOCONF, autoconf)
+AM_MISSING_PROG(AUTOMAKE, automake-${am__api_version})
+AM_MISSING_PROG(AUTOHEADER, autoheader)
+AM_MISSING_PROG(MAKEINFO, makeinfo)
+AM_PROG_INSTALL_SH
+AM_PROG_INSTALL_STRIP
+AC_REQUIRE([AM_PROG_MKDIR_P])dnl
+# We need awk for the "check" target.  The system "awk" is bad on
+# some platforms.
+AC_REQUIRE([AC_PROG_AWK])dnl
+AC_REQUIRE([AC_PROG_MAKE_SET])dnl
+AC_REQUIRE([AM_SET_LEADING_DOT])dnl
+_AM_IF_OPTION([tar-ustar], [_AM_PROG_TAR([ustar])],
+              [_AM_IF_OPTION([tar-pax], [_AM_PROG_TAR([pax])],
+	      		     [_AM_PROG_TAR([v7])])])
+_AM_IF_OPTION([no-dependencies],,
+[AC_PROVIDE_IFELSE([AC_PROG_CC],
+                  [_AM_DEPENDENCIES(CC)],
+                  [define([AC_PROG_CC],
+                          defn([AC_PROG_CC])[_AM_DEPENDENCIES(CC)])])dnl
+AC_PROVIDE_IFELSE([AC_PROG_CXX],
+                  [_AM_DEPENDENCIES(CXX)],
+                  [define([AC_PROG_CXX],
+                          defn([AC_PROG_CXX])[_AM_DEPENDENCIES(CXX)])])dnl
+])
+])
+
+
+# When config.status generates a header, we must update the stamp-h file.
+# This file resides in the same directory as the config header
+# that is generated.  The stamp files are numbered to have different names.
+
+# Autoconf calls _AC_AM_CONFIG_HEADER_HOOK (when defined) in the
+# loop where config.status creates the headers, so we can generate
+# our stamp files there.
+AC_DEFUN([_AC_AM_CONFIG_HEADER_HOOK],
+[# Compute $1's index in $config_headers.
+_am_stamp_count=1
+for _am_header in $config_headers :; do
+  case $_am_header in
+    $1 | $1:* )
+      break ;;
+    * )
+      _am_stamp_count=`expr $_am_stamp_count + 1` ;;
+  esac
+done
+echo "timestamp for $1" >`AS_DIRNAME([$1])`/stamp-h[]$_am_stamp_count])
+
+# Copyright (C) 2001, 2003, 2005  Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# AM_PROG_INSTALL_SH
+# ------------------
+# Define $install_sh.
+AC_DEFUN([AM_PROG_INSTALL_SH],
+[AC_REQUIRE([AM_AUX_DIR_EXPAND])dnl
+install_sh=${install_sh-"$am_aux_dir/install-sh"}
+AC_SUBST(install_sh)])
+
+# Copyright (C) 2003, 2005  Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# serial 2
+
+# Check whether the underlying file-system supports filenames
+# with a leading dot.  For instance MS-DOS doesn't.
+AC_DEFUN([AM_SET_LEADING_DOT],
+[rm -rf .tst 2>/dev/null
+mkdir .tst 2>/dev/null
+if test -d .tst; then
+  am__leading_dot=.
+else
+  am__leading_dot=_
+fi
+rmdir .tst 2>/dev/null
+AC_SUBST([am__leading_dot])])
+
+# Add --enable-maintainer-mode option to configure.         -*- Autoconf -*-
+# From Jim Meyering
+
+# Copyright (C) 1996, 1998, 2000, 2001, 2002, 2003, 2004, 2005
+# Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# serial 4
+
+AC_DEFUN([AM_MAINTAINER_MODE],
+[AC_MSG_CHECKING([whether to enable maintainer-specific portions of Makefiles])
+  dnl maintainer-mode is disabled by default
+  AC_ARG_ENABLE(maintainer-mode,
+[  --enable-maintainer-mode  enable make rules and dependencies not useful
+			  (and sometimes confusing) to the casual installer],
+      USE_MAINTAINER_MODE=$enableval,
+      USE_MAINTAINER_MODE=no)
+  AC_MSG_RESULT([$USE_MAINTAINER_MODE])
+  AM_CONDITIONAL(MAINTAINER_MODE, [test $USE_MAINTAINER_MODE = yes])
+  MAINT=$MAINTAINER_MODE_TRUE
+  AC_SUBST(MAINT)dnl
+]
+)
+
+AU_DEFUN([jm_MAINTAINER_MODE], [AM_MAINTAINER_MODE])
+
+# Check to see how 'make' treats includes.	            -*- Autoconf -*-
+
+# Copyright (C) 2001, 2002, 2003, 2005  Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# serial 3
+
+# AM_MAKE_INCLUDE()
+# -----------------
+# Check to see how make treats includes.
+AC_DEFUN([AM_MAKE_INCLUDE],
+[am_make=${MAKE-make}
+cat > confinc << 'END'
+am__doit:
+	@echo done
+.PHONY: am__doit
+END
+# If we don't find an include directive, just comment out the code.
+AC_MSG_CHECKING([for style of include used by $am_make])
+am__include="#"
+am__quote=
+_am_result=none
+# First try GNU make style include.
+echo "include confinc" > confmf
+# We grep out `Entering directory' and `Leaving directory'
+# messages which can occur if `w' ends up in MAKEFLAGS.
+# In particular we don't look at `^make:' because GNU make might
+# be invoked under some other name (usually "gmake"), in which
+# case it prints its new name instead of `make'.
+if test "`$am_make -s -f confmf 2> /dev/null | grep -v 'ing directory'`" = "done"; then
+   am__include=include
+   am__quote=
+   _am_result=GNU
+fi
+# Now try BSD make style include.
+if test "$am__include" = "#"; then
+   echo '.include "confinc"' > confmf
+   if test "`$am_make -s -f confmf 2> /dev/null`" = "done"; then
+      am__include=.include
+      am__quote="\""
+      _am_result=BSD
+   fi
+fi
+AC_SUBST([am__include])
+AC_SUBST([am__quote])
+AC_MSG_RESULT([$_am_result])
+rm -f confinc confmf
+])
+
+# Copyright (C) 1999, 2000, 2001, 2003, 2005  Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# serial 3
+
+# AM_PROG_CC_C_O
+# --------------
+# Like AC_PROG_CC_C_O, but changed for automake.
+AC_DEFUN([AM_PROG_CC_C_O],
+[AC_REQUIRE([AC_PROG_CC_C_O])dnl
+AC_REQUIRE([AM_AUX_DIR_EXPAND])dnl
+# FIXME: we rely on the cache variable name because
+# there is no other way.
+set dummy $CC
+ac_cc=`echo $[2] | sed ['s/[^a-zA-Z0-9_]/_/g;s/^[0-9]/_/']`
+if eval "test \"`echo '$ac_cv_prog_cc_'${ac_cc}_c_o`\" != yes"; then
+   # Losing compiler, so override with the script.
+   # FIXME: It is wrong to rewrite CC.
+   # But if we don't then we get into trouble of one sort or another.
+   # A longer-term fix would be to have automake use am__CC in this case,
+   # and then we could set am__CC="\$(top_srcdir)/compile \$(CC)"
+   CC="$am_aux_dir/compile $CC"
+fi
+])
+
+# Fake the existence of programs that GNU maintainers use.  -*- Autoconf -*-
+
+# Copyright (C) 1997, 1999, 2000, 2001, 2003, 2005
+# Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# serial 4
+
+# AM_MISSING_PROG(NAME, PROGRAM)
+# ------------------------------
+AC_DEFUN([AM_MISSING_PROG],
+[AC_REQUIRE([AM_MISSING_HAS_RUN])
+$1=${$1-"${am_missing_run}$2"}
+AC_SUBST($1)])
+
+
+# AM_MISSING_HAS_RUN
+# ------------------
+# Define MISSING if not defined so far and test if it supports --run.
+# If it does, set am_missing_run to use it, otherwise, to nothing.
+AC_DEFUN([AM_MISSING_HAS_RUN],
+[AC_REQUIRE([AM_AUX_DIR_EXPAND])dnl
+test x"${MISSING+set}" = xset || MISSING="\${SHELL} $am_aux_dir/missing"
+# Use eval to expand $SHELL
+if eval "$MISSING --run true"; then
+  am_missing_run="$MISSING --run "
+else
+  am_missing_run=
+  AC_MSG_WARN([`missing' script is too old or missing])
+fi
+])
+
+# Copyright (C) 2003, 2004, 2005  Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# AM_PROG_MKDIR_P
+# ---------------
+# Check whether `mkdir -p' is supported, fallback to mkinstalldirs otherwise.
+#
+# Automake 1.8 used `mkdir -m 0755 -p --' to ensure that directories
+# created by `make install' are always world readable, even if the
+# installer happens to have an overly restrictive umask (e.g. 077).
+# This was a mistake.  There are at least two reasons why we must not
+# use `-m 0755':
+#   - it causes special bits like SGID to be ignored,
+#   - it may be too restrictive (some setups expect 775 directories).
+#
+# Do not use -m 0755 and let people choose whatever they expect by
+# setting umask.
+#
+# We cannot accept any implementation of `mkdir' that recognizes `-p'.
+# Some implementations (such as Solaris 8's) are not thread-safe: if a
+# parallel make tries to run `mkdir -p a/b' and `mkdir -p a/c'
+# concurrently, both version can detect that a/ is missing, but only
+# one can create it and the other will error out.  Consequently we
+# restrict ourselves to GNU make (using the --version option ensures
+# this.)
+AC_DEFUN([AM_PROG_MKDIR_P],
+[if mkdir -p --version . >/dev/null 2>&1 && test ! -d ./--version; then
+  # We used to keeping the `.' as first argument, in order to
+  # allow $(mkdir_p) to be used without argument.  As in
+  #   $(mkdir_p) $(somedir)
+  # where $(somedir) is conditionally defined.  However this is wrong
+  # for two reasons:
+  #  1. if the package is installed by a user who cannot write `.'
+  #     make install will fail,
+  #  2. the above comment should most certainly read
+  #     $(mkdir_p) $(DESTDIR)$(somedir)
+  #     so it does not work when $(somedir) is undefined and
+  #     $(DESTDIR) is not.
+  #  To support the latter case, we have to write
+  #     test -z "$(somedir)" || $(mkdir_p) $(DESTDIR)$(somedir),
+  #  so the `.' trick is pointless.
+  mkdir_p='mkdir -p --'
+else
+  # On NextStep and OpenStep, the `mkdir' command does not
+  # recognize any option.  It will interpret all options as
+  # directories to create, and then abort because `.' already
+  # exists.
+  for d in ./-p ./--version;
+  do
+    test -d $d && rmdir $d
+  done
+  # $(mkinstalldirs) is defined by Automake if mkinstalldirs exists.
+  if test -f "$ac_aux_dir/mkinstalldirs"; then
+    mkdir_p='$(mkinstalldirs)'
+  else
+    mkdir_p='$(install_sh) -d'
+  fi
+fi
+AC_SUBST([mkdir_p])])
+
+# Helper functions for option handling.                     -*- Autoconf -*-
+
+# Copyright (C) 2001, 2002, 2003, 2005  Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# serial 3
+
+# _AM_MANGLE_OPTION(NAME)
+# -----------------------
+AC_DEFUN([_AM_MANGLE_OPTION],
+[[_AM_OPTION_]m4_bpatsubst($1, [[^a-zA-Z0-9_]], [_])])
+
+# _AM_SET_OPTION(NAME)
+# ------------------------------
+# Set option NAME.  Presently that only means defining a flag for this option.
+AC_DEFUN([_AM_SET_OPTION],
+[m4_define(_AM_MANGLE_OPTION([$1]), 1)])
+
+# _AM_SET_OPTIONS(OPTIONS)
+# ----------------------------------
+# OPTIONS is a space-separated list of Automake options.
+AC_DEFUN([_AM_SET_OPTIONS],
+[AC_FOREACH([_AM_Option], [$1], [_AM_SET_OPTION(_AM_Option)])])
+
+# _AM_IF_OPTION(OPTION, IF-SET, [IF-NOT-SET])
+# -------------------------------------------
+# Execute IF-SET if OPTION is set, IF-NOT-SET otherwise.
+AC_DEFUN([_AM_IF_OPTION],
+[m4_ifset(_AM_MANGLE_OPTION([$1]), [$2], [$3])])
+
+# Check to make sure that the build environment is sane.    -*- Autoconf -*-
+
+# Copyright (C) 1996, 1997, 2000, 2001, 2003, 2005
+# Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# serial 4
+
+# AM_SANITY_CHECK
+# ---------------
+AC_DEFUN([AM_SANITY_CHECK],
+[AC_MSG_CHECKING([whether build environment is sane])
+# Just in case
+sleep 1
+echo timestamp > conftest.file
+# Do `set' in a subshell so we don't clobber the current shell's
+# arguments.  Must try -L first in case configure is actually a
+# symlink; some systems play weird games with the mod time of symlinks
+# (eg FreeBSD returns the mod time of the symlink's containing
+# directory).
+if (
+   set X `ls -Lt $srcdir/configure conftest.file 2> /dev/null`
+   if test "$[*]" = "X"; then
+      # -L didn't work.
+      set X `ls -t $srcdir/configure conftest.file`
+   fi
+   rm -f conftest.file
+   if test "$[*]" != "X $srcdir/configure conftest.file" \
+      && test "$[*]" != "X conftest.file $srcdir/configure"; then
+
+      # If neither matched, then we have a broken ls.  This can happen
+      # if, for instance, CONFIG_SHELL is bash and it inherits a
+      # broken ls alias from the environment.  This has actually
+      # happened.  Such a system could not be considered "sane".
+      AC_MSG_ERROR([ls -t appears to fail.  Make sure there is not a broken
+alias in your environment])
+   fi
+
+   test "$[2]" = conftest.file
+   )
+then
+   # Ok.
+   :
+else
+   AC_MSG_ERROR([newly created file is older than distributed files!
+Check your system clock])
+fi
+AC_MSG_RESULT(yes)])
+
+# Copyright (C) 2001, 2003, 2005  Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# AM_PROG_INSTALL_STRIP
+# ---------------------
+# One issue with vendor `install' (even GNU) is that you can't
+# specify the program used to strip binaries.  This is especially
+# annoying in cross-compiling environments, where the build's strip
+# is unlikely to handle the host's binaries.
+# Fortunately install-sh will honor a STRIPPROG variable, so we
+# always use install-sh in `make install-strip', and initialize
+# STRIPPROG with the value of the STRIP variable (set by the user).
+AC_DEFUN([AM_PROG_INSTALL_STRIP],
+[AC_REQUIRE([AM_PROG_INSTALL_SH])dnl
+# Installed binaries are usually stripped using `strip' when the user
+# run `make install-strip'.  However `strip' might not be the right
+# tool to use in cross-compilation environments, therefore Automake
+# will honor the `STRIP' environment variable to overrule this program.
+dnl Don't test for $cross_compiling = yes, because it might be `maybe'.
+if test "$cross_compiling" != no; then
+  AC_CHECK_TOOL([STRIP], [strip], :)
+fi
+INSTALL_STRIP_PROGRAM="\${SHELL} \$(install_sh) -c -s"
+AC_SUBST([INSTALL_STRIP_PROGRAM])])
+
+# Check how to create a tarball.                            -*- Autoconf -*-
+
+# Copyright (C) 2004, 2005  Free Software Foundation, Inc.
+#
+# This file is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# serial 2
+
+# _AM_PROG_TAR(FORMAT)
+# --------------------
+# Check how to create a tarball in format FORMAT.
+# FORMAT should be one of `v7', `ustar', or `pax'.
+#
+# Substitute a variable $(am__tar) that is a command
+# writing to stdout a FORMAT-tarball containing the directory
+# $tardir.
+#     tardir=directory && $(am__tar) > result.tar
+#
+# Substitute a variable $(am__untar) that extract such
+# a tarball read from stdin.
+#     $(am__untar) < result.tar
+AC_DEFUN([_AM_PROG_TAR],
+[# Always define AMTAR for backward compatibility.
+AM_MISSING_PROG([AMTAR], [tar])
+m4_if([$1], [v7],
+     [am__tar='${AMTAR} chof - "$$tardir"'; am__untar='${AMTAR} xf -'],
+     [m4_case([$1], [ustar],, [pax],,
+              [m4_fatal([Unknown tar format])])
+AC_MSG_CHECKING([how to create a $1 tar archive])
+# Loop over all known methods to create a tar archive until one works.
+_am_tools='gnutar m4_if([$1], [ustar], [plaintar]) pax cpio none'
+_am_tools=${am_cv_prog_tar_$1-$_am_tools}
+# Do not fold the above two line into one, because Tru64 sh and
+# Solaris sh will not grok spaces in the rhs of `-'.
+for _am_tool in $_am_tools
+do
+  case $_am_tool in
+  gnutar)
+    for _am_tar in tar gnutar gtar;
+    do
+      AM_RUN_LOG([$_am_tar --version]) && break
+    done
+    am__tar="$_am_tar --format=m4_if([$1], [pax], [posix], [$1]) -chf - "'"$$tardir"'
+    am__tar_="$_am_tar --format=m4_if([$1], [pax], [posix], [$1]) -chf - "'"$tardir"'
+    am__untar="$_am_tar -xf -"
+    ;;
+  plaintar)
+    # Must skip GNU tar: if it does not support --format= it doesn't create
+    # ustar tarball either.
+    (tar --version) >/dev/null 2>&1 && continue
+    am__tar='tar chf - "$$tardir"'
+    am__tar_='tar chf - "$tardir"'
+    am__untar='tar xf -'
+    ;;
+  pax)
+    am__tar='pax -L -x $1 -w "$$tardir"'
+    am__tar_='pax -L -x $1 -w "$tardir"'
+    am__untar='pax -r'
+    ;;
+  cpio)
+    am__tar='find "$$tardir" -print | cpio -o -H $1 -L'
+    am__tar_='find "$tardir" -print | cpio -o -H $1 -L'
+    am__untar='cpio -i -H $1 -d'
+    ;;
+  none)
+    am__tar=false
+    am__tar_=false
+    am__untar=false
+    ;;
+  esac
+
+  # If the value was cached, stop now.  We just wanted to have am__tar
+  # and am__untar set.
+  test -n "${am_cv_prog_tar_$1}" && break
+
+  # tar/untar a dummy directory, and stop if the command works
+  rm -rf conftest.dir
+  mkdir conftest.dir
+  echo GrepMe > conftest.dir/file
+  AM_RUN_LOG([tardir=conftest.dir && eval $am__tar_ >conftest.tar])
+  rm -rf conftest.dir
+  if test -s conftest.tar; then
+    AM_RUN_LOG([$am__untar <conftest.tar])
+    grep GrepMe conftest.dir/file >/dev/null 2>&1 && break
+  fi
+done
+rm -rf conftest.dir
+
+AC_CACHE_VAL([am_cv_prog_tar_$1], [am_cv_prog_tar_$1=$_am_tool])
+AC_MSG_RESULT([$am_cv_prog_tar_$1])])
+AC_SUBST([am__tar])
+AC_SUBST([am__untar])
+]) # _AM_PROG_TAR
+
+m4_include([acinclude.m4])

ccmain/Makefile.am

@@ -0,0 +1,41 @@
+SUBDIRS =
+AM_CPPFLAGS = \
+    -I$(top_srcdir)/ccutil -I$(top_srcdir)/ccstruct \
+    -I$(top_srcdir)/image -I$(top_srcdir)/viewer \
+    -I$(top_srcdir)/ccops -I$(top_srcdir)/dict \
+    -I$(top_srcdir)/classify -I$(top_srcdir)/display \
+    -I$(top_srcdir)/wordrec -I$(top_srcdir)/cutil \
+    -I$(top_srcdir)/textord
+
+EXTRA_DIST = \
+    adaptions.h applybox.h baseapi.h blobcmp.h \
+    callnet.h charcut.h \
+    control.h docqual.h expandblob.h fixspace.h fixxht.h \
+    imgscale.h matmatch.h output.h paircmp.h reject.h scaleimg.h \
+    tessbox.h tessedit.h tesseractmain.h tessvars.h tfacep.h \
+    tessembedded.h tfacepp.h tstruct.h werdit.h
+
+noinst_LIBRARIES = libtesseract_main.a
+libtesseract_main_a_SOURCES = \
+    tessedit.cpp adaptions.cpp applybox.cpp \
+    baseapi.cpp blobcmp.cpp \
+    callnet.cpp charcut.cpp charsample.cpp control.cpp \
+    docqual.cpp expandblob.cpp fixspace.cpp fixxht.cpp \
+    imgscale.cpp matmatch.cpp output.cpp paircmp.cpp \
+    reject.cpp scaleimg.cpp tessbox.cpp tessvars.cpp \
+    tfacepp.cpp tstruct.cpp werdit.cpp
+
+bin_PROGRAMS = tesseract
+tesseract_SOURCES = tesseractmain.cpp
+tesseract_LDADD = \
+    libtesseract_main.a \
+    ../display/libtesseract_display.a \
+    ../textord/libtesseract_textord.a \
+    ../wordrec/libtesseract_wordrec.a \
+    ../classify/libtesseract_classify.a \
+    ../dict/libtesseract_dict.a \
+    ../viewer/libtesseract_viewer.a \
+    ../image/libtesseract_image.a \
+    ../cutil/libtesseract_cutil.a \
+    ../ccstruct/libtesseract_ccstruct.a \
+    ../ccutil/libtesseract_ccutil.a

ccmain/Makefile.in

@@ -0,0 +1,636 @@
+# Makefile.in generated by automake 1.9.6 from Makefile.am.
+# @configure_input@
+
+# Copyright (C) 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002,
+# 2003, 2004, 2005  Free Software Foundation, Inc.
+# This Makefile.in is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY, to the extent permitted by law; without
+# even the implied warranty of MERCHANTABILITY or FITNESS FOR A
+# PARTICULAR PURPOSE.
+
+@SET_MAKE@
+
+
+srcdir = @srcdir@
+top_srcdir = @top_srcdir@
+VPATH = @srcdir@
+pkgdatadir = $(datadir)/@PACKAGE@
+pkglibdir = $(libdir)/@PACKAGE@
+pkgincludedir = $(includedir)/@PACKAGE@
+top_builddir = ..
+am__cd = CDPATH="$${ZSH_VERSION+.}$(PATH_SEPARATOR)" && cd
+INSTALL = @INSTALL@
+install_sh_DATA = $(install_sh) -c -m 644
+install_sh_PROGRAM = $(install_sh) -c
+install_sh_SCRIPT = $(install_sh) -c
+INSTALL_HEADER = $(INSTALL_DATA)
+transform = $(program_transform_name)
+NORMAL_INSTALL = :
+PRE_INSTALL = :
+POST_INSTALL = :
+NORMAL_UNINSTALL = :
+PRE_UNINSTALL = :
+POST_UNINSTALL = :
+build_triplet = @build@
+host_triplet = @host@
+bin_PROGRAMS = tesseract$(EXEEXT)
+subdir = ccmain
+DIST_COMMON = $(srcdir)/Makefile.am $(srcdir)/Makefile.in
+ACLOCAL_M4 = $(top_srcdir)/aclocal.m4
+am__aclocal_m4_deps = $(top_srcdir)/acinclude.m4 \
+	$(top_srcdir)/config/ac_define_versionlevel.m4 \
+	$(top_srcdir)/config/acinclude_custom.m4 \
+	$(top_srcdir)/configure.ac
+am__configure_deps = $(am__aclocal_m4_deps) $(CONFIGURE_DEPENDENCIES) \
+	$(ACLOCAL_M4)
+mkinstalldirs = $(SHELL) $(top_srcdir)/config/mkinstalldirs
+CONFIG_HEADER = $(top_builddir)/config_auto.h
+CONFIG_CLEAN_FILES =
+LIBRARIES = $(noinst_LIBRARIES)
+AR = ar
+ARFLAGS = cru
+libtesseract_main_a_AR = $(AR) $(ARFLAGS)
+libtesseract_main_a_LIBADD =
+am_libtesseract_main_a_OBJECTS = tessedit.$(OBJEXT) \
+	adaptions.$(OBJEXT) applybox.$(OBJEXT) baseapi.$(OBJEXT) \
+	blobcmp.$(OBJEXT) callnet.$(OBJEXT) charcut.$(OBJEXT) \
+	charsample.$(OBJEXT) control.$(OBJEXT) docqual.$(OBJEXT) \
+	expandblob.$(OBJEXT) fixspace.$(OBJEXT) fixxht.$(OBJEXT) \
+	imgscale.$(OBJEXT) matmatch.$(OBJEXT) output.$(OBJEXT) \
+	paircmp.$(OBJEXT) reject.$(OBJEXT) scaleimg.$(OBJEXT) \
+	tessbox.$(OBJEXT) tessvars.$(OBJEXT) tfacepp.$(OBJEXT) \
+	tstruct.$(OBJEXT) werdit.$(OBJEXT)
+libtesseract_main_a_OBJECTS = $(am_libtesseract_main_a_OBJECTS)
+am__installdirs = "$(DESTDIR)$(bindir)"
+binPROGRAMS_INSTALL = $(INSTALL_PROGRAM)
+PROGRAMS = $(bin_PROGRAMS)
+am_tesseract_OBJECTS = tesseractmain.$(OBJEXT)
+tesseract_OBJECTS = $(am_tesseract_OBJECTS)
+tesseract_DEPENDENCIES = libtesseract_main.a \
+	../display/libtesseract_display.a \
+	../textord/libtesseract_textord.a \
+	../wordrec/libtesseract_wordrec.a \
+	../classify/libtesseract_classify.a \
+	../dict/libtesseract_dict.a ../viewer/libtesseract_viewer.a \
+	../image/libtesseract_image.a ../cutil/libtesseract_cutil.a \
+	../ccstruct/libtesseract_ccstruct.a \
+	../ccutil/libtesseract_ccutil.a
+DEFAULT_INCLUDES = -I. -I$(srcdir) -I$(top_builddir)
+depcomp = $(SHELL) $(top_srcdir)/config/depcomp
+am__depfiles_maybe = depfiles
+CXXCOMPILE = $(CXX) $(DEFS) $(DEFAULT_INCLUDES) $(INCLUDES) \
+	$(AM_CPPFLAGS) $(CPPFLAGS) $(AM_CXXFLAGS) $(CXXFLAGS)
+CXXLD = $(CXX)
+CXXLINK = $(CXXLD) $(AM_CXXFLAGS) $(CXXFLAGS) $(AM_LDFLAGS) $(LDFLAGS) \
+	-o $@
+SOURCES = $(libtesseract_main_a_SOURCES) $(tesseract_SOURCES)
+DIST_SOURCES = $(libtesseract_main_a_SOURCES) $(tesseract_SOURCES)
+RECURSIVE_TARGETS = all-recursive check-recursive dvi-recursive \
+	html-recursive info-recursive install-data-recursive \
+	install-exec-recursive install-info-recursive \
+	install-recursive installcheck-recursive installdirs-recursive \
+	pdf-recursive ps-recursive uninstall-info-recursive \
+	uninstall-recursive
+ETAGS = etags
+CTAGS = ctags
+DIST_SUBDIRS = $(SUBDIRS)
+DISTFILES = $(DIST_COMMON) $(DIST_SOURCES) $(TEXINFOS) $(EXTRA_DIST)
+ACLOCAL = @ACLOCAL@
+AMDEP_FALSE = @AMDEP_FALSE@
+AMDEP_TRUE = @AMDEP_TRUE@
+AMTAR = @AMTAR@
+AUTOCONF = @AUTOCONF@
+AUTOHEADER = @AUTOHEADER@
+AUTOMAKE = @AUTOMAKE@
+AWK = @AWK@
+CC = @CC@
+CCDEPMODE = @CCDEPMODE@
+CFLAGS = @CFLAGS@
+CPPFLAGS = @CPPFLAGS@
+CXX = @CXX@
+CXXCPP = @CXXCPP@
+CXXDEPMODE = @CXXDEPMODE@
+CXXFLAGS = @CXXFLAGS@
+CXXRPOFLAGS = @CXXRPOFLAGS@
+CYGPATH_W = @CYGPATH_W@
+DEFS = @DEFS@
+DEPDIR = @DEPDIR@
+ECHO_C = @ECHO_C@
+ECHO_N = @ECHO_N@
+ECHO_T = @ECHO_T@
+EGREP = @EGREP@
+EXEEXT = @EXEEXT@
+GNUWIN32_DIR = @GNUWIN32_DIR@
+HAVE_GNUWIN32_FALSE = @HAVE_GNUWIN32_FALSE@
+HAVE_GNUWIN32_TRUE = @HAVE_GNUWIN32_TRUE@
+HAVE_LIBTIFF_FALSE = @HAVE_LIBTIFF_FALSE@
+HAVE_LIBTIFF_TRUE = @HAVE_LIBTIFF_TRUE@
+INSTALL_DATA = @INSTALL_DATA@
+INSTALL_PROGRAM = @INSTALL_PROGRAM@
+INSTALL_SCRIPT = @INSTALL_SCRIPT@
+INSTALL_STRIP_PROGRAM = @INSTALL_STRIP_PROGRAM@
+LDFLAGS = @LDFLAGS@
+LIBOBJS = @LIBOBJS@
+LIBS = @LIBS@
+LIBTIFF_CFLAGS = @LIBTIFF_CFLAGS@
+LIBTIFF_LIBS = @LIBTIFF_LIBS@
+LTLIBOBJS = @LTLIBOBJS@
+MAINT = @MAINT@
+MAINTAINER_MODE_FALSE = @MAINTAINER_MODE_FALSE@
+MAINTAINER_MODE_TRUE = @MAINTAINER_MODE_TRUE@
+MAKEINFO = @MAKEINFO@
+OBJEXT = @OBJEXT@
+OPTS = @OPTS@
+PACKAGE = @PACKAGE@
+PACKAGE_BUGREPORT = @PACKAGE_BUGREPORT@
+PACKAGE_DATE = @PACKAGE_DATE@
+PACKAGE_NAME = @PACKAGE_NAME@
+PACKAGE_STRING = @PACKAGE_STRING@
+PACKAGE_TARNAME = @PACKAGE_TARNAME@
+PACKAGE_VERSION = @PACKAGE_VERSION@
+PACKAGE_YEAR = @PACKAGE_YEAR@
+PATH_SEPARATOR = @PATH_SEPARATOR@
+RANLIB = @RANLIB@
+RPO_NO = @RPO_NO@
+RPO_YES = @RPO_YES@
+SET_MAKE = @SET_MAKE@
+SHELL = @SHELL@
+STRIP = @STRIP@
+USING_CL_FALSE = @USING_CL_FALSE@
+USING_CL_TRUE = @USING_CL_TRUE@
+VERSION = @VERSION@
+ac_ct_CC = @ac_ct_CC@
+ac_ct_CXX = @ac_ct_CXX@
+ac_ct_RANLIB = @ac_ct_RANLIB@
+ac_ct_STRIP = @ac_ct_STRIP@
+am__fastdepCC_FALSE = @am__fastdepCC_FALSE@
+am__fastdepCC_TRUE = @am__fastdepCC_TRUE@
+am__fastdepCXX_FALSE = @am__fastdepCXX_FALSE@
+am__fastdepCXX_TRUE = @am__fastdepCXX_TRUE@
+am__include = @am__include@
+am__leading_dot = @am__leading_dot@
+am__quote = @am__quote@
+am__tar = @am__tar@
+am__untar = @am__untar@
+bindir = @bindir@
+build = @build@
+build_alias = @build_alias@
+build_cpu = @build_cpu@
+build_os = @build_os@
+build_vendor = @build_vendor@
+datadir = @datadir@
+exec_prefix = @exec_prefix@
+host = @host@
+host_alias = @host_alias@
+host_cpu = @host_cpu@
+host_os = @host_os@
+host_vendor = @host_vendor@
+includedir = @includedir@
+infodir = @infodir@
+install_sh = @install_sh@
+libdir = @libdir@
+libexecdir = @libexecdir@
+localstatedir = @localstatedir@
+mandir = @mandir@
+mkdir_p = @mkdir_p@
+oldincludedir = @oldincludedir@
+prefix = @prefix@
+program_transform_name = @program_transform_name@
+sbindir = @sbindir@
+sharedstatedir = @sharedstatedir@
+sysconfdir = @sysconfdir@
+target_alias = @target_alias@
+SUBDIRS = 
+AM_CPPFLAGS = \
+    -I$(top_srcdir)/ccutil -I$(top_srcdir)/ccstruct \
+    -I$(top_srcdir)/image -I$(top_srcdir)/viewer \
+    -I$(top_srcdir)/ccops -I$(top_srcdir)/dict \
+    -I$(top_srcdir)/classify -I$(top_srcdir)/display \
+    -I$(top_srcdir)/wordrec -I$(top_srcdir)/cutil \
+    -I$(top_srcdir)/textord
+
+EXTRA_DIST = \
+    adaptions.h applybox.h baseapi.h blobcmp.h \
+    callnet.h charcut.h \
+    control.h docqual.h expandblob.h fixspace.h fixxht.h \
+    imgscale.h matmatch.h output.h paircmp.h reject.h scaleimg.h \
+    tessbox.h tessedit.h tesseractmain.h tessvars.h tfacep.h \
+    tessembedded.h tfacepp.h tstruct.h werdit.h
+
+noinst_LIBRARIES = libtesseract_main.a
+libtesseract_main_a_SOURCES = \
+    tessedit.cpp adaptions.cpp applybox.cpp \
+    baseapi.cpp blobcmp.cpp \
+    callnet.cpp charcut.cpp charsample.cpp control.cpp \
+    docqual.cpp expandblob.cpp fixspace.cpp fixxht.cpp \
+    imgscale.cpp matmatch.cpp output.cpp paircmp.cpp \
+    reject.cpp scaleimg.cpp tessbox.cpp tessvars.cpp \
+    tfacepp.cpp tstruct.cpp werdit.cpp
+
+tesseract_SOURCES = tesseractmain.cpp
+tesseract_LDADD = \
+    libtesseract_main.a \
+    ../display/libtesseract_display.a \
+    ../textord/libtesseract_textord.a \
+    ../wordrec/libtesseract_wordrec.a \
+    ../classify/libtesseract_classify.a \
+    ../dict/libtesseract_dict.a \
+    ../viewer/libtesseract_viewer.a \
+    ../image/libtesseract_image.a \
+    ../cutil/libtesseract_cutil.a \
+    ../ccstruct/libtesseract_ccstruct.a \
+    ../ccutil/libtesseract_ccutil.a
+
+all: all-recursive
+
+.SUFFIXES:
+.SUFFIXES: .cpp .o .obj
+$(srcdir)/Makefile.in: @MAINTAINER_MODE_TRUE@ $(srcdir)/Makefile.am  $(am__configure_deps)
+	@for dep in $?; do \
+	  case '$(am__configure_deps)' in \
+	    *$$dep*) \
+	      cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh \
+		&& exit 0; \
+	      exit 1;; \
+	  esac; \
+	done; \
+	echo ' cd $(top_srcdir) && $(AUTOMAKE) --gnu  ccmain/Makefile'; \
+	cd $(top_srcdir) && \
+	  $(AUTOMAKE) --gnu  ccmain/Makefile
+.PRECIOUS: Makefile
+Makefile: $(srcdir)/Makefile.in $(top_builddir)/config.status
+	@case '$?' in \
+	  *config.status*) \
+	    cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh;; \
+	  *) \
+	    echo ' cd $(top_builddir) && $(SHELL) ./config.status $(subdir)/$@ $(am__depfiles_maybe)'; \
+	    cd $(top_builddir) && $(SHELL) ./config.status $(subdir)/$@ $(am__depfiles_maybe);; \
+	esac;
+
+$(top_builddir)/config.status: $(top_srcdir)/configure $(CONFIG_STATUS_DEPENDENCIES)
+	cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh
+
+$(top_srcdir)/configure: @MAINTAINER_MODE_TRUE@ $(am__configure_deps)
+	cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh
+$(ACLOCAL_M4): @MAINTAINER_MODE_TRUE@ $(am__aclocal_m4_deps)
+	cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh
+
+clean-noinstLIBRARIES:
+	-test -z "$(noinst_LIBRARIES)" || rm -f $(noinst_LIBRARIES)
+libtesseract_main.a: $(libtesseract_main_a_OBJECTS) $(libtesseract_main_a_DEPENDENCIES) 
+	-rm -f libtesseract_main.a
+	$(libtesseract_main_a_AR) libtesseract_main.a $(libtesseract_main_a_OBJECTS) $(libtesseract_main_a_LIBADD)
+	$(RANLIB) libtesseract_main.a
+install-binPROGRAMS: $(bin_PROGRAMS)
+	@$(NORMAL_INSTALL)
+	test -z "$(bindir)" || $(mkdir_p) "$(DESTDIR)$(bindir)"
+	@list='$(bin_PROGRAMS)'; for p in $$list; do \
+	  p1=`echo $$p|sed 's/$(EXEEXT)$$//'`; \
+	  if test -f $$p \
+	  ; then \
+	    f=`echo "$$p1" | sed 's,^.*/,,;$(transform);s/$$/$(EXEEXT)/'`; \
+	   echo " $(INSTALL_PROGRAM_ENV) $(binPROGRAMS_INSTALL) '$$p' '$(DESTDIR)$(bindir)/$$f'"; \
+	   $(INSTALL_PROGRAM_ENV) $(binPROGRAMS_INSTALL) "$$p" "$(DESTDIR)$(bindir)/$$f" || exit 1; \
+	  else :; fi; \
+	done
+
+uninstall-binPROGRAMS:
+	@$(NORMAL_UNINSTALL)
+	@list='$(bin_PROGRAMS)'; for p in $$list; do \
+	  f=`echo "$$p" | sed 's,^.*/,,;s/$(EXEEXT)$$//;$(transform);s/$$/$(EXEEXT)/'`; \
+	  echo " rm -f '$(DESTDIR)$(bindir)/$$f'"; \
+	  rm -f "$(DESTDIR)$(bindir)/$$f"; \
+	done
+
+clean-binPROGRAMS:
+	-test -z "$(bin_PROGRAMS)" || rm -f $(bin_PROGRAMS)
+tesseract$(EXEEXT): $(tesseract_OBJECTS) $(tesseract_DEPENDENCIES) 
+	@rm -f tesseract$(EXEEXT)
+	$(CXXLINK) $(tesseract_LDFLAGS) $(tesseract_OBJECTS) $(tesseract_LDADD) $(LIBS)
+
+mostlyclean-compile:
+	-rm -f *.$(OBJEXT)
+
+distclean-compile:
+	-rm -f *.tab.c
+
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/adaptions.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/applybox.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/baseapi.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/blobcmp.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/callnet.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/charcut.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/charsample.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/control.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/docqual.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/expandblob.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/fixspace.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/fixxht.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/imgscale.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/matmatch.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/output.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/paircmp.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/reject.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/scaleimg.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/tessbox.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/tessedit.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/tesseractmain.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/tessvars.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/tfacepp.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/tstruct.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/werdit.Po@am__quote@
+
+.cpp.o:
+@am__fastdepCXX_TRUE@	if $(CXXCOMPILE) -MT $@ -MD -MP -MF "$(DEPDIR)/$*.Tpo" -c -o $@ $<; \
+@am__fastdepCXX_TRUE@	then mv -f "$(DEPDIR)/$*.Tpo" "$(DEPDIR)/$*.Po"; else rm -f "$(DEPDIR)/$*.Tpo"; exit 1; fi
+@AMDEP_TRUE@@am__fastdepCXX_FALSE@	source='$<' object='$@' libtool=no @AMDEPBACKSLASH@
+@AMDEP_TRUE@@am__fastdepCXX_FALSE@	DEPDIR=$(DEPDIR) $(CXXDEPMODE) $(depcomp) @AMDEPBACKSLASH@
+@am__fastdepCXX_FALSE@	$(CXXCOMPILE) -c -o $@ $<
+
+.cpp.obj:
+@am__fastdepCXX_TRUE@	if $(CXXCOMPILE) -MT $@ -MD -MP -MF "$(DEPDIR)/$*.Tpo" -c -o $@ `$(CYGPATH_W) '$<'`; \
+@am__fastdepCXX_TRUE@	then mv -f "$(DEPDIR)/$*.Tpo" "$(DEPDIR)/$*.Po"; else rm -f "$(DEPDIR)/$*.Tpo"; exit 1; fi
+@AMDEP_TRUE@@am__fastdepCXX_FALSE@	source='$<' object='$@' libtool=no @AMDEPBACKSLASH@
+@AMDEP_TRUE@@am__fastdepCXX_FALSE@	DEPDIR=$(DEPDIR) $(CXXDEPMODE) $(depcomp) @AMDEPBACKSLASH@
+@am__fastdepCXX_FALSE@	$(CXXCOMPILE) -c -o $@ `$(CYGPATH_W) '$<'`
+uninstall-info-am:
+
+# This directory's subdirectories are mostly independent; you can cd
+# into them and run `make' without going through this Makefile.
+# To change the values of `make' variables: instead of editing Makefiles,
+# (1) if the variable is set in `config.status', edit `config.status'
+#     (which will cause the Makefiles to be regenerated when you run `make');
+# (2) otherwise, pass the desired values on the `make' command line.
+$(RECURSIVE_TARGETS):
+	@failcom='exit 1'; \
+	for f in x $$MAKEFLAGS; do \
+	  case $$f in \
+	    *=* | --[!k]*);; \
+	    *k*) failcom='fail=yes';; \
+	  esac; \
+	done; \
+	dot_seen=no; \
+	target=`echo $@ | sed s/-recursive//`; \
+	list='$(SUBDIRS)'; for subdir in $$list; do \
+	  echo "Making $$target in $$subdir"; \
+	  if test "$$subdir" = "."; then \
+	    dot_seen=yes; \
+	    local_target="$$target-am"; \
+	  else \
+	    local_target="$$target"; \
+	  fi; \
+	  (cd $$subdir && $(MAKE) $(AM_MAKEFLAGS) $$local_target) \
+	  || eval $$failcom; \
+	done; \
+	if test "$$dot_seen" = "no"; then \
+	  $(MAKE) $(AM_MAKEFLAGS) "$$target-am" || exit 1; \
+	fi; test -z "$$fail"
+
+mostlyclean-recursive clean-recursive distclean-recursive \
+maintainer-clean-recursive:
+	@failcom='exit 1'; \
+	for f in x $$MAKEFLAGS; do \
+	  case $$f in \
+	    *=* | --[!k]*);; \
+	    *k*) failcom='fail=yes';; \
+	  esac; \
+	done; \
+	dot_seen=no; \
+	case "$@" in \
+	  distclean-* | maintainer-clean-*) list='$(DIST_SUBDIRS)' ;; \
+	  *) list='$(SUBDIRS)' ;; \
+	esac; \
+	rev=''; for subdir in $$list; do \
+	  if test "$$subdir" = "."; then :; else \
+	    rev="$$subdir $$rev"; \
+	  fi; \
+	done; \
+	rev="$$rev ."; \
+	target=`echo $@ | sed s/-recursive//`; \
+	for subdir in $$rev; do \
+	  echo "Making $$target in $$subdir"; \
+	  if test "$$subdir" = "."; then \
+	    local_target="$$target-am"; \
+	  else \
+	    local_target="$$target"; \
+	  fi; \
+	  (cd $$subdir && $(MAKE) $(AM_MAKEFLAGS) $$local_target) \
+	  || eval $$failcom; \
+	done && test -z "$$fail"
+tags-recursive:
+	list='$(SUBDIRS)'; for subdir in $$list; do \
+	  test "$$subdir" = . || (cd $$subdir && $(MAKE) $(AM_MAKEFLAGS) tags); \
+	done
+ctags-recursive:
+	list='$(SUBDIRS)'; for subdir in $$list; do \
+	  test "$$subdir" = . || (cd $$subdir && $(MAKE) $(AM_MAKEFLAGS) ctags); \
+	done
+
+ID: $(HEADERS) $(SOURCES) $(LISP) $(TAGS_FILES)
+	list='$(SOURCES) $(HEADERS) $(LISP) $(TAGS_FILES)'; \
+	unique=`for i in $$list; do \
+	    if test -f "$$i"; then echo $$i; else echo $(srcdir)/$$i; fi; \
+	  done | \
+	  $(AWK) '    { files[$$0] = 1; } \
+	       END { for (i in files) print i; }'`; \
+	mkid -fID $$unique
+tags: TAGS
+
+TAGS: tags-recursive $(HEADERS) $(SOURCES)  $(TAGS_DEPENDENCIES) \
+		$(TAGS_FILES) $(LISP)
+	tags=; \
+	here=`pwd`; \
+	if ($(ETAGS) --etags-include --version) >/dev/null 2>&1; then \
+	  include_option=--etags-include; \
+	  empty_fix=.; \
+	else \
+	  include_option=--include; \
+	  empty_fix=; \
+	fi; \
+	list='$(SUBDIRS)'; for subdir in $$list; do \
+	  if test "$$subdir" = .; then :; else \
+	    test ! -f $$subdir/TAGS || \
+	      tags="$$tags $$include_option=$$here/$$subdir/TAGS"; \
+	  fi; \
+	done; \
+	list='$(SOURCES) $(HEADERS)  $(LISP) $(TAGS_FILES)'; \
+	unique=`for i in $$list; do \
+	    if test -f "$$i"; then echo $$i; else echo $(srcdir)/$$i; fi; \
+	  done | \
+	  $(AWK) '    { files[$$0] = 1; } \
+	       END { for (i in files) print i; }'`; \
+	if test -z "$(ETAGS_ARGS)$$tags$$unique"; then :; else \
+	  test -n "$$unique" || unique=$$empty_fix; \
+	  $(ETAGS) $(ETAGSFLAGS) $(AM_ETAGSFLAGS) $(ETAGS_ARGS) \
+	    $$tags $$unique; \
+	fi
+ctags: CTAGS
+CTAGS: ctags-recursive $(HEADERS) $(SOURCES)  $(TAGS_DEPENDENCIES) \
+		$(TAGS_FILES) $(LISP)
+	tags=; \
+	here=`pwd`; \
+	list='$(SOURCES) $(HEADERS)  $(LISP) $(TAGS_FILES)'; \
+	unique=`for i in $$list; do \
+	    if test -f "$$i"; then echo $$i; else echo $(srcdir)/$$i; fi; \
+	  done | \
+	  $(AWK) '    { files[$$0] = 1; } \
+	       END { for (i in files) print i; }'`; \
+	test -z "$(CTAGS_ARGS)$$tags$$unique" \
+	  || $(CTAGS) $(CTAGSFLAGS) $(AM_CTAGSFLAGS) $(CTAGS_ARGS) \
+	     $$tags $$unique
+
+GTAGS:
+	here=`$(am__cd) $(top_builddir) && pwd` \
+	  && cd $(top_srcdir) \
+	  && gtags -i $(GTAGS_ARGS) $$here
+
+distclean-tags:
+	-rm -f TAGS ID GTAGS GRTAGS GSYMS GPATH tags
+
+distdir: $(DISTFILES)
+	@srcdirstrip=`echo "$(srcdir)" | sed 's|.|.|g'`; \
+	topsrcdirstrip=`echo "$(top_srcdir)" | sed 's|.|.|g'`; \
+	list='$(DISTFILES)'; for file in $$list; do \
+	  case $$file in \
+	    $(srcdir)/*) file=`echo "$$file" | sed "s|^$$srcdirstrip/||"`;; \
+	    $(top_srcdir)/*) file=`echo "$$file" | sed "s|^$$topsrcdirstrip/|$(top_builddir)/|"`;; \
+	  esac; \
+	  if test -f $$file || test -d $$file; then d=.; else d=$(srcdir); fi; \
+	  dir=`echo "$$file" | sed -e 's,/[^/]*$$,,'`; \
+	  if test "$$dir" != "$$file" && test "$$dir" != "."; then \
+	    dir="/$$dir"; \
+	    $(mkdir_p) "$(distdir)$$dir"; \
+	  else \
+	    dir=''; \
+	  fi; \
+	  if test -d $$d/$$file; then \
+	    if test -d $(srcdir)/$$file && test $$d != $(srcdir); then \
+	      cp -pR $(srcdir)/$$file $(distdir)$$dir || exit 1; \
+	    fi; \
+	    cp -pR $$d/$$file $(distdir)$$dir || exit 1; \
+	  else \
+	    test -f $(distdir)/$$file \
+	    || cp -p $$d/$$file $(distdir)/$$file \
+	    || exit 1; \
+	  fi; \
+	done
+	list='$(DIST_SUBDIRS)'; for subdir in $$list; do \
+	  if test "$$subdir" = .; then :; else \
+	    test -d "$(distdir)/$$subdir" \
+	    || $(mkdir_p) "$(distdir)/$$subdir" \
+	    || exit 1; \
+	    distdir=`$(am__cd) $(distdir) && pwd`; \
+	    top_distdir=`$(am__cd) $(top_distdir) && pwd`; \
+	    (cd $$subdir && \
+	      $(MAKE) $(AM_MAKEFLAGS) \
+	        top_distdir="$$top_distdir" \
+	        distdir="$$distdir/$$subdir" \
+	        distdir) \
+	      || exit 1; \
+	  fi; \
+	done
+check-am: all-am
+check: check-recursive
+all-am: Makefile $(LIBRARIES) $(PROGRAMS)
+installdirs: installdirs-recursive
+installdirs-am:
+	for dir in "$(DESTDIR)$(bindir)"; do \
+	  test -z "$$dir" || $(mkdir_p) "$$dir"; \
+	done
+install: install-recursive
+install-exec: install-exec-recursive
+install-data: install-data-recursive
+uninstall: uninstall-recursive
+
+install-am: all-am
+	@$(MAKE) $(AM_MAKEFLAGS) install-exec-am install-data-am
+
+installcheck: installcheck-recursive
+install-strip:
+	$(MAKE) $(AM_MAKEFLAGS) INSTALL_PROGRAM="$(INSTALL_STRIP_PROGRAM)" \
+	  install_sh_PROGRAM="$(INSTALL_STRIP_PROGRAM)" INSTALL_STRIP_FLAG=-s \
+	  `test -z '$(STRIP)' || \
+	    echo "INSTALL_PROGRAM_ENV=STRIPPROG='$(STRIP)'"` install
+mostlyclean-generic:
+
+clean-generic:
+
+distclean-generic:
+	-test -z "$(CONFIG_CLEAN_FILES)" || rm -f $(CONFIG_CLEAN_FILES)
+
+maintainer-clean-generic:
+	@echo "This command is intended for maintainers to use"
+	@echo "it deletes files that may require special tools to rebuild."
+clean: clean-recursive
+
+clean-am: clean-binPROGRAMS clean-generic clean-noinstLIBRARIES \
+	mostlyclean-am
+
+distclean: distclean-recursive
+	-rm -rf ./$(DEPDIR)
+	-rm -f Makefile
+distclean-am: clean-am distclean-compile distclean-generic \
+	distclean-tags
+
+dvi: dvi-recursive
+
+dvi-am:
+
+html: html-recursive
+
+info: info-recursive
+
+info-am:
+
+install-data-am:
+
+install-exec-am: install-binPROGRAMS
+
+install-info: install-info-recursive
+
+install-man:
+
+installcheck-am:
+
+maintainer-clean: maintainer-clean-recursive
+	-rm -rf ./$(DEPDIR)
+	-rm -f Makefile
+maintainer-clean-am: distclean-am maintainer-clean-generic
+
+mostlyclean: mostlyclean-recursive
+
+mostlyclean-am: mostlyclean-compile mostlyclean-generic
+
+pdf: pdf-recursive
+
+pdf-am:
+
+ps: ps-recursive
+
+ps-am:
+
+uninstall-am: uninstall-binPROGRAMS uninstall-info-am
+
+uninstall-info: uninstall-info-recursive
+
+.PHONY: $(RECURSIVE_TARGETS) CTAGS GTAGS all all-am check check-am \
+	clean clean-binPROGRAMS clean-generic clean-noinstLIBRARIES \
+	clean-recursive ctags ctags-recursive distclean \
+	distclean-compile distclean-generic distclean-recursive \
+	distclean-tags distdir dvi dvi-am html html-am info info-am \
+	install install-am install-binPROGRAMS install-data \
+	install-data-am install-exec install-exec-am install-info \
+	install-info-am install-man install-strip installcheck \
+	installcheck-am installdirs installdirs-am maintainer-clean \
+	maintainer-clean-generic maintainer-clean-recursive \
+	mostlyclean mostlyclean-compile mostlyclean-generic \
+	mostlyclean-recursive pdf pdf-am ps ps-am tags tags-recursive \
+	uninstall uninstall-am uninstall-binPROGRAMS uninstall-info-am
+
+# Tell versions [3.59,3.63) of GNU make to not export all variables.
+# Otherwise a system limit (for SysV at least) may be exceeded.
+.NOEXPORT:

ccmain/adaptions.h

@@ -0,0 +1,109 @@
+/**********************************************************************
+ * File:        adaptions.h  (Formerly adaptions.h)
+ * Description: Functions used to adapt to blobs already confidently
+ *					identified
+ * Author:		Chris Newton
+ * Created:		Thu Oct  7 10:17:28 BST 1993
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           ADAPTIONS_H
+#define           ADAPTIONS_H
+
+#include          "charsample.h"
+#include          "charcut.h"
+#include          "notdll.h"
+
+extern BOOL_VAR_H (tessedit_reject_ems, FALSE, "Reject all m's");
+extern BOOL_VAR_H (tessedit_reject_suspect_ems, FALSE, "Reject suspect m's");
+extern double_VAR_H (tessedit_cluster_t1, 0.20,
+"t1 threshold for clustering samples");
+extern double_VAR_H (tessedit_cluster_t2, 0.40,
+"t2 threshold for clustering samples");
+extern double_VAR_H (tessedit_cluster_t3, 0.12,
+"Extra threshold for clustering samples, only keep a new sample if best score greater than this value");
+extern double_VAR_H (tessedit_cluster_accept_fraction, 0.80,
+"Largest fraction of characters in cluster for it to be used for adaption");
+extern INT_VAR_H (tessedit_cluster_min_size, 3,
+"Smallest number of samples in a cluster for it to be used for adaption");
+extern BOOL_VAR_H (tessedit_cluster_debug, FALSE,
+"Generate and print debug information for adaption by clustering");
+extern BOOL_VAR_H (tessedit_use_best_sample, FALSE,
+"Use best sample from cluster when adapting");
+extern BOOL_VAR_H (tessedit_test_cluster_input, FALSE,
+"Set reject map to enable cluster input to be measured");
+extern BOOL_VAR_H (tessedit_matrix_match, TRUE, "Use matrix matcher");
+extern BOOL_VAR_H (tessedit_old_matrix_match, FALSE, "Use matrix matcher");
+extern BOOL_VAR_H (tessedit_mm_use_non_adaption_set, FALSE,
+"Don't try to adapt to characters on this list");
+extern STRING_VAR_H (tessedit_non_adaption_set, ",.;:'~@*",
+"Characters to be avoided when adapting");
+extern BOOL_VAR_H (tessedit_mm_adapt_using_prototypes, TRUE,
+"Use prototypes when adapting");
+extern BOOL_VAR_H (tessedit_mm_use_prototypes, TRUE,
+"Use prototypes as clusters are built");
+extern BOOL_VAR_H (tessedit_mm_use_rejmap, FALSE,
+"Adapt to characters using reject map");
+extern BOOL_VAR_H (tessedit_mm_all_rejects, FALSE,
+"Adapt to all characters using, matrix matcher");
+extern BOOL_VAR_H (tessedit_mm_only_match_same_char, FALSE,
+"Only match samples against clusters for the same character");
+extern BOOL_VAR_H (tessedit_process_rns, FALSE, "Handle m - rn ambigs");
+extern BOOL_VAR_H (tessedit_demo_adaption, FALSE,
+"Display cut images and matrix match for demo purposes");
+extern INT_VAR_H (tessedit_demo_word1, 62,
+"Word number of first word to display");
+extern INT_VAR_H (tessedit_demo_word2, 64,
+"Word number of second word to display");
+extern STRING_VAR_H (tessedit_demo_file, "academe",
+"Name of document containing demo words");
+BOOL8 word_adaptable(  //should we adapt?
+                     WERD_RES *word,
+                     UINT16 mode);
+void collect_ems_for_adaption(WERD_RES *word,
+                              CHAR_SAMPLES_LIST *char_clusters,
+                              CHAR_SAMPLE_LIST *chars_waiting);
+void collect_characters_for_adaption(WERD_RES *word,
+                                     CHAR_SAMPLES_LIST *char_clusters,
+                                     CHAR_SAMPLE_LIST *chars_waiting);
+void cluster_sample(CHAR_SAMPLE *sample,
+                    CHAR_SAMPLES_LIST *char_clusters,
+                    CHAR_SAMPLE_LIST *chars_waiting);
+void check_wait_list(CHAR_SAMPLE_LIST *chars_waiting,
+                     CHAR_SAMPLE *sample,
+                     CHAR_SAMPLES *best_cluster);
+void complete_clustering(CHAR_SAMPLES_LIST *char_clusters,
+                         CHAR_SAMPLE_LIST *chars_waiting);
+void adapt_to_good_ems(WERD_RES *word,
+                       CHAR_SAMPLES_LIST *char_clusters,
+                       CHAR_SAMPLE_LIST *chars_waiting);
+void adapt_to_good_samples(WERD_RES *word,
+                           CHAR_SAMPLES_LIST *char_clusters,
+                           CHAR_SAMPLE_LIST *chars_waiting);
+void print_em_stats(CHAR_SAMPLES_LIST *char_clusters,
+                    CHAR_SAMPLE_LIST *chars_waiting);
+                                 //lines of the image
+CHAR_SAMPLE *clip_sample(PIXROW *pixrow,
+                         IMAGELINE *imlines,
+                         BOX pix_box,  //box of imlines extent
+                         BOOL8 white_on_black,
+                         char c);
+void display_cluster_prototypes(CHAR_SAMPLES_LIST *char_clusters); 
+void reject_all_ems(WERD_RES *word); 
+void reject_all_fullstops(WERD_RES *word); 
+void reject_suspect_ems(WERD_RES *word); 
+void reject_suspect_fullstops(WERD_RES *word); 
+BOOL8 suspect_em(WERD_RES *word, INT16 index); 
+BOOL8 suspect_fullstop(WERD_RES *word, INT16 i); 
+#endif

ccmain/applybox.cpp

@@ -0,0 +1,859 @@
+/**********************************************************************
+ * File:        applybox.cpp  (Formerly applybox.c)
+ * Description: Re segment rows according to box file data
+ * Author:		Phil Cheatle
+ * Created:		Wed Nov 24 09:11:23 GMT 1993
+ *
+ * (C) Copyright 1993, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+/*
+define SECURE_NAMES for code versions which go to UNLV to stop tessedit
+including all the newdiff stuff (which contains lots of text indicating
+what measures we are interested in.
+*/
+/* #define SECURE_NAMES done in secnames.h when necessary*/
+
+#include "mfcpch.h"
+#include          "applybox.h"
+#include          <ctype.h>
+#include          <string.h>
+#ifdef __UNIX__
+#include          <assert.h>
+#include                    <errno.h>
+#endif
+#include          "mainblk.h"
+#include                   "genblob.h"
+#include                   "fixxht.h"
+#include          "control.h"
+#include          "tessbox.h"
+#include          "globals.h"
+#include          "secname.h"
+
+#define SECURE_NAMES
+#ifndef SECURE_NAMES
+#include          "wordstats.h"
+#endif
+
+#define EXTERN
+EXTERN BOOL_VAR (applybox_rebalance, TRUE, "Drop dead");
+EXTERN INT_VAR (applybox_debug, 0, "Debug level");
+EXTERN STRING_VAR (applybox_test_exclusions, "|",
+"Chars ignored for testing");
+EXTERN double_VAR (applybox_error_band, 0.15, "Err band as fract of xht");
+
+/*************************************************************************
+ * The code re-assigns outlines to form words each with ONE labelled blob.
+ * Noise is left in UNLABELLED words. The chars on the page are checked crudely
+ * for sensible position relative to baseline and xht. Failed boxes are
+ * compensated for by duplicating other believable instances of the character.
+ *
+ * The box file is assumed to contain box definitions, one per line, of the
+ * following format:
+ *   <Char> <left> <bottom> <right> <top> ... arbitrary trailing fields unused
+ *
+ * The approach taken is to search the WHOLE page for stuff overlapping each box.
+ *	- This is not too inefficient and is SAFE.
+ *    - We can detect overlapping blobs as we will be attempting to put a blob
+ *      from a LABELLED word into the current word.
+ *    - When all the boxes have been processed we can detect any stuff which is
+ *      being ignored - it is the unlabelled words left on the page.
+ *
+ * A box should only overlap one row.
+ *
+ * A warning is given if the box is on the same row as the previous box, but NOT
+ * on the same row as the previous blob.
+ *
+ * Any OUTLINE which overlaps the box is put into the new word.
+ *
+ * ascender chars must ascend above xht significantly
+ * xht chars must not rise above row xht significantly
+ * bl chars must not descend below baseline significantly
+ * descender chars must descend below baseline significantly
+ *
+ * ?? Certain chars are DROPPED - to limit the training data.
+ *
+ *************************************************************************/
+
+void apply_boxes(BLOCK_LIST *block_list    //real blocks
+                ) {
+  INT16 boxfile_lineno = 0;
+  INT16 boxfile_charno = 0;
+  BOX box;                       //boxfile box
+  char ch[2];                    //correct ch from boxfile
+  ROW *row;
+  ROW *prev_row = NULL;
+  INT16 prev_box_right = MAX_INT16;
+  INT16 block_id;
+  INT16 row_id;
+  INT16 box_count = 0;
+  INT16 box_failures = 0;
+  INT16 labels_ok;
+  INT16 rows_ok;
+  INT16 bad_blobs;
+  INT16 tgt_char_counts[128];    //No. of box samples
+  //      INT16                                   labelled_char_counts[128];      //No. of unique labelled samples
+  INT16 i;
+  INT16 rebalance_count = 0;
+  char min_char;
+  INT16 min_samples;
+  INT16 final_labelled_blob_count;
+
+  for (i = 0; i < 128; i++)
+    tgt_char_counts[i] = 0;
+
+  FILE* box_file;
+  STRING filename = imagefile;
+  filename += ".box";
+  if (!(box_file = fopen (filename.string(), "r"))) {
+    CANTOPENFILE.error ("read_next_box", EXIT,
+      "Cant open box file %s %d",
+      filename.string(), errno);
+  }
+
+  ch[1] = '\0';
+  clear_any_old_text(block_list);
+  while (read_next_box (box_file, &box, &ch[0])) {
+    box_count++;
+    tgt_char_counts[ch[0]]++;
+    row = find_row_of_box (block_list, box, block_id, row_id);
+    if (box.left () < prev_box_right) {
+      boxfile_lineno++;
+      boxfile_charno = 1;
+    }
+    else
+      boxfile_charno++;
+
+    if (row == NULL) {
+      box_failures++;
+      report_failed_box (boxfile_lineno, boxfile_charno, box, ch,
+        "FAILURE! box overlaps no blobs or blobs in multiple rows");
+    }
+    else {
+      if ((box.left () >= prev_box_right) && (row != prev_row))
+        report_failed_box (boxfile_lineno, boxfile_charno, box, ch,
+          "WARNING! false row break");
+      box_failures += resegment_box (row, box, ch, block_id, row_id,
+        boxfile_lineno, boxfile_charno);
+      prev_row = row;
+    }
+    prev_box_right = box.right ();
+  }
+  tidy_up(block_list,
+          labels_ok,
+          rows_ok,
+          bad_blobs,
+          tgt_char_counts,
+          rebalance_count,
+          min_char,
+          min_samples,
+          final_labelled_blob_count);
+  tprintf ("APPLY_BOXES:\n");
+  tprintf ("   Boxes read from boxfile:  %6d\n", box_count);
+  tprintf ("   Initially labelled blobs: %6d in %d rows\n",
+    labels_ok, rows_ok);
+  tprintf ("   Box failures detected:		%6d\n", box_failures);
+  tprintf ("   Duped blobs for rebalance:%6d\n", rebalance_count);
+  tprintf ("   \"%c\" has fewest samples:%6d\n", min_char, min_samples);
+  tprintf ("				Total unlabelled words:   %6d\n",
+    bad_blobs);
+  tprintf ("				Final labelled words:     %6d\n",
+    final_labelled_blob_count);
+}
+
+
+void clear_any_old_text(                        //remove correct text
+                        BLOCK_LIST *block_list  //real blocks
+                       ) {
+  BLOCK_IT block_it(block_list);
+  ROW_IT row_it;
+  WERD_IT word_it;
+
+  for (block_it.mark_cycle_pt ();
+  !block_it.cycled_list (); block_it.forward ()) {
+    row_it.set_to_list (block_it.data ()->row_list ());
+    for (row_it.mark_cycle_pt (); !row_it.cycled_list (); row_it.forward ()) {
+      word_it.set_to_list (row_it.data ()->word_list ());
+      for (word_it.mark_cycle_pt ();
+      !word_it.cycled_list (); word_it.forward ()) {
+        word_it.data ()->set_text ("");
+      }
+    }
+  }
+}
+
+
+BOOL8 read_next_box(FILE* box_file,  //
+                    BOX *box,
+                    char *ch) {
+  char buff[256];                //boxfile read buffer
+  char *buffptr = buff;
+  STRING box_filename;
+  static INT16 line = 0;
+  INT32 x_min;
+  INT32 y_min;
+  INT32 x_max;
+  INT32 y_max;
+  INT32 count = 0;
+
+  while (!feof (box_file)) {
+    fgets (buff, sizeof (buff) - 1, box_file);
+    line++;
+
+    /* Check for blank lines in box file */
+    for (buffptr = buff; isspace (*buffptr); buffptr++)
+      ;
+    if (*buffptr != '\0') {
+      count =
+        sscanf (buff,
+        "%c " INT32FORMAT " " INT32FORMAT " " INT32FORMAT " "
+        INT32FORMAT, ch, &x_min, &y_min, &x_max, &y_max);
+      if (count != 5) {
+        tprintf ("Box file format error on line %i ignored\n", line);
+      }
+      else {
+        *box = BOX (ICOORD (x_min, y_min), ICOORD (x_max, y_max));
+        return TRUE;             //read a box ok
+      }
+    }
+  }
+  return FALSE;                  //EOF
+}
+
+
+ROW *find_row_of_box(                         //
+                     BLOCK_LIST *block_list,  //real blocks
+                     BOX box,                 //from boxfile
+                     INT16 &block_id,
+                     INT16 &row_id_to_process) {
+  BLOCK_IT block_it(block_list);
+  BLOCK *block;
+  ROW_IT row_it;
+  ROW *row;
+  ROW *row_to_process = NULL;
+  INT16 row_id;
+  WERD_IT word_it;
+  WERD *word;
+  BOOL8 polyg;
+  PBLOB_IT blob_it;
+  PBLOB *blob;
+  OUTLINE_IT outline_it;
+  OUTLINE *outline;
+
+  /*
+    Find row to process - error if box REALLY overlaps more than one row. (I.e
+    it overlaps blobs in the row - not just overlaps the bounding box of the
+    whole row.)
+  */
+
+  block_id = 0;
+  for (block_it.mark_cycle_pt ();
+  !block_it.cycled_list (); block_it.forward ()) {
+    block_id++;
+    row_id = 0;
+    block = block_it.data ();
+    if (block->bounding_box ().overlap (box)) {
+      row_it.set_to_list (block->row_list ());
+      for (row_it.mark_cycle_pt ();
+      !row_it.cycled_list (); row_it.forward ()) {
+        row_id++;
+        row = row_it.data ();
+        if (row->bounding_box ().overlap (box)) {
+          word_it.set_to_list (row->word_list ());
+          for (word_it.mark_cycle_pt ();
+          !word_it.cycled_list (); word_it.forward ()) {
+            word = word_it.data ();
+            polyg = word->flag (W_POLYGON);
+            if (word->bounding_box ().overlap (box)) {
+              blob_it.set_to_list (word->gblob_list ());
+              for (blob_it.mark_cycle_pt ();
+              !blob_it.cycled_list (); blob_it.forward ()) {
+                blob = blob_it.data ();
+                if (gblob_bounding_box (blob, polyg).
+                overlap (box)) {
+                  outline_it.
+                    set_to_list (gblob_out_list
+                    (blob, polyg));
+                  for (outline_it.mark_cycle_pt ();
+                    !outline_it.cycled_list ();
+                  outline_it.forward ()) {
+                    outline = outline_it.data ();
+                    if (goutline_bounding_box
+                    (outline, polyg).major_overlap (box)) {
+                      if ((row_to_process == NULL) ||
+                      (row_to_process == row)) {
+                        row_to_process = row;
+                        row_id_to_process = row_id;
+                      }
+                      else
+                        /* RETURN ERROR Box overlaps blobs in more than one row  */
+                        return NULL;
+                    }
+                  }
+                }
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+  return row_to_process;
+}
+
+
+INT16 resegment_box(  //
+                    ROW *row,
+                    BOX box,
+                    char *ch,
+                    INT16 block_id,
+                    INT16 row_id,
+                    INT16 boxfile_lineno,
+                    INT16 boxfile_charno) {
+  WERD_IT word_it;
+  WERD *word;
+  WERD *new_word = NULL;
+  BOOL8 polyg = false;
+  PBLOB_IT blob_it;
+  PBLOB_IT new_blob_it;
+  PBLOB *blob;
+  PBLOB *new_blob;
+  OUTLINE_IT outline_it;
+  OUTLINE_LIST dummy;  // Just to initialize new_outline_it.
+  OUTLINE_IT new_outline_it = &dummy;
+  OUTLINE *outline;
+  BOX new_word_box;
+  float word_x_centre;
+  float baseline;
+  INT16 error_count = 0;         //number of chars lost
+
+  word_it.set_to_list (row->word_list ());
+  for (word_it.mark_cycle_pt (); !word_it.cycled_list (); word_it.forward ()) {
+    word = word_it.data ();
+    polyg = word->flag (W_POLYGON);
+    if (word->bounding_box ().overlap (box)) {
+      blob_it.set_to_list (word->gblob_list ());
+      for (blob_it.mark_cycle_pt ();
+      !blob_it.cycled_list (); blob_it.forward ()) {
+        blob = blob_it.data ();
+        if (gblob_bounding_box (blob, polyg).overlap (box)) {
+          outline_it.set_to_list (gblob_out_list (blob, polyg));
+          for (outline_it.mark_cycle_pt ();
+          !outline_it.cycled_list (); outline_it.forward ()) {
+            outline = outline_it.data ();
+            if (goutline_bounding_box (outline, polyg).
+            major_overlap (box)) {
+              if (strlen (word->text ()) > 0) {
+                if (error_count == 0) {
+                  error_count = 1;
+                  if (applybox_debug > 4)
+                    report_failed_box (boxfile_lineno,
+                      boxfile_charno,
+                      box, ch,
+                      "FAILURE! box overlaps blob in labelled word");
+                }
+                if (applybox_debug > 4)
+                  tprintf
+                    ("APPLY_BOXES: ALSO ignoring corrupted char blk:%d row:%d \"%s\"\n",
+                    block_id, row_id,
+                    word_it.data ()->text ());
+                word_it.data ()->set_text ("");
+                //UN label it
+                error_count++;
+              }
+
+              if (error_count == 0) {
+                if (new_word == NULL) {
+                                 /* Make a new word with a single blob */
+                  new_word = word->shallow_copy ();
+                  new_word->set_text (ch);
+                  if (polyg)
+                    new_blob = new PBLOB;
+                  else
+                    new_blob = (PBLOB *) new C_BLOB;
+                  new_blob_it.set_to_list (new_word->
+                    gblob_list ());
+                  new_blob_it.add_to_end (new_blob);
+                  new_outline_it.
+                    set_to_list (gblob_out_list
+                    (new_blob, polyg));
+                }
+                new_outline_it.add_to_end (outline_it.
+                  extract ());
+                //move blob
+              }
+            }
+          }
+                                 //no outlines in blob
+          if (outline_it.empty ())
+                                 //so delete blob
+            delete blob_it.extract ();
+        }
+      }
+      if (blob_it.empty ())      //no blobs in word
+                                 //so delete word
+          delete word_it.extract ();
+    }
+  }
+  if (error_count > 0)
+    return error_count;
+
+  if (new_word != NULL) {
+    gblob_sort_list (new_word->gblob_list (), polyg);
+    word_it.add_to_end (new_word);
+    new_word_box = new_word->bounding_box ();
+    word_x_centre = (new_word_box.left () + new_word_box.right ()) / 2.0f;
+    baseline = row->base_line (word_x_centre);
+
+    if (STRING (chs_caps_ht).contains (ch[0]) &&
+      (new_word_box.top () <
+    baseline + (1 + applybox_error_band) * row->x_height ())) {
+      report_failed_box (boxfile_lineno, boxfile_charno, box, ch,
+        "FAILURE! caps-ht char didn't ascend");
+      new_word->set_text ("");
+      return 1;
+    }
+    if (STRING (chs_odd_top).contains (ch[0]) &&
+      (new_word_box.top () <
+    baseline + (1 - applybox_error_band) * row->x_height ())) {
+      report_failed_box (boxfile_lineno, boxfile_charno, box, ch,
+        "FAILURE! Odd top char below xht");
+      new_word->set_text ("");
+      return 1;
+    }
+    if (STRING (chs_x_ht).contains (ch[0]) &&
+      ((new_word_box.top () >
+      baseline + (1 + applybox_error_band) * row->x_height ()) ||
+      (new_word_box.top () <
+    baseline + (1 - applybox_error_band) * row->x_height ()))) {
+      report_failed_box (boxfile_lineno, boxfile_charno, box, ch,
+        "FAILURE! x-ht char didn't have top near xht");
+      new_word->set_text ("");
+      return 1;
+    }
+    if (STRING (chs_non_ambig_bl).contains (ch[0]) &&
+      ((new_word_box.bottom () <
+      baseline - applybox_error_band * row->x_height ()) ||
+      (new_word_box.bottom () >
+    baseline + applybox_error_band * row->x_height ()))) {
+      report_failed_box (boxfile_lineno, boxfile_charno, box, ch,
+        "FAILURE! non ambig BL char didnt have bottom near baseline");
+      new_word->set_text ("");
+      return 1;
+    }
+    if (STRING (chs_odd_bot).contains (ch[0]) &&
+      (new_word_box.bottom () >
+    baseline + applybox_error_band * row->x_height ())) {
+      report_failed_box (boxfile_lineno, boxfile_charno, box, ch,
+        "FAILURE! Odd bottom char above baseline");
+      new_word->set_text ("");
+      return 1;
+    }
+    if (STRING (chs_desc).contains (ch[0]) &&
+      (new_word_box.bottom () >
+    baseline - applybox_error_band * row->x_height ())) {
+      report_failed_box (boxfile_lineno, boxfile_charno, box, ch,
+        "FAILURE! Descender doesn't descend");
+      new_word->set_text ("");
+      return 1;
+    }
+    return 0;
+  }
+  else {
+    report_failed_box (boxfile_lineno, boxfile_charno, box, ch,
+      "FAILURE! Couldn't find any blobs");
+    return 1;
+  }
+}
+
+
+/*************************************************************************
+ * tidy_up()
+ *   - report >1 block
+ *   - sort the words in each row.
+ *   - report any rows with no labelled words.
+ *   - report any remaining unlabelled words
+ *		- report total labelled words
+ *
+ *************************************************************************/
+void tidy_up(                         //
+             BLOCK_LIST *block_list,  //real blocks
+             INT16 &ok_char_count,
+             INT16 &ok_row_count,
+             INT16 &unlabelled_words,
+             INT16 *tgt_char_counts,
+             INT16 &rebalance_count,
+             char &min_char,
+             INT16 &min_samples,
+             INT16 &final_labelled_blob_count) {
+  BLOCK_IT block_it(block_list);
+  ROW_IT row_it;
+  ROW *row;
+  WERD_IT word_it;
+  WERD *word;
+  WERD *duplicate_word;
+  INT16 block_idx = 0;
+  INT16 row_idx;
+  INT16 all_row_idx = 0;
+  BOOL8 row_ok;
+  BOOL8 rebalance_needed = FALSE;
+                                 //No. of unique labelled samples
+  INT16 labelled_char_counts[128];
+  INT16 i;
+  char ch;
+  char prev_ch = '\0';
+  BOOL8 at_dupe_of_prev_word;
+  ROW *prev_row = NULL;
+  INT16 left;
+  INT16 prev_left = -1;
+
+  for (i = 0; i < 128; i++)
+    labelled_char_counts[i] = 0;
+
+  ok_char_count = 0;
+  ok_row_count = 0;
+  unlabelled_words = 0;
+  if ((applybox_debug > 4) && (block_it.length () != 1))
+
+    tprintf ("APPLY_BOXES: More than one block??\n");
+
+  for (block_it.mark_cycle_pt ();
+  !block_it.cycled_list (); block_it.forward ()) {
+    block_idx++;
+    row_idx = 0;
+    row_ok = FALSE;
+    row_it.set_to_list (block_it.data ()->row_list ());
+    for (row_it.mark_cycle_pt (); !row_it.cycled_list (); row_it.forward ()) {
+      row_idx++;
+      all_row_idx++;
+      row = row_it.data ();
+      word_it.set_to_list (row->word_list ());
+      word_it.sort (word_comparator);
+      for (word_it.mark_cycle_pt ();
+      !word_it.cycled_list (); word_it.forward ()) {
+        word = word_it.data ();
+        if (strlen (word->text ()) == 0) {
+          unlabelled_words++;
+          if (applybox_debug > 4) {
+            tprintf
+              ("APPLY_BOXES: Unlabelled word blk:%d row:%d allrows:%d\n",
+              block_idx, row_idx, all_row_idx);
+          }
+        }
+        else {
+          if (word->gblob_list ()->length () != 1)
+            tprintf
+              ("APPLY_BOXES: FATALITY - MULTIBLOB Labelled word blk:%d row:%d allrows:%d\n",
+              block_idx, row_idx, all_row_idx);
+
+          ok_char_count++;
+          labelled_char_counts[*word->text ()]++;
+          row_ok = TRUE;
+        }
+      }
+      if ((applybox_debug > 4) && (!row_ok)) {
+        tprintf
+          ("APPLY_BOXES: Row with no labelled words blk:%d row:%d allrows:%d\n",
+          block_idx, row_idx, all_row_idx);
+      }
+      else
+        ok_row_count++;
+    }
+  }
+
+  min_samples = 9999;
+  for (i = 0; i < 128; i++) {
+    if (tgt_char_counts[i] > labelled_char_counts[i]) {
+      if (labelled_char_counts[i] <= 1) {
+        tprintf
+          ("APPLY_BOXES: FATALITY - %d labelled samples of \"%c\" - target is %d\n",
+          labelled_char_counts[i], (char) i, tgt_char_counts[i]);
+      }
+      else {
+        rebalance_needed = TRUE;
+        if (applybox_debug > 0)
+          tprintf
+            ("APPLY_BOXES: REBALANCE REQD \"%c\" - target of %d from %d labelled samples\n",
+            (char) i, tgt_char_counts[i], labelled_char_counts[i]);
+      }
+    }
+    if ((min_samples > labelled_char_counts[i]) && (tgt_char_counts[i] > 0)) {
+      min_samples = labelled_char_counts[i];
+      min_char = (char) i;
+    }
+  }
+
+  while (applybox_rebalance && rebalance_needed) {
+    block_it.set_to_list (block_list);
+    for (block_it.mark_cycle_pt ();
+    !block_it.cycled_list (); block_it.forward ()) {
+      row_it.set_to_list (block_it.data ()->row_list ());
+      for (row_it.mark_cycle_pt ();
+      !row_it.cycled_list (); row_it.forward ()) {
+        row = row_it.data ();
+        word_it.set_to_list (row->word_list ());
+        for (word_it.mark_cycle_pt ();
+        !word_it.cycled_list (); word_it.forward ()) {
+          word = word_it.data ();
+          left = word->bounding_box ().left ();
+          ch = *word->text ();
+          at_dupe_of_prev_word = ((row == prev_row) &&
+            (left = prev_left) &&
+            (ch == prev_ch));
+          if ((ch != '\0') &&
+            (labelled_char_counts[ch] > 1) &&
+            (tgt_char_counts[ch] > labelled_char_counts[ch]) &&
+          (!at_dupe_of_prev_word)) {
+            /* Duplicate the word to rebalance the labelled samples */
+            if (applybox_debug > 9) {
+              tprintf ("Duping \"%c\" from ", ch);
+              word->bounding_box ().print ();
+            }
+            duplicate_word = new WERD;
+            *duplicate_word = *word;
+            word_it.add_after_then_move (duplicate_word);
+            rebalance_count++;
+            labelled_char_counts[ch]++;
+          }
+          prev_row = row;
+          prev_left = left;
+          prev_ch = ch;
+        }
+      }
+    }
+    rebalance_needed = FALSE;
+    for (i = 0; i < 128; i++) {
+      if ((tgt_char_counts[i] > labelled_char_counts[i]) &&
+      (labelled_char_counts[i] > 1)) {
+        rebalance_needed = TRUE;
+        break;
+      }
+    }
+  }
+
+  /* Now final check - count labelled blobs */
+  final_labelled_blob_count = 0;
+  block_it.set_to_list (block_list);
+  for (block_it.mark_cycle_pt ();
+  !block_it.cycled_list (); block_it.forward ()) {
+    row_it.set_to_list (block_it.data ()->row_list ());
+    for (row_it.mark_cycle_pt (); !row_it.cycled_list (); row_it.forward ()) {
+      row = row_it.data ();
+      word_it.set_to_list (row->word_list ());
+      word_it.sort (word_comparator);
+      for (word_it.mark_cycle_pt ();
+      !word_it.cycled_list (); word_it.forward ()) {
+        word = word_it.data ();
+        if ((strlen (word->text ()) == 1) &&
+          (word->gblob_list ()->length () == 1))
+          final_labelled_blob_count++;
+      }
+    }
+  }
+}
+
+
+void report_failed_box(INT16 boxfile_lineno,
+                       INT16 boxfile_charno,
+                       BOX box,
+                       char *box_ch,
+                       const char *err_msg) {
+  if (applybox_debug > 4)
+    tprintf ("APPLY_BOXES: boxfile %1d/%1d/%s ((%1d,%1d),(%1d,%1d)): %s\n",
+      boxfile_lineno,
+      boxfile_charno,
+      box_ch,
+      box.left (), box.bottom (), box.right (), box.top (), err_msg);
+}
+
+
+void apply_box_training(BLOCK_LIST *block_list) {
+  BLOCK_IT block_it(block_list);
+  ROW_IT row_it;
+  ROW *row;
+  WERD_IT word_it;
+  WERD *word;
+  WERD *bln_word;
+  WERD copy_outword;             // copy to denorm
+  PBLOB_IT blob_it;
+  DENORM denorm;
+  INT16 count = 0;
+  char ch[2];
+
+  ch[1] = '\0';
+
+  tprintf ("Generating training data\n");
+  for (block_it.mark_cycle_pt ();
+  !block_it.cycled_list (); block_it.forward ()) {
+    row_it.set_to_list (block_it.data ()->row_list ());
+    for (row_it.mark_cycle_pt (); !row_it.cycled_list (); row_it.forward ()) {
+      row = row_it.data ();
+      word_it.set_to_list (row->word_list ());
+      for (word_it.mark_cycle_pt ();
+      !word_it.cycled_list (); word_it.forward ()) {
+        word = word_it.data ();
+        if ((strlen (word->text ()) == 1) &&
+        (word->gblob_list ()->length () == 1)) {
+          /* Here is a word with a single char label and a single blob so train on it */
+          bln_word =
+            make_bln_copy (word, row, row->x_height (), &denorm);
+          blob_it.set_to_list (bln_word->blob_list ());
+          ch[0] = *word->text ();
+          tess_training_tester (blob_it.data (),
+                                 //single blob
+            &denorm, TRUE,       //correct
+            ch,                  //correct ASCII char
+            1,                   //ASCII length
+            NULL);
+          copy_outword = *(bln_word);
+          copy_outword.baseline_denormalise (&denorm);
+          blob_it.set_to_list (copy_outword.blob_list ());
+          ch[0] = *word->text ();
+          delete bln_word;
+          count++;
+        }
+      }
+    }
+  }
+  tprintf ("Generated training data for %d blobs\n", count);
+}
+
+
+void apply_box_testing(BLOCK_LIST *block_list) {
+  BLOCK_IT block_it(block_list);
+  ROW_IT row_it;
+  ROW *row;
+  INT16 row_count = 0;
+  WERD_IT word_it;
+  WERD *word;
+  WERD *bln_word;
+  INT16 word_count = 0;
+  PBLOB_IT blob_it;
+  DENORM denorm;
+  INT16 count = 0;
+  char ch[2];
+  WERD *outword;                 //bln best choice
+  //segmentation
+  WERD_CHOICE *best_choice;      //tess output
+  WERD_CHOICE *raw_choice;       //top choice permuter
+                                 //detailed results
+  BLOB_CHOICE_LIST_CLIST blob_choices;
+  INT16 char_count = 0;
+  INT16 correct_count = 0;
+  INT16 err_count = 0;
+  INT16 rej_count = 0;
+  #ifndef SECURE_NAMES
+  WERDSTATS wordstats;           //As from newdiff
+  #endif
+  char tess_rej_str[3];
+  char tess_long_str[3];
+
+  ch[1] = '\0';
+  strcpy (tess_rej_str, "|A");
+  strcpy (tess_long_str, "|B");
+
+  for (block_it.mark_cycle_pt ();
+  !block_it.cycled_list (); block_it.forward ()) {
+    row_it.set_to_list (block_it.data ()->row_list ());
+    for (row_it.mark_cycle_pt (); !row_it.cycled_list (); row_it.forward ()) {
+      row = row_it.data ();
+      row_count++;
+      word_count = 0;
+      word_it.set_to_list (row->word_list ());
+      for (word_it.mark_cycle_pt ();
+      !word_it.cycled_list (); word_it.forward ()) {
+        word = word_it.data ();
+        word_count++;
+        if ((strlen (word->text ()) == 1) &&
+          !STRING (applybox_test_exclusions).contains (*word->text ())
+        && (word->gblob_list ()->length () == 1)) {
+          /* Here is a word with a single char label and a single blob so test it */
+          bln_word =
+            make_bln_copy (word, row, row->x_height (), &denorm);
+          blob_it.set_to_list (bln_word->blob_list ());
+          ch[0] = *word->text ();
+          char_count++;
+          best_choice = tess_segment_pass1 (bln_word,
+            &denorm,
+            tess_default_matcher,
+            raw_choice,
+            &blob_choices, outword);
+
+          /*
+            Test for TESS screw up on word. Recog_word has already ensured that the
+            choice list, outword blob lists and best_choice string are the same
+            length. A TESS screw up is indicated by a blank filled or 0 length string.
+          */
+          if ((best_choice->string ().length () == 0) ||
+            (strspn (best_choice->string ().string (), " ") ==
+          best_choice->string ().length ())) {
+            rej_count++;
+            tprintf ("%d:%d: \"%s\" -> TESS FAILED\n",
+              row_count, word_count, ch);
+            #ifndef SECURE_NAMES
+            wordstats.word (tess_rej_str, 2, ch, 1);
+            #endif
+          }
+          else {
+            if ((best_choice->string ().length () !=
+              outword->blob_list ()->length ()) ||
+              (best_choice->string ().length () !=
+            blob_choices.length ())) {
+              tprintf
+                ("ASSERT FAIL String:\"%s\"; Strlen=%d; #Blobs=%d; #Choices=%d\n",
+                best_choice->string ().string (),
+                best_choice->string ().length (),
+                outword->blob_list ()->length (),
+                blob_choices.length ());
+            }
+            ASSERT_HOST (best_choice->string ().length () ==
+              outword->blob_list ()->length ());
+            ASSERT_HOST (best_choice->string ().length () ==
+              blob_choices.length ());
+            fix_quotes ((char *) best_choice->string ().string (),
+                                 //turn to double
+              outword, &blob_choices);
+            if (strcmp (best_choice->string ().string (), ch) != 0) {
+              err_count++;
+              tprintf ("%d:%d: \"%s\" -> \"%s\"\n",
+                row_count, word_count, ch,
+                best_choice->string ().string ());
+            }
+            else
+              correct_count++;
+            #ifndef SECURE_NAMES
+            if (best_choice->string ().length () > 2)
+              wordstats.word (tess_long_str, 2, ch, 1);
+            else
+              wordstats.word ((char *) best_choice->string ().
+                string (),
+                best_choice->string ().length (), ch,
+                1);
+            #endif
+          }
+          delete bln_word;
+          delete outword;
+          delete best_choice;
+          delete raw_choice;
+          blob_choices.deep_clear ();
+          count++;
+        }
+      }
+    }
+  }
+  #ifndef SECURE_NAMES
+  wordstats.print (1, 100.0);
+  wordstats.conf_matrix ();
+  tprintf ("Tested %d chars: %d correct; %d rejected by tess; %d errs\n",
+    char_count, correct_count, rej_count, err_count);
+  #endif
+}

ccmain/applybox.h

@@ -0,0 +1,71 @@
+/**********************************************************************
+ * File:        applybox.h  (Formerly applybox.h)
+ * Description: Re segment rows according to box file data
+ * Author:		Phil Cheatle
+ * Created:		Wed Nov 24 09:11:23 GMT 1993
+ *
+ * (C) Copyright 1993, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           APPLYBOX_H
+#define           APPLYBOX_H
+
+#include          "varable.h"
+#include          "ocrblock.h"
+#include          "ocrrow.h"
+#include          "notdll.h"
+
+extern BOOL_VAR_H (applybox_rebalance, TRUE, "Drop dead");
+extern INT_VAR_H (applybox_debug, 0, "Debug level");
+extern STRING_VAR_H (applybox_test_exclusions, "|",
+"Chars ignored for testing");
+extern double_VAR_H (applybox_error_band, 0.15, "Err band as fract of xht");
+void apply_boxes(BLOCK_LIST *block_list    //real blocks
+                );
+void clear_any_old_text(                        //remove correct text
+                        BLOCK_LIST *block_list  //real blocks
+                       );
+BOOL8 read_next_box(FILE* box_file,  //
+                    BOX *box,
+                    char *ch);
+ROW *find_row_of_box(                         //
+                     BLOCK_LIST *block_list,  //real blocks
+                     BOX box,                 //from boxfile
+                     INT16 &block_id,
+                     INT16 &row_id_to_process);
+INT16 resegment_box(  //
+                    ROW *row,
+                    BOX box,
+                    char *ch,
+                    INT16 block_id,
+                    INT16 row_id,
+                    INT16 boxfile_lineno,
+                    INT16 boxfile_charno);
+void tidy_up(                         //
+             BLOCK_LIST *block_list,  //real blocks
+             INT16 &ok_char_count,
+             INT16 &ok_row_count,
+             INT16 &unlabelled_words,
+             INT16 *tgt_char_counts,
+             INT16 &rebalance_count,
+             char &min_char,
+             INT16 &min_samples,
+             INT16 &final_labelled_blob_count);
+void report_failed_box(INT16 boxfile_lineno,
+                       INT16 boxfile_charno,
+                       BOX box,
+                       char *box_ch,
+                       const char *err_msg);
+void apply_box_training(BLOCK_LIST *block_list);
+void apply_box_testing(BLOCK_LIST *block_list);
+#endif

ccmain/baseapi.cpp

@@ -0,0 +1,395 @@
+/**********************************************************************
+ * File:        baseapi.cpp
+ * Description: Simple API for calling tesseract.
+ * Author:      Ray Smith
+ * Created:     Fri Oct 06 15:35:01 PDT 2006
+ *
+ * (C) Copyright 2006, Google Inc.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "baseapi.h"
+
+#include "tessedit.h"
+#include "pageres.h"
+#include "tessvars.h"
+#include "control.h"
+#include "applybox.h"
+#include "pgedit.h"
+#include "varabled.h"
+#include "adaptmatch.h"
+
+BOOL_VAR(tessedit_resegment_from_boxes, FALSE,
+         "Take segmentation and labeling from box file");
+BOOL_VAR(tessedit_train_from_boxes, FALSE,
+         "Generate training data from boxed chars");
+
+// Minimum sensible image size to be worth running tesseract.
+const int kMinRectSize = 10;
+
+// Start tesseract.
+// The datapath must be the name of the data directory or some other file
+// in which the data directory resides (for instance argv[0].)
+// The configfile is the name of a file in the tessconfigs directory
+// (eg batch) or NULL to run on defaults.
+// Outputbase may also be NULL, and is the basename of various output files.
+// If the output of any of these files is enabled, then a name nmust be given.
+// If numeric_mode is true, only possible digits and roman numbers are
+// returned. Returns 0 if successful. Crashes if not.
+// The argc and argv may be 0 and NULL respectively. They are used for
+// providing config files for debug/display purposes.
+// TODO(rays) get the facts straight. Is it OK to call
+// it more than once? Make it properly check for errors and return them.
+int TessBaseAPI::Init(const char* datapath, const char* outputbase,
+                      const char* configfile, bool numeric_mode,
+                      int argc, char* argv[]) {
+  int result = init_tesseract(datapath, outputbase, configfile, argc, argv);
+  bln_numericmode.set_value(numeric_mode);
+  return result;
+}
+
+// Recognize a rectangle from an image and return the result as a string.
+// May be called many times for a single Init.
+// Currently has no error checking.
+// Greyscale of 8 and color of 24 or 32 bits per pixel may be given.
+// Palette color images will not work properly and must be converted to
+// 24 bit.
+// Binary images of 1 bit per pixel may also be given but they must be
+// byte packed with the MSB of the first byte being the first pixel, and a
+// one pixel is WHITE. For binary images set bytes_per_pixel=0.
+// The recognized text is returned as a char* which (in future will be coded
+// as UTF8 and) must be freed with the delete [] operator.
+char* TessBaseAPI::TesseractRect(const UINT8* imagedata,
+                                 int bytes_per_pixel,
+                                 int bytes_per_line,
+                                 int left, int top,
+                                 int width, int height) {
+  if (width < kMinRectSize || height < kMinRectSize)
+    return NULL;  // Nothing worth doing.
+
+  // Copy/Threshold the image to the tesseract global page_image.
+  CopyImageToTesseract(imagedata, bytes_per_pixel, bytes_per_line,
+                       left, top, width, height);
+
+  return RecognizeToString();
+}
+
+// Call between pages or documents etc to free up memory and forget
+// adaptive data.
+void TessBaseAPI::ClearAdaptiveClassifier() {
+  ResetAdaptiveClassifier();
+}
+
+// Close down tesseract and free up memory.
+void TessBaseAPI::End() {
+  ResetAdaptiveClassifier();
+  end_tesseract();
+}
+
+// Dump the internal binary image to a PGM file.
+void TessBaseAPI::DumpPGM(const char* filename) {
+  IMAGELINE line;
+  line.init(page_image.get_xsize());
+  FILE *fp = fopen(filename, "w");
+  fprintf(fp, "P5 " INT32FORMAT " " INT32FORMAT " 255\n", page_image.get_xsize(),
+          page_image.get_ysize());
+  for (int j = page_image.get_ysize()-1; j >= 0 ; --j) {
+    page_image.get_line(0, j, page_image.get_xsize(), &line, 0);
+    for (int i = 0; i < page_image.get_xsize(); ++i) {
+      UINT8 b = line.pixels[i] ? 255 : 0;
+      fwrite(&b, 1, 1, fp);
+    }
+  }
+  fclose(fp);
+}
+
+// Copy the given image rectangle to Tesseract, with adaptive thresholding
+// if the image is not already binary.
+void TessBaseAPI::CopyImageToTesseract(const UINT8* imagedata,
+                                       int bytes_per_pixel,
+                                       int bytes_per_line,
+                                       int left, int top,
+                                       int width, int height) {
+  if (bytes_per_pixel > 0) {
+    // Threshold grey or color.
+    int* thresholds = new int[bytes_per_pixel];
+    int* hi_values = new int[bytes_per_pixel];
+
+    // Compute the thresholds.
+    OtsuThreshold(imagedata, bytes_per_pixel, bytes_per_line,
+                  left, top, left + width, top + height,
+                  thresholds, hi_values);
+
+    // Threshold the image to the tesseract global page_image.
+    ThresholdRect(imagedata, bytes_per_pixel, bytes_per_line,
+                  left, top, width, height,
+                  thresholds, hi_values);
+    delete [] thresholds;
+    delete [] hi_values;
+  } else {
+    CopyBinaryRect(imagedata, bytes_per_line, left, top, width, height);
+  }
+}
+
+// Compute the Otsu threshold(s) for the given image rectangle, making one
+// for each channel. Each channel is always one byte per pixel.
+// Returns an array of threshold values and an array of hi_values, such
+// that a pixel value >threshold[channel] is considered foreground if
+// hi_values[channel] is 0 or background if 1. A hi_value of -1 indicates
+// that there is no apparent foreground. At least one hi_value will not be -1.
+// thresholds and hi_values are assumed to be of bytes_per_pixel size.
+void TessBaseAPI::OtsuThreshold(const UINT8* imagedata,
+                                int bytes_per_pixel,
+                                int bytes_per_line,
+                                int left, int top, int right, int bottom,
+                                int* thresholds,
+                                int* hi_values) {
+  // Of all channels with no good hi_value, keep the best so we can always
+  // produce at least one answer.
+  int best_hi_value = 0;
+  int best_hi_index = 0;
+  bool any_good_hivalue = false;
+  double best_hi_dist = 0.0;
+
+  for (int ch = 0; ch < bytes_per_pixel; ++ch) {
+    thresholds[ch] = 0;
+    hi_values[ch] = -1;
+    // Compute the histogram of the image rectangle.
+    int histogram[256];
+    HistogramRect(imagedata + ch, bytes_per_pixel, bytes_per_line,
+                  left, top, right, bottom, histogram);
+    int H;
+    int best_omega_0;
+    int best_t = OtsuStats(histogram, &H, &best_omega_0);
+    // To be a convincing foreground we must have a small fraction of H
+    // or to be a convincing background we must have a large fraction of H.
+    // In between we assume this channel contains no thresholding information.
+    int hi_value = best_omega_0 < H * 0.5;
+    thresholds[ch] = best_t;
+    if (best_omega_0 > H * 0.75) {
+      any_good_hivalue = true;
+      hi_values[ch] = 0;
+    }
+    else if (best_omega_0 < H * 0.25) {
+      any_good_hivalue = true;
+      hi_values[ch] = 1;
+    }
+    else {
+      // In case all channels are like this, keep the best of the bad lot.
+      double hi_dist = hi_value ? (H - best_omega_0) : best_omega_0;
+      if (hi_dist > best_hi_dist) {
+        best_hi_dist = hi_dist;
+        best_hi_value = hi_value;
+        best_hi_index = ch;
+      }
+    }
+  }
+  if (!any_good_hivalue) {
+    // Use the best of the ones that were not good enough.
+    hi_values[best_hi_index] = best_hi_value;
+  }
+}
+
+// Compute the histogram for the given image rectangle, and the given
+// channel. (Channel pointed to by imagedata.) Each channel is always
+// one byte per pixel.
+// Bytes per pixel is used to skip channels not being
+// counted with this call in a multi-channel (pixel-major) image.
+// Histogram is always a 256 element array to count occurrences of
+// each pixel value.
+void TessBaseAPI::HistogramRect(const UINT8* imagedata,
+                                int bytes_per_pixel,
+                                int bytes_per_line,
+                                int left, int top, int right, int bottom,
+                                int* histogram) {
+  int width = right - left;
+  memset(histogram, 0, sizeof(*histogram) * 256);
+  const UINT8* pix = imagedata +
+                     top*bytes_per_line +
+                     left*bytes_per_pixel;
+  for (int y = top; y < bottom; ++y) {
+    for (int x = 0; x < width; ++x) {
+      ++histogram[pix[x * bytes_per_pixel]];
+    }
+    pix += bytes_per_line;
+  }
+}
+
+// Compute the Otsu threshold(s) for the given histogram.
+// Also returns H = total count in histogram, and
+// omega0 = count of histogram below threshold.
+int TessBaseAPI::OtsuStats(const int* histogram,
+                           int* H_out,
+                           int* omega0_out) {
+  int H = 0;
+  double mu_T = 0.0;
+  for (int i = 0; i < 256; ++i) {
+    H += histogram[i];
+    mu_T += i * histogram[i];
+  }
+
+  // Now maximize sig_sq_B over t.
+  // http://www.ctie.monash.edu.au/hargreave/Cornall_Terry_328.pdf
+  int best_t = -1;
+  int omega_0, omega_1;
+  int best_omega_0 = 0;
+  double best_sig_sq_B = 0.0;
+  double mu_0, mu_1, mu_t;
+  omega_0 = 0;
+  mu_t = 0.0;
+  for (int t = 0; t < 255; ++t) {
+    omega_0 += histogram[t];
+    mu_t += t * static_cast<double>(histogram[t]);
+    if (omega_0 == 0)
+      continue;
+    omega_1 = H - omega_0;
+    mu_0 = mu_t / omega_0;
+    mu_1 = (mu_T - mu_t) / omega_1;
+    double sig_sq_B = mu_1 - mu_0;
+    sig_sq_B *= sig_sq_B * omega_0 * omega_1;
+    if (best_t < 0 || sig_sq_B > best_sig_sq_B) {
+      best_sig_sq_B = sig_sq_B;
+      best_t = t;
+      best_omega_0 = omega_0;
+    }
+  }
+  if (H_out != NULL) *H_out = H;
+  if (omega0_out != NULL) *omega0_out = best_omega_0;
+  return best_t;
+}
+
+// Threshold the given grey or color image into the tesseract global
+// image ready for recognition. Requires thresholds and hi_value
+// produced by OtsuThreshold above.
+void TessBaseAPI::ThresholdRect(const UINT8* imagedata,
+                                int bytes_per_pixel,
+                                int bytes_per_line,
+                                int left, int top,
+                                int width, int height,
+                                const int* thresholds,
+                                const int* hi_values) {
+  IMAGELINE line;
+  page_image.create(width, height, 1);
+  line.init(width);
+  // For each line in the image, fill the IMAGELINE class and put it into the
+  // Tesseract global page_image. Note that Tesseract stores images with the
+  // bottom at y=0 and 0 is black, so we need 2 kinds of inversion.
+  const UINT8* data = imagedata + top*bytes_per_line + left*bytes_per_pixel;
+  for (int y = height - 1 ; y >= 0; --y) {
+    const UINT8* pix = data;
+    for (int x = 0; x < width; ++x, pix += bytes_per_pixel) {
+      line.pixels[x] = 1;
+      for (int ch = 0; ch < bytes_per_pixel; ++ch) {
+        if (hi_values[ch] >= 0 &&
+            (pix[ch] > thresholds[ch]) == (hi_values[ch] == 0)) {
+          line.pixels[x] = 0;
+          break;
+        }
+      }
+    }
+    page_image.put_line(0, y, width, &line, 0);
+    data += bytes_per_line;
+  }
+}
+
+// Cut out the requested rectangle of the binary image to the
+// tesseract global image ready for recognition.
+void TessBaseAPI::CopyBinaryRect(const UINT8* imagedata,
+                                 int bytes_per_line,
+                                 int left, int top,
+                                 int width, int height) {
+  // Copy binary image, cutting out the required rectangle.
+  IMAGE image;
+  image.capture(const_cast<UINT8*>(imagedata),
+                bytes_per_line*8, top + height, 1);
+  page_image.create(width, height, 1);
+  copy_sub_image(&image, left, top, width, height, &page_image, 0, 0, false);
+}
+
+// Low-level function to recognize the current global image to a string.
+char* TessBaseAPI::RecognizeToString() {
+  BLOCK_LIST    block_list;
+
+  FindLines(&block_list);
+
+  // Now run the main recognition.
+  PAGE_RES* page_res = Recognize(&block_list, NULL);
+
+  return TesseractToText(page_res);
+}
+
+// Find lines from the image making the BLOCK_LIST.
+void TessBaseAPI::FindLines(BLOCK_LIST* block_list) {
+  STRING input_file = "noname.tif";
+  // The following call creates a full-page block and then runs connected
+  // component analysis and text line creation.
+  pgeditor_read_file(input_file, block_list);
+}
+
+// Recognize the tesseract global image and return the result as Tesseract
+// internal structures.
+PAGE_RES* TessBaseAPI::Recognize(BLOCK_LIST* block_list, ETEXT_DESC* monitor) {
+  if (tessedit_resegment_from_boxes)
+    apply_boxes(block_list);
+  if (edit_variables)
+    start_variables_editor();
+
+  PAGE_RES* page_res = new PAGE_RES(block_list);
+  if (interactive_mode) {
+    pgeditor_main(block_list);                  //pgeditor user I/F
+  } else if (tessedit_train_from_boxes) {
+    apply_box_training(block_list);
+  } else {
+    // Now run the main recognition.
+    recog_all_words(page_res, monitor);
+  }
+  return page_res;
+}
+
+// Make a text string from the internal data structures.
+// The input page_res is deleted.
+char* TessBaseAPI::TesseractToText(PAGE_RES* page_res) {
+  if (page_res != NULL) {
+    int total_length = 2;
+    PAGE_RES_IT   page_res_it(page_res);
+    // Iterate over the data structures to extract the recognition result.
+    for (page_res_it.restart_page(); page_res_it.word () != NULL;
+         page_res_it.forward()) {
+      WERD_RES *word = page_res_it.word();
+      WERD_CHOICE* choice = word->best_choice;
+      if (choice != NULL) {
+        total_length += choice->string().length() + 1;
+      }
+    }
+    char* result = new char[total_length];
+    char* ptr = result;
+    for (page_res_it.restart_page(); page_res_it.word () != NULL;
+         page_res_it.forward()) {
+      WERD_RES *word = page_res_it.word();
+      WERD_CHOICE* choice = word->best_choice;
+      if (choice != NULL) {
+        strcpy(ptr, choice->string().string());
+        ptr += strlen(ptr);
+        if (word->word->flag(W_EOL))
+          *ptr++ = '\n';
+        else
+          *ptr++ = ' ';
+      }
+    }
+    *ptr++ = '\n';
+    *ptr = '\0';
+    delete page_res;
+    return result;
+  }
+  return NULL;
+}
+

ccmain/baseapi.h

@@ -0,0 +1,154 @@
+///////////////////////////////////////////////////////////////////////
+// File:        baseapi.h
+// Description: Simple API for calling tesseract.
+// Author:      Ray Smith
+// Created:     Fri Oct 06 15:35:01 PDT 2006
+//
+// (C) Copyright 2006, Google Inc.
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+// http://www.apache.org/licenses/LICENSE-2.0
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+//
+///////////////////////////////////////////////////////////////////////
+
+#ifndef THIRD_PARTY_TESSERACT_CCMAIN_BASEAPI_H__
+#define THIRD_PARTY_TESSERACT_CCMAIN_BASEAPI_H__
+
+#include <string>
+
+#include "host.h"
+#include "ocrclass.h"
+
+class PAGE_RES;
+class BLOCK_LIST;
+
+// Base class for all tesseract APIs.
+// Specific classes can add ability to work on different inputs or produce
+// different outputs.
+
+class TessBaseAPI {
+ public:
+  // Start tesseract.
+  // The datapath must be the name of the data directory or some other file
+  // in which the data directory resides (for instance argv[0].)
+  // The configfile is the name of a file in the tessconfigs directory
+  // (eg batch) or NULL to run on defaults.
+  // Outputbase may also be NULL, and is the basename of various output files.
+  // If the output of any of these files is enabled, then a name must be given.
+  // If numeric_mode is true, only possible digits and roman numbers are
+  // returned. Returns 0 if successful. Crashes if not.
+  // The argc and argv may be 0 and NULL respectively. They are used for
+  // providing config files for debug/display purposes.
+  // TODO(rays) get the facts straight. Is it OK to call
+  // it more than once? Make it properly check for errors and return them.
+  static int Init(const char* datapath, const char* outputbase,
+                  const char* configfile, bool numeric_mode,
+                  int argc, char* argv[]);
+
+  // Recognize a rectangle from an image and return the result as a string.
+  // May be called many times for a single Init.
+  // Currently has no error checking.
+  // Greyscale of 8 and color of 24 or 32 bits per pixel may be given.
+  // Palette color images will not work properly and must be converted to
+  // 24 bit.
+  // Binary images of 1 bit per pixel may also be given but they must be
+  // byte packed with the MSB of the first byte being the first pixel, and a
+  // 1 represents WHITE. For binary images set bytes_per_pixel=0.
+  // The recognized text is returned as a char* which (in future will be coded
+  // as UTF8 and) must be freed with the delete [] operator.
+  static char* TesseractRect(const UINT8* imagedata,
+                             int bytes_per_pixel,
+                             int bytes_per_line,
+                             int left, int top, int width, int height);
+
+  // Call between pages or documents etc to free up memory and forget
+  // adaptive data.
+  static void ClearAdaptiveClassifier();
+
+  // Close down tesseract and free up memory.
+  static void End();
+
+  // Dump the internal binary image to a PGM file.
+  static void DumpPGM(const char* filename);
+
+ protected:
+  // Copy the given image rectangle to Tesseract, with adaptive thresholding
+  // if the image is not already binary.
+  static void CopyImageToTesseract(const UINT8* imagedata,
+                                   int bytes_per_pixel,
+                                   int bytes_per_line,
+                                   int left, int top, int width, int height);
+
+  // Compute the Otsu threshold(s) for the given image rectangle, making one
+  // for each channel. Each channel is always one byte per pixel.
+  // Returns an array of threshold values and an array of hi_values, such
+  // that a pixel value >threshold[channel] is considered foreground if
+  // hi_values[channel] is 0 or background if 1. A hi_value of -1 indicates
+  // that there is no apparent foreground. At least one hi_value will not be -1.
+  // thresholds and hi_values are assumed to be of bytes_per_pixel size.
+  static void OtsuThreshold(const UINT8* imagedata,
+                           int bytes_per_pixel,
+                           int bytes_per_line,
+                           int left, int top, int right, int bottom,
+                           int* thresholds,
+                           int* hi_values);
+
+  // Compute the histogram for the given image rectangle, and the given
+  // channel. (Channel pointed to by imagedata.) Each channel is always
+  // one byte per pixel.
+  // Bytes per pixel is used to skip channels not being
+  // counted with this call in a multi-channel (pixel-major) image.
+  // Histogram is always a 256 element array to count occurrences of
+  // each pixel value.
+  static void HistogramRect(const UINT8* imagedata,
+                            int bytes_per_pixel,
+                            int bytes_per_line,
+                            int left, int top, int right, int bottom,
+                            int* histogram);
+
+  // Compute the Otsu threshold(s) for the given histogram.
+  // Also returns H = total count in histogram, and
+  // omega0 = count of histogram below threshold.
+  static int OtsuStats(const int* histogram,
+                       int* H_out,
+                       int* omega0_out);
+
+  // Threshold the given grey or color image into the tesseract global
+  // image ready for recognition. Requires thresholds and hi_value
+  // produced by OtsuThreshold above.
+  static void ThresholdRect(const UINT8* imagedata,
+                            int bytes_per_pixel,
+                            int bytes_per_line,
+                            int left, int top,
+                            int width, int height,
+                            const int* thresholds,
+                            const int* hi_values);
+
+  // Cut out the requested rectangle of the binary image to the
+  // tesseract global image ready for recognition.
+  static void CopyBinaryRect(const UINT8* imagedata,
+                             int bytes_per_line,
+                             int left, int top,
+                             int width, int height);
+
+  // Low-level function to recognize the current global image to a string.
+  static char* RecognizeToString();
+
+  // Find lines from the image making the BLOCK_LIST.
+  static void FindLines(BLOCK_LIST* block_list);
+
+  // Recognize the tesseract global image and return the result as Tesseract
+  // internal structures.
+  static PAGE_RES* Recognize(BLOCK_LIST* block_list, ETEXT_DESC* monitor);
+
+  // Convert (and free) the internal data structures into a text string.
+  static char* TesseractToText(PAGE_RES* page_res);
+};
+
+#endif  // THIRD_PARTY_TESSERACT_CCMAIN_BASEAPI_H__

ccmain/blobcmp.cpp

@@ -0,0 +1,76 @@
+/**********************************************************************
+ * File:        blobcmp.c  (Formerly blobcmp.c)
+ * Description: Code to compare blobs using the adaptive matcher.
+ * Author:		Ray Smith
+ * Created:		Wed Apr 21 09:28:51 BST 1993
+ *
+ * (C) Copyright 1993, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include "fxdefs.h"
+#include "ocrfeatures.h"
+#include "intmatcher.h"
+#include "intproto.h"
+#include "adaptive.h"
+#include "adaptmatch.h"
+#include "const.h"
+#include "tessvars.h"
+
+#define CMP_CLASS       'x'
+
+/**********************************************************************
+ * compare_tess_blobs
+ *
+ * Match 2 blobs using the adaptive classifier.
+ **********************************************************************/
+float compare_tess_blobs(TBLOB *blob1,
+                         TEXTROW *row1,
+                         TBLOB *blob2,
+                         TEXTROW *row2) {
+  int fcount;                    /*number of features */
+  ADAPT_TEMPLATES ad_templates;
+  LINE_STATS line_stats1, line_stats2;
+  INT_FEATURE_ARRAY int_features;
+  FEATURE_SET float_features;
+  INT_RESULT_STRUCT int_result;  /*output */
+
+  BIT_VECTOR AllProtosOn = NewBitVector (MAX_NUM_PROTOS);
+  BIT_VECTOR AllConfigsOn = NewBitVector (MAX_NUM_CONFIGS);
+  set_all_bits (AllProtosOn, WordsInVectorOfSize (MAX_NUM_PROTOS));
+  set_all_bits (AllConfigsOn, WordsInVectorOfSize (MAX_NUM_CONFIGS));
+
+  EnterClassifyMode;
+  ad_templates = NewAdaptedTemplates ();
+  GetLineStatsFromRow(row1, &line_stats1);
+                                 /*copy baseline stuff */
+  GetLineStatsFromRow(row2, &line_stats2);
+  MakeNewAdaptedClass(blob1, &line_stats1, CMP_CLASS, ad_templates);
+  fcount = GetAdaptiveFeatures (blob2, &line_stats2,
+    int_features, &float_features);
+  if (fcount > 0) {
+    SetBaseLineMatch();
+    IntegerMatcher (ClassForClassId (ad_templates->Templates, CMP_CLASS),
+      AllProtosOn, AllConfigsOn, fcount, fcount,
+      int_features, 0, 0, &int_result, testedit_match_debug);
+    FreeFeatureSet(float_features);
+    if (int_result.Rating < 0)
+      int_result.Rating = MAX_FLOAT32;
+  }
+
+  free_adapted_templates(ad_templates);
+  FreeBitVector(AllConfigsOn);
+  FreeBitVector(AllProtosOn);
+
+  return fcount > 0 ? int_result.Rating * fcount : MAX_FLOAT32;
+}

ccmain/blobcmp.h

@@ -0,0 +1,29 @@
+/**********************************************************************
+ * File:			blobcmp.c
+ * Description: Code to compare blobs using the adaptive matcher.
+ * Author:		Ray Smith
+ * Created:		Wed Apr 21 09:28:51 BST 1993
+ *
+ * (C) Copyright 1993, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           BLOBCMP_H
+#define           BLOBCMP_H
+
+#include          "tstruct.h"
+
+float compare_tess_blobs(TBLOB *blob1,
+                         TEXTROW *row1,
+                         TBLOB *blob2,
+                         TEXTROW *row2);
+#endif

ccmain/callnet.cpp

@@ -0,0 +1,93 @@
+/**********************************************************************
+ * File:        callnet.cpp  (Formerly callnet.c)
+ * Description: Interface to Neural Net matcher
+ * Author:      Phil Cheatle
+ * Created:     Wed Nov 18 10:35:00 GMT 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include "errcode.h"
+//#include "nmatch.h"
+#include "globals.h"
+
+#define OUTPUT_NODES 94
+
+const ERRCODE NETINIT = "NN init error";
+
+//extern "C"
+//{
+//extern char*                          demodir;                                        /* where program lives */
+
+void init_net() {  /* Initialise net */
+#ifdef ASPIRIN_INCLUDED
+  char wts_filename[256];
+
+  if (nmatch_init_network () != 0) {
+    NETINIT.error ("Init_net", EXIT, "Errcode %s", nmatch_error_string ());
+  }
+  strcpy(wts_filename, demodir);
+  strcat (wts_filename, "tessdata/netwts");
+
+  if (nmatch_load_network (wts_filename) != 0) {
+    NETINIT.error ("Init_net", EXIT, "Weights failed, Errcode %s",
+      nmatch_error_string ());
+  }
+#endif
+}
+
+
+void callnet(  /* Apply image to net */
+             float *input_vector,
+             char *top,
+             float *top_score,
+             char *next,
+             float *next_score) {
+#ifdef ASPIRIN_INCLUDED
+  float *output_vector;
+  int i;
+  int max_out_i = 0;
+  int next_max_out_i = 0;
+  float max_out = -9;
+  float next_max_out = -9;
+  
+  nmatch_set_input(input_vector);
+  nmatch_propagate_forward();
+  output_vector = nmatch_get_output ();
+  
+  /* Now find top two choices */
+
+  for (i = 0; i < OUTPUT_NODES; i++) {
+    if (output_vector[i] > max_out) {
+      next_max_out = max_out;
+      max_out = output_vector[i];
+      next_max_out_i = max_out_i;
+      max_out_i = i;
+    }
+    else {
+      if (output_vector[i] > next_max_out) {
+        next_max_out = output_vector[i];
+        next_max_out_i = i;
+      }
+    }
+  }
+  *top = max_out_i + '!';
+  *next = next_max_out_i + '!';
+  *top_score = max_out;
+  *next_score = next_max_out;
+#endif
+}
+
+
+//};

ccmain/callnet.h

@@ -0,0 +1,32 @@
+/**********************************************************************
+ * File:        callnet.h  (Formerly callnet.h)
+ * Description: Interface to Neural Net matcher
+ * Author:      Phil Cheatle
+ * Created:     Wed Nov 18 10:35:00 GMT 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           CALLNET_H
+#define           CALLNET_H
+
+//      extern "C" {
+void init_net();  /* Initialise net */
+void callnet(  /* Apply image to net */
+             float *input_vector,
+             char *top,
+             float *top_score,
+             char *next,
+             float *next_score);
+//      };
+#endif

ccmain/charcut.cpp

@@ -0,0 +1,710 @@
+/**********************************************************************
+ * File:        charcut.cpp  (Formerly charclip.c)
+ * Description: Code for character clipping
+ * Author:      Phil Cheatle
+ * Created:     Wed Nov 11 08:35:15 GMT 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          "charcut.h"
+#include          "imgs.h"
+#include          "showim.h"
+#include          "evnts.h"
+#include          "notdll.h"
+
+#define LARGEST(a,b) ( (a) > (b) ? (a) : (b) )
+#define SMALLEST(a,b) ( (a) > (b) ? (b) : (a) )
+#define BUG_OFFSET 1
+#define EXTERN
+
+EXTERN INT_VAR (pix_word_margin, 3, "How far outside word BB to grow");
+
+extern IMAGE page_image;
+
+ELISTIZE (PIXROW)
+/*************************************************************************
+ * PIXROW::PIXROW()
+ *
+ * Constructor for a specified size PIXROW from a blob
+ *************************************************************************/
+PIXROW::PIXROW(INT16 pos, INT16 count, PBLOB *blob) { 
+  OUTLINE_LIST *outline_list;
+  OUTLINE_IT outline_it;
+  POLYPT_LIST *pts_list;
+  POLYPT_IT pts_it;
+  INT16 i;
+  FCOORD pt;
+  FCOORD vec;
+  float y_coord;
+  INT16 x_coord;
+
+  row_offset = pos;
+  row_count = count;
+  min = (INT16 *) alloc_mem (count * sizeof (INT16));
+  max = (INT16 *) alloc_mem (count * sizeof (INT16));
+  outline_list = blob->out_list ();
+  outline_it.set_to_list (outline_list);
+
+  for (i = 0; i < count; i++) {
+    min[i] = MAX_INT16 - 1;
+    max[i] = -MAX_INT16 + 1;
+    y_coord = row_offset + i + 0.5;
+    for (outline_it.mark_cycle_pt ();
+    !outline_it.cycled_list (); outline_it.forward ()) {
+      pts_list = outline_it.data ()->polypts ();
+      pts_it.set_to_list (pts_list);
+      for (pts_it.mark_cycle_pt ();
+      !pts_it.cycled_list (); pts_it.forward ()) {
+        pt = pts_it.data ()->pos;
+        vec = pts_it.data ()->vec;
+        if ((vec.y () != 0) &&
+          (((pt.y () <= y_coord) && (pt.y () + vec.y () >= y_coord))
+          || ((pt.y () >= y_coord)
+        && (pt.y () + vec.y () <= y_coord)))) {
+          /* The segment crosses y_coord so find x-point and check for min/max. */
+          x_coord = (INT16) floor ((y_coord -
+            pt.y ()) * vec.x () / vec.y () +
+            pt.x () + 0.5);
+          if (x_coord < min[i])
+            min[i] = x_coord;
+          x_coord--;             //to get pix to left of line
+          if (x_coord > max[i])
+            max[i] = x_coord;
+        }
+      }
+    }
+  }
+}
+
+
+/*************************************************************************
+ * PIXROW::plot()
+ *
+ * Draw the PIXROW
+ *************************************************************************/
+
+#ifndef GRAPHICS_DISABLED
+void PIXROW::plot(WINDOW fd  //where to paint
+                 ) const {
+  INT16 i;
+  INT16 y_coord;
+
+  for (i = 0; i < row_count; i++) {
+    y_coord = row_offset + i;
+    if (min[i] <= max[i]) {
+      rectangle (fd, min[i], y_coord, max[i] + 1, y_coord + 1);
+    }
+  }
+}
+#endif
+
+/*************************************************************************
+ * PIXROW::bounding_box()
+ *
+ * Generate bounding box for blob image
+ *************************************************************************/
+
+bool PIXROW::bad_box(  //return true if box exceeds image
+                     int xsize,
+                     int ysize) const {
+  BOX bbox = bounding_box ();
+  if (bbox.left () < 0 || bbox.right () > xsize
+  || bbox.top () > ysize || bbox.bottom () < 0) {
+    tprintf("Box (%d,%d)->(%d,%d) bad compared to %d,%d\n",
+            bbox.left(),bbox.bottom(), bbox.right(), bbox.top(),
+            xsize, ysize);
+    return true;
+  }
+  return false;
+}
+
+
+/*************************************************************************
+ * PIXROW::bounding_box()
+ *
+ * Generate bounding box for blob image
+ *************************************************************************/
+
+BOX PIXROW::bounding_box() const { 
+  INT16 i;
+  INT16 y_coord;
+  INT16 min_x = MAX_INT16 - 1;
+  INT16 min_y = MAX_INT16 - 1;
+  INT16 max_x = -MAX_INT16 + 1;
+  INT16 max_y = -MAX_INT16 + 1;
+
+  for (i = 0; i < row_count; i++) {
+    y_coord = row_offset + i;
+    if (min[i] <= max[i]) {
+      if (y_coord < min_y)
+        min_y = y_coord;
+      if (y_coord + 1 > max_y)
+        max_y = y_coord + 1;
+      if (min[i] < min_x)
+        min_x = min[i];
+      if (max[i] + 1 > max_x)
+        max_x = max[i] + 1;
+    }
+  }
+  if (min_x > max_x || min_y > max_y)
+    return BOX ();
+  else
+    return BOX (ICOORD (min_x, min_y), ICOORD (max_x, max_y));
+}
+
+
+/*************************************************************************
+ * PIXROW::contract()
+ *
+ * Reduce the mins and maxs so that they end on black pixels
+ *************************************************************************/
+
+void PIXROW::contract(                         //image array
+                      IMAGELINE *imlines,
+                      INT16 x_offset,          //of pixels[0]
+                      INT16 foreground_colour  //0 or 1
+                     ) {
+  INT16 i;
+  UINT8 *line_pixels;
+
+  for (i = 0; i < row_count; i++) {
+    if (min[i] > max[i])
+      continue;
+
+    line_pixels = imlines[i].pixels;
+    while (line_pixels[min[i] - x_offset] != foreground_colour) {
+      if (min[i] == max[i]) {
+        min[i] = MAX_INT16 - 1;
+        max[i] = -MAX_INT16 + 1;
+        goto nextline;
+      }
+      else
+        min[i]++;
+    }
+    while (line_pixels[max[i] - x_offset] != foreground_colour) {
+      if (min[i] == max[i]) {
+        min[i] = MAX_INT16 - 1;
+        max[i] = -MAX_INT16 + 1;
+        goto nextline;
+      }
+      else
+        max[i]--;
+    }
+    nextline:;
+    //goto label!
+  }
+}
+
+
+/*************************************************************************
+ * PIXROW::extend()
+ *
+ * 1 pixel extension in each direction to cover extra black area
+ *************************************************************************/
+
+BOOL8 PIXROW::extend(               //image array
+                     IMAGELINE *imlines,
+                     BOX &imbox,
+                     PIXROW *prev,  //for prev blob
+                     PIXROW *next,  //for next blob
+                     INT16 foreground_colour) {
+  INT16 i;
+  INT16 x_offset = imbox.left ();
+  INT16 limit;
+  INT16 left_limit;
+  INT16 right_limit;
+  UINT8 *pixels = NULL;
+  UINT8 *pixels_below = NULL;    //row below current
+  UINT8 *pixels_above = NULL;    //row above current
+  BOOL8 changed = FALSE;
+
+  pixels_above = imlines[0].pixels;
+  for (i = 0; i < row_count; i++) {
+    pixels_below = pixels;
+    pixels = pixels_above;
+    if (i < (row_count - 1))
+      pixels_above = imlines[i + 1].pixels;
+    else
+      pixels_above = NULL;
+
+    /* Extend Left by one pixel*/
+    if (prev == NULL || prev->max[i] < prev->min[i])
+      limit = imbox.left ();
+    else
+      limit = prev->max[i] + 1;
+    if ((min[i] <= max[i]) &&
+      (min[i] > limit) &&
+    (pixels[min[i] - 1 - x_offset] == foreground_colour)) {
+      min[i]--;
+      changed = TRUE;
+    }
+
+    /* Extend Right by one pixel*/
+    if (next == NULL || next->min[i] > next->max[i])
+      limit = imbox.right () - 1;//-1 to index inside pix
+    else
+      limit = next->min[i] - 1;
+    if ((min[i] <= max[i]) &&
+      (max[i] < limit) &&
+    (pixels[max[i] + 1 - x_offset] == foreground_colour)) {
+      max[i]++;
+      changed = TRUE;
+    }
+
+    /* Extend down by one row */
+    if (pixels_below != NULL) {
+      if (min[i] < min[i - 1]) { //row goes left of row below
+        if (prev == NULL || prev->max[i - 1] < prev->min[i - 1])
+          left_limit = min[i];
+        else
+          left_limit = LARGEST (min[i], prev->max[i - 1] + 1);
+      }
+      else
+        left_limit = min[i - 1];
+
+      if (max[i] > max[i - 1]) { //row goes right of row below
+        if (next == NULL || next->min[i - 1] > next->max[i - 1])
+          right_limit = max[i];
+        else
+          right_limit = SMALLEST (max[i], next->min[i - 1] - 1);
+      }
+      else
+        right_limit = max[i - 1];
+
+      while ((left_limit <= right_limit) &&
+        (pixels_below[left_limit - x_offset] != foreground_colour))
+        left_limit++;            //find black extremity
+
+      if ((left_limit <= right_limit) && (left_limit < min[i - 1])) {
+        min[i - 1] = left_limit; //widen left if poss
+        changed = TRUE;
+      }
+
+      while ((left_limit <= right_limit) &&
+        (pixels_below[right_limit - x_offset] != foreground_colour))
+        right_limit--;           //find black extremity
+
+      if ((left_limit <= right_limit) && (right_limit > max[i - 1])) {
+        max[i - 1] = right_limit;//widen right if poss
+        changed = TRUE;
+      }
+    }
+
+    /* Extend up by one row */
+    if (pixels_above != NULL) {
+      if (min[i] < min[i + 1]) { //row goes left of row above
+        if (prev == NULL || prev->min[i + 1] > prev->max[i + 1])
+          left_limit = min[i];
+        else
+          left_limit = LARGEST (min[i], prev->max[i + 1] + 1);
+      }
+      else
+        left_limit = min[i + 1];
+
+      if (max[i] > max[i + 1]) { //row goes right of row above
+        if (next == NULL || next->min[i + 1] > next->max[i + 1])
+          right_limit = max[i];
+        else
+          right_limit = SMALLEST (max[i], next->min[i + 1] - 1);
+      }
+      else
+        right_limit = max[i + 1];
+
+      while ((left_limit <= right_limit) &&
+        (pixels_above[left_limit - x_offset] != foreground_colour))
+        left_limit++;            //find black extremity
+
+      if ((left_limit <= right_limit) && (left_limit < min[i + 1])) {
+        min[i + 1] = left_limit; //widen left if poss
+        changed = TRUE;
+      }
+
+      while ((left_limit <= right_limit) &&
+        (pixels_above[right_limit - x_offset] != foreground_colour))
+        right_limit--;           //find black extremity
+
+      if ((left_limit <= right_limit) && (right_limit > max[i + 1])) {
+        max[i + 1] = right_limit;//widen right if poss
+        changed = TRUE;
+      }
+    }
+  }
+  return changed;
+}
+
+
+/*************************************************************************
+ * PIXROW::char_clip_image()
+ * Cut out a sub image for a character
+ *************************************************************************/
+
+void PIXROW::char_clip_image(                     //box of imlines extnt
+                             IMAGELINE *imlines,
+                             BOX &im_box,
+                             ROW *row,            //row containing word
+                             IMAGE &clip_image,   //unscaled sq subimage
+                             float &baseline_pos  //baseline ht in image
+                            ) {
+  INT16 clip_image_xsize;        //sub image x size
+  INT16 clip_image_ysize;        //sub image y size
+  INT16 x_shift;                 //from pixrow to subim
+  INT16 y_shift;                 //from pixrow to subim
+  BOX char_pix_box;              //bbox of char pixels
+  INT16 y_dest;
+  INT16 x_min;
+  INT16 x_max;
+  INT16 x_min_dest;
+  INT16 x_max_dest;
+  INT16 x_width;
+  INT16 y;
+
+  clip_image_xsize = clip_image.get_xsize ();
+  clip_image_ysize = clip_image.get_ysize ();
+
+  char_pix_box = bounding_box ();
+  /*
+    The y shift is calculated by first finding the coord of the bottom of the
+    image relative to the image lines. Then reducing this so by the amount
+    relative to the clip image size, necessary to vertically position the
+    character.
+  */
+  y_shift = char_pix_box.bottom () - row_offset -
+    (INT16) floor ((clip_image_ysize - char_pix_box.height () + 0.5) / 2);
+
+  /*
+    The x_shift is the shift to be applied to the page coord in the pixrow to
+    generate a centred char in the clip image.  Thus the left hand edge of the
+    char is shifted to the margin width of the centred character.
+  */
+  x_shift = char_pix_box.left () -
+    (INT16) floor ((clip_image_xsize - char_pix_box.width () + 0.5) / 2);
+
+  for (y = 0; y < row_count; y++) {
+    /*
+      Check that there is something in this row of the source that will fit in the
+      sub image.  If there is, reduce x range if necessary, then copy it
+    */
+    y_dest = y - y_shift;
+    if ((min[y] <= max[y]) && (y_dest >= 0) && (y_dest < clip_image_ysize)) {
+      x_min = min[y];
+      x_min_dest = x_min - x_shift;
+      if (x_min_dest < 0) {
+        x_min = x_min - x_min_dest;
+        x_min_dest = 0;
+      }
+      x_max = max[y];
+      x_max_dest = x_max - x_shift;
+      if (x_max_dest > clip_image_xsize - 1) {
+        x_max = x_max - (x_max_dest - (clip_image_xsize - 1));
+        x_max_dest = clip_image_xsize - 1;
+      }
+      x_width = x_max - x_min + 1;
+      if (x_width > 0) {
+        x_min -= im_box.left ();
+                                 //offset pixel ptr
+        imlines[y].pixels += x_min;
+        clip_image.put_line (x_min_dest, y_dest, x_width, imlines + y,
+          0);
+        imlines[y].init ();      //reset pixel ptr
+      }
+    }
+  }
+  /*
+    Baseline position relative to clip image: First find the baseline relative
+    to the page origin at the x coord of the centre of the character. Then
+    make this relative to the character bottom. Finally shift by the margin
+    between the bottom of the character and the bottom of the clip image.
+  */
+  if (row == NULL)
+    baseline_pos = 0;            //Not needed
+  else
+    baseline_pos = row->base_line ((char_pix_box.left () +
+      char_pix_box.right ()) / 2.0)
+      - char_pix_box.bottom ()
+      + ((clip_image_ysize - char_pix_box.height ()) / 2);
+}
+
+
+/*************************************************************************
+ * char_clip_word()
+ *
+ * Generate a PIXROW_LIST with one element for each blob in the word, together
+ * with the image lines for the whole word.
+ *************************************************************************/
+
+void char_clip_word(                            //
+                    WERD *word,                 //word to be processed
+                    IMAGE &bin_image,           //whole image
+                    PIXROW_LIST *&pixrow_list,  //pixrows built
+                    IMAGELINE *&imlines,        //lines cut from image
+                    BOX &pix_box                //box defining imlines
+                   ) {
+  BOX word_box = word->bounding_box ();
+  PBLOB_LIST *blob_list;
+  PBLOB_IT blob_it;
+  PIXROW_IT pixrow_it;
+  INT16 pix_offset;              //Y pos of pixrow[0]
+  INT16 row_height;              //No of pix rows
+  INT16 imlines_x_offset;
+  PIXROW *prev;
+  PIXROW *next;
+  PIXROW *current;
+  BOOL8 changed;                 //still improving
+  BOOL8 just_changed;            //still improving
+  INT16 iteration_count = 0;
+  INT16 foreground_colour;
+
+  if (word->flag (W_INVERSE))
+    foreground_colour = 1;
+  else
+    foreground_colour = 0;
+
+  /* Define region for max pixrow expansion */
+  pix_box = word_box;
+  pix_box.move_bottom_edge (-pix_word_margin);
+  pix_box.move_top_edge (pix_word_margin);
+  pix_box.move_left_edge (-pix_word_margin);
+  pix_box.move_right_edge (pix_word_margin);
+  pix_box -= BOX (ICOORD (0, 0 + BUG_OFFSET),
+    ICOORD (bin_image.get_xsize (),
+    bin_image.get_ysize () - BUG_OFFSET));
+
+  /* Generate pixrows list */
+
+  pix_offset = pix_box.bottom ();
+  row_height = pix_box.height ();
+  blob_list = word->blob_list ();
+  blob_it.set_to_list (blob_list);
+
+  pixrow_list = new PIXROW_LIST;
+  pixrow_it.set_to_list (pixrow_list);
+
+  for (blob_it.mark_cycle_pt (); !blob_it.cycled_list (); blob_it.forward ()) {
+    PIXROW *row = new PIXROW (pix_offset, row_height, blob_it.data ());
+    ASSERT_HOST (!row->
+      bad_box (bin_image.get_xsize (), bin_image.get_ysize ()));
+    pixrow_it.add_after_then_move (row);
+  }
+
+  imlines = generate_imlines (bin_image, pix_box);
+
+  /* Contract pixrows - shrink min and max back to black pixels */
+
+  imlines_x_offset = pix_box.left ();
+
+  pixrow_it.move_to_first ();
+  for (pixrow_it.mark_cycle_pt ();
+  !pixrow_it.cycled_list (); pixrow_it.forward ()) {
+    ASSERT_HOST (!pixrow_it.data ()->
+      bad_box (bin_image.get_xsize (), bin_image.get_ysize ()));
+    pixrow_it.data ()->contract (imlines, imlines_x_offset,
+      foreground_colour);
+    ASSERT_HOST (!pixrow_it.data ()->
+      bad_box (bin_image.get_xsize (), bin_image.get_ysize ()));
+  }
+
+  /* Expand pixrows iteratively 1 pixel at a time */
+  do {
+    changed = FALSE;
+    pixrow_it.move_to_first ();
+    prev = NULL;
+    current = NULL;
+    next = pixrow_it.data ();
+    for (pixrow_it.mark_cycle_pt ();
+    !pixrow_it.cycled_list (); pixrow_it.forward ()) {
+      prev = current;
+      current = next;
+      if (pixrow_it.at_last ())
+        next = NULL;
+      else
+        next = pixrow_it.data_relative (1);
+      just_changed = current->extend (imlines, pix_box, prev, next,
+        foreground_colour);
+      ASSERT_HOST (!current->
+        bad_box (bin_image.get_xsize (),
+        bin_image.get_ysize ()));
+      changed = changed || just_changed;
+    }
+    iteration_count++;
+  }
+  while (changed);
+}
+
+
+/*************************************************************************
+ * generate_imlines()
+ * Get an array of IMAGELINES  holding a portion of an image
+ *************************************************************************/
+
+IMAGELINE *generate_imlines(                   //get some imagelines
+                            IMAGE &bin_image,  //from here
+                            BOX &pix_box) {
+  IMAGELINE *imlines;            //array of lines
+  int i;
+
+  imlines = new IMAGELINE[pix_box.height ()];
+  for (i = 0; i < pix_box.height (); i++) {
+    imlines[i].init (pix_box.width ());
+                                 //coord to start at
+    bin_image.fast_get_line (pix_box.left (),
+      pix_box.bottom () + i + BUG_OFFSET,
+    //line to get
+      pix_box.width (),          //width to get
+      imlines + i);              //dest imline
+  }
+  return imlines;
+}
+
+
+/*************************************************************************
+ * display_clip_image()
+ * All the boring user interface bits to let you see what's going on
+ *************************************************************************/
+
+#ifndef GRAPHICS_DISABLED
+WINDOW display_clip_image(WERD *word,                //word to be processed
+                          IMAGE &bin_image,          //whole image
+                          PIXROW_LIST *pixrow_list,  //pixrows built
+                          BOX &pix_box               //box of subimage
+                         ) {
+  WINDOW clip_window;            //window for debug
+  BOX word_box = word->bounding_box ();
+  int border = word_box.height () / 2;
+  BOX display_box = word_box;
+
+  display_box.move_bottom_edge (-border);
+  display_box.move_top_edge (border);
+  display_box.move_left_edge (-border);
+  display_box.move_right_edge (border);
+  display_box -= BOX (ICOORD (0, 0 - BUG_OFFSET),
+    ICOORD (bin_image.get_xsize (),
+    bin_image.get_ysize () - BUG_OFFSET));
+
+  pgeditor_msg ("Creating Clip window...");
+  clip_window =
+    create_window ("Clipped Blobs",
+    SCROLLINGWIN,
+    editor_word_xpos, editor_word_ypos,
+    3 * (word_box.width () + 2 * border),
+    3 * (word_box.height () + 2 * border),
+  //window width,height
+                                 // xmin, xmax
+    display_box.left (), display_box.right (),
+    display_box.bottom () - BUG_OFFSET,
+    display_box.top () - BUG_OFFSET,
+  // ymin, ymax
+    TRUE, FALSE, FALSE, TRUE);   // down event & key only
+  pgeditor_msg ("Creating Clip window...Done");
+
+  clear_view_surface(clip_window); 
+  show_sub_image (&bin_image,
+    display_box.left (),
+    display_box.bottom (),
+    display_box.width (),
+    display_box.height (),
+    clip_window,
+    display_box.left (), display_box.bottom () - BUG_OFFSET);
+
+  word->plot (clip_window, RED);
+  word_box.plot (clip_window, INT_HOLLOW, TRUE, BLUE, BLUE);
+  pix_box.plot (clip_window, INT_HOLLOW, TRUE, BLUE, BLUE);
+  plot_pixrows(pixrow_list, clip_window); 
+  overlap_picture_ops(TRUE); 
+  return clip_window;
+}
+
+
+/*************************************************************************
+ * display_images()
+ * Show a pair of clip and scaled character images and wait for key before
+ * continuing.
+ *************************************************************************/
+
+void display_images(IMAGE &clip_image, IMAGE &scaled_image) { 
+  WINDOW clip_im_window;         //window for debug
+  WINDOW scale_im_window;        //window for debug
+  INT16 i;
+  GRAPHICS_EVENT event;          //                                                      c;
+
+                                 // xmin xmax ymin ymax
+  clip_im_window = create_window ("Clipped Blob", SCROLLINGWIN, editor_word_xpos - 20, editor_word_ypos - 100, 5 * clip_image.get_xsize (), 5 * clip_image.get_ysize (), 0, clip_image.get_xsize (), 0, clip_image.get_ysize (),
+    TRUE, FALSE, FALSE, TRUE);   // down event & key only
+
+  clear_view_surface(clip_im_window); 
+  show_sub_image (&clip_image,
+    0, 0,
+    clip_image.get_xsize (), clip_image.get_ysize (),
+    clip_im_window, 0, 0);
+
+  line_color_index(clip_im_window, RED); 
+  for (i = 1; i < clip_image.get_xsize (); i++) {
+    move2d (clip_im_window, i, 0);
+    draw2d (clip_im_window, i, clip_image.get_xsize ());
+  }
+  for (i = 1; i < clip_image.get_ysize (); i++) {
+    move2d (clip_im_window, 0, i);
+    draw2d (clip_im_window, clip_image.get_xsize (), i);
+  }
+
+                                 // xmin xmax ymin ymax
+  scale_im_window = create_window ("Scaled Blob", SCROLLINGWIN, editor_word_xpos + 300, editor_word_ypos - 100, 5 * scaled_image.get_xsize (), 5 * scaled_image.get_ysize (), 0, scaled_image.get_xsize (), 0, scaled_image.get_ysize (),
+    TRUE, FALSE, FALSE, TRUE);   // down event & key only
+
+  clear_view_surface(scale_im_window); 
+  show_sub_image (&scaled_image,
+    0, 0,
+    scaled_image.get_xsize (), scaled_image.get_ysize (),
+    scale_im_window, 0, 0);
+
+  line_color_index(scale_im_window, RED); 
+  for (i = 1; i < scaled_image.get_xsize (); i++) {
+    move2d (scale_im_window, i, 0);
+    draw2d (scale_im_window, i, scaled_image.get_xsize ());
+  }
+  for (i = 1; i < scaled_image.get_ysize (); i++) {
+    move2d (scale_im_window, 0, i);
+    draw2d (scale_im_window, scaled_image.get_xsize (), i);
+  }
+
+  overlap_picture_ops(TRUE); 
+  await_event(scale_im_window, TRUE, ANY_EVENT, &event); 
+  destroy_window(clip_im_window); 
+  destroy_window(scale_im_window); 
+}
+
+
+/*************************************************************************
+ * plot_pixrows()
+ * Display a list of pixrows
+ *************************************************************************/
+
+void plot_pixrows(  //plot for all blobs
+                  PIXROW_LIST *pixrow_list,
+                  WINDOW win) {
+  PIXROW_IT pixrow_it(pixrow_list); 
+  INT16 colour = RED;
+
+  for (pixrow_it.mark_cycle_pt ();
+  !pixrow_it.cycled_list (); pixrow_it.forward ()) {
+    if (colour > RED + 7)
+      colour = RED;
+
+    perimeter_color_index (win, (COLOUR) colour);
+    interior_style(win, INT_HOLLOW, TRUE); 
+    pixrow_it.data ()->plot (win);
+    colour++;
+  }
+}
+#endif

ccmain/charcut.h

@@ -0,0 +1,119 @@
+/**********************************************************************
+ * File:        charcut.h  (Formerly charclip.h)
+ * Description: Code for character clipping
+ * Author:      Phil Cheatle
+ * Created:     Wed Nov 11 08:35:15 GMT 1992
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           CHARCUT_H
+#define           CHARCUT_H
+
+#include          "pgedit.h"
+#include          "notdll.h"
+#include          "notdll.h"
+
+/*************************************************************************
+ * CLASS PIXROW
+ *
+ * This class describes the pixels occupied by a blob. It uses two arrays, (min
+ * and max), each with one element per row, to identify the min and max x
+ * coordinates of the black pixels in the character on that row of the image.
+ * The number of rows used to describe the blob is held in row_count - note that
+ * some rows may be unoccupied - signified by max < min. The page coordinate of
+ * the row defined by min[0] and max[0] is held in row_offset.
+ *************************************************************************/
+
+class PIXROW:public ELIST_LINK
+{
+  public:
+    INT16 row_offset;            //y coord of min[0]
+    INT16 row_count;             //length of arrays
+    INT16 *min;                  //array of min x
+    INT16 *max;                  //array of max x
+
+    PIXROW() {  //empty constructor
+      row_offset = 0;
+      row_count = 0;
+      min = NULL;
+      max = NULL;
+    }
+    PIXROW(  //specified size
+           INT16 pos,
+           INT16 count,
+           PBLOB *blob);
+
+    ~PIXROW () {                 //destructor
+      if (min != NULL)
+        free_mem(min); 
+      if (max != NULL)
+        free_mem(max); 
+      max = NULL;
+    }
+
+    void plot(                   //use current settings
+              WINDOW fd) const;  //where to paint
+
+    BOX bounding_box() const;  //return bounding box
+                                 //return true if box exceeds image
+    bool bad_box(int xsize, int ysize) const; 
+
+    void contract(                           //force end on black
+                  IMAGELINE *imlines,        //image array
+                  INT16 x_offset,            //of pixels[0]
+                  INT16 foreground_colour);  //0 or 1
+
+                                 //image array
+    BOOL8 extend(IMAGELINE *imlines,
+                 BOX &imbox,
+                 PIXROW *prev,              //for prev blob
+                 PIXROW *next,              //for next blob
+                 INT16 foreground_colour);  //0 or 1
+
+                                 //box of imlines extnt
+    void char_clip_image(IMAGELINE *imlines,
+                         BOX &im_box,
+                         ROW *row,              //row containing word
+                         IMAGE &clip_image,     //unscaled char image
+                         float &baseline_pos);  //baseline ht in image
+
+};
+
+ELISTIZEH (PIXROW)
+extern INT_VAR_H (pix_word_margin, 3, "How far outside word BB to grow");
+extern BOOL_VAR_H (show_char_clipping, TRUE, "Show clip image window?");
+extern INT_VAR_H (net_image_width, 40, "NN input image width");
+extern INT_VAR_H (net_image_height, 36, "NN input image height");
+extern INT_VAR_H (net_image_x_height, 22, "NN input image x_height");
+void char_clip_word(                            //
+                    WERD *word,                 //word to be processed
+                    IMAGE &bin_image,           //whole image
+                    PIXROW_LIST *&pixrow_list,  //pixrows built
+                    IMAGELINE *&imlines,        //lines cut from image
+                    BOX &pix_box                //box defining imlines
+                   );
+IMAGELINE *generate_imlines(                   //get some imagelines
+                            IMAGE &bin_image,  //from here
+                            BOX &pix_box);
+                                 //word to be processed
+WINDOW display_clip_image(WERD *word,
+                          IMAGE &bin_image,          //whole image
+                          PIXROW_LIST *pixrow_list,  //pixrows built
+                          BOX &pix_box               //box of subimage
+                         );
+void display_images(IMAGE &clip_image, IMAGE &scaled_image); 
+void plot_pixrows(  //plot for all blobs
+                  PIXROW_LIST *pixrow_list,
+                  WINDOW win);
+#endif

ccmain/charsample.cpp

@@ -0,0 +1,698 @@
+/**********************************************************************
+ * File:        charsample.cpp  (Formerly charsample.c)
+ * Description: Class to contain character samples and match scores
+ *					to be used for adaption
+ * Author:      Chris Newton
+ * Created:     Thu Oct  7 13:40:37 BST 1993
+ *
+ * (C) Copyright 1993, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include <stdio.h>
+#include          <ctype.h>
+#include          <math.h>
+#ifdef __UNIX__
+#include <assert.h>
+#include          <unistd.h>
+#endif
+#include "memry.h"
+#include          "tessvars.h"
+#include "statistc.h"
+#include          "charsample.h"
+#include "paircmp.h"
+#include "matmatch.h"
+#include          "adaptions.h"
+#include          "secname.h"
+#include          "notdll.h"
+
+extern INT32 demo_word;          // Hack for demos
+
+ELISTIZE (CHAR_SAMPLE) ELISTIZE (CHAR_SAMPLES) CHAR_SAMPLE::CHAR_SAMPLE () {
+  sample_blob = NULL;
+  sample_denorm = NULL;
+  sample_image = NULL;
+  ch = '\0';
+  n_samples_matched = 0;
+  total_match_scores = 0.0;
+  sumsq_match_scores = 0.0;
+}
+
+
+CHAR_SAMPLE::CHAR_SAMPLE(PBLOB *blob, DENORM *denorm, char c) { 
+  sample_blob = blob;
+  sample_denorm = denorm;
+  sample_image = NULL;
+  ch = c;
+  n_samples_matched = 0;
+  total_match_scores = 0.0;
+  sumsq_match_scores = 0.0;
+}
+
+
+CHAR_SAMPLE::CHAR_SAMPLE(IMAGE *image, char c) { 
+  sample_blob = NULL;
+  sample_denorm = NULL;
+  sample_image = image;
+  ch = c;
+  n_samples_matched = 0;
+  total_match_scores = 0.0;
+  sumsq_match_scores = 0.0;
+}
+
+
+float CHAR_SAMPLE::match_sample(  // Update match scores
+                                CHAR_SAMPLE *test_sample,
+                                BOOL8 updating) {
+  float score1;
+  float score2;
+  IMAGE *image = test_sample->image ();
+
+  if (sample_blob != NULL && test_sample->blob () != NULL) {
+    PBLOB *blob = test_sample->blob ();
+    DENORM *denorm = test_sample->denorm ();
+
+    score1 = compare_bln_blobs (sample_blob, sample_denorm, blob, denorm);
+    score2 = compare_bln_blobs (blob, denorm, sample_blob, sample_denorm);
+
+    score1 = (score1 > score2) ? score1 : score2;
+  }
+  else if (sample_image != NULL && image != NULL) {
+    CHAR_PROTO *sample = new CHAR_PROTO (this);
+
+    score1 = matrix_match (sample_image, image);
+    delete sample;
+  }
+  else
+    return BAD_SCORE;
+
+  if ((tessedit_use_best_sample || tessedit_cluster_debug) && updating) {
+    n_samples_matched++;
+    total_match_scores += score1;
+    sumsq_match_scores += score1 * score1;
+  }
+  return score1;
+}
+
+
+double CHAR_SAMPLE::mean_score() { 
+  if (n_samples_matched > 0)
+    return (total_match_scores / n_samples_matched);
+  else
+    return BAD_SCORE;
+}
+
+
+double CHAR_SAMPLE::variance() { 
+  double mean = mean_score ();
+
+  if (n_samples_matched > 0) {
+    return (sumsq_match_scores / n_samples_matched) - mean * mean;
+  }
+  else
+    return BAD_SCORE;
+}
+
+
+void CHAR_SAMPLE::print(FILE *f) { 
+  if (!tessedit_cluster_debug)
+    return;
+
+  if (n_samples_matched > 0)
+    fprintf (f,
+      "%c - sample matched against " INT32FORMAT
+      " blobs, mean: %f, var: %f\n", ch, n_samples_matched,
+      mean_score (), variance ());
+  else
+    fprintf (f, "No matches for this sample (%c)\n", ch);
+}
+
+
+void CHAR_SAMPLE::reset_match_statistics() { 
+  n_samples_matched = 0;
+  total_match_scores = 0.0;
+  sumsq_match_scores = 0.0;
+}
+
+
+CHAR_SAMPLES::CHAR_SAMPLES() { 
+  type = UNKNOWN;
+  samples.clear ();
+  ch = '\0';
+  best_sample = NULL;
+  proto = NULL;
+}
+
+
+CHAR_SAMPLES::CHAR_SAMPLES(CHAR_SAMPLE *sample) { 
+  CHAR_SAMPLE_IT sample_it = &samples;
+
+  ASSERT_HOST (sample->image () != NULL || sample->blob () != NULL);
+
+  if (sample->image () != NULL)
+    type = IMAGE_CLUSTER;
+  else if (sample->blob () != NULL)
+    type = BLOB_CLUSTER;
+
+  samples.clear ();
+  sample_it.add_to_end (sample);
+  if (tessedit_mm_only_match_same_char)
+    ch = sample->character ();
+  else
+    ch = '\0';
+  best_sample = NULL;
+  proto = NULL;
+}
+
+
+void CHAR_SAMPLES::add_sample(CHAR_SAMPLE *sample) { 
+  CHAR_SAMPLE_IT sample_it = &samples;
+
+  if (tessedit_use_best_sample || tessedit_cluster_debug)
+    for (sample_it.mark_cycle_pt ();
+  !sample_it.cycled_list (); sample_it.forward ()) {
+    sample_it.data ()->match_sample (sample, TRUE);
+    sample->match_sample (sample_it.data (), TRUE);
+  }
+
+  sample_it.add_to_end (sample);
+
+  if (tessedit_mm_use_prototypes && type == IMAGE_CLUSTER)
+    if (samples.length () == tessedit_mm_prototype_min_size)
+      this->build_prototype ();
+  else if (samples.length () > tessedit_mm_prototype_min_size)
+    this->add_sample_to_prototype (sample);
+}
+
+
+void CHAR_SAMPLES::add_sample_to_prototype(CHAR_SAMPLE *sample) { 
+  BOOL8 rebuild = FALSE;
+  INT32 new_xsize = proto->x_size ();
+  INT32 new_ysize = proto->y_size ();
+  INT32 sample_xsize = sample->image ()->get_xsize ();
+  INT32 sample_ysize = sample->image ()->get_ysize ();
+
+  if (sample_xsize > new_xsize) {
+    new_xsize = sample_xsize;
+    rebuild = TRUE;
+  }
+  if (sample_ysize > new_ysize) {
+    new_ysize = sample_ysize;
+    rebuild = TRUE;
+  }
+
+  if (rebuild)
+    proto->enlarge_prototype (new_xsize, new_ysize);
+
+  proto->add_sample (sample);
+}
+
+
+void CHAR_SAMPLES::build_prototype() { 
+  CHAR_SAMPLE_IT sample_it = &samples;
+  CHAR_SAMPLE *sample;
+  INT32 proto_xsize = 0;
+  INT32 proto_ysize = 0;
+
+  if (type != IMAGE_CLUSTER
+    || samples.length () < tessedit_mm_prototype_min_size)
+    return;
+
+  for (sample_it.mark_cycle_pt ();
+  !sample_it.cycled_list (); sample_it.forward ()) {
+    sample = sample_it.data ();
+    if (sample->image ()->get_xsize () > proto_xsize)
+      proto_xsize = sample->image ()->get_xsize ();
+    if (sample->image ()->get_ysize () > proto_ysize)
+      proto_ysize = sample->image ()->get_ysize ();
+  }
+
+  proto = new CHAR_PROTO (proto_xsize, proto_ysize, 0, 0, '\0');
+
+  for (sample_it.mark_cycle_pt ();
+    !sample_it.cycled_list (); sample_it.forward ())
+  this->add_sample_to_prototype (sample_it.data ());
+
+}
+
+
+void CHAR_SAMPLES::find_best_sample() { 
+  CHAR_SAMPLE_IT sample_it = &samples;
+  double score;
+  double best_score = MAX_INT32;
+
+  if (ch == '\0' || samples.length () < tessedit_mm_prototype_min_size)
+    return;
+
+  for (sample_it.mark_cycle_pt ();
+  !sample_it.cycled_list (); sample_it.forward ()) {
+    score = sample_it.data ()->mean_score ();
+    if (score < best_score) {
+      best_score = score;
+      best_sample = sample_it.data ();
+    }
+  }
+  #ifndef SECURE_NAMES
+  if (tessedit_cluster_debug) {
+    tprintf ("Best sample for this %c cluster:\n", ch);
+    best_sample->print (debug_fp);
+  }
+  #endif
+}
+
+
+float CHAR_SAMPLES::match_score(CHAR_SAMPLE *sample) { 
+  if (tessedit_mm_only_match_same_char && sample->character () != ch)
+    return BAD_SCORE;
+
+  if (tessedit_use_best_sample && best_sample != NULL)
+    return best_sample->match_sample (sample, FALSE);
+  else if ((tessedit_mm_use_prototypes
+    || tessedit_mm_adapt_using_prototypes) && proto != NULL)
+    return proto->match_sample (sample);
+  else
+    return this->nn_match_score (sample);
+}
+
+
+float CHAR_SAMPLES::nn_match_score(CHAR_SAMPLE *sample) { 
+  CHAR_SAMPLE_IT sample_it = &samples;
+  float score;
+  float min_score = MAX_INT32;
+
+  for (sample_it.mark_cycle_pt ();
+  !sample_it.cycled_list (); sample_it.forward ()) {
+    score = sample_it.data ()->match_sample (sample, FALSE);
+    if (score < min_score)
+      min_score = score;
+  }
+
+  return min_score;
+}
+
+
+void CHAR_SAMPLES::assign_to_char() { 
+  STATS char_frequency(FIRST_CHAR, LAST_CHAR); 
+  CHAR_SAMPLE_IT sample_it = &samples;
+  INT32 i;
+  INT32 max_index = 0;
+  INT32 max_freq = 0;
+
+  if (samples.length () == 0 || tessedit_mm_only_match_same_char)
+    return;
+
+  for (sample_it.mark_cycle_pt ();
+    !sample_it.cycled_list (); sample_it.forward ())
+  char_frequency.add ((INT32) sample_it.data ()->character (), 1);
+
+  for (i = FIRST_CHAR; i <= LAST_CHAR; i++)
+  if (char_frequency.pile_count (i) > max_freq) {
+    max_index = i;
+    max_freq = char_frequency.pile_count (i);
+  }
+
+  if (samples.length () >= tessedit_cluster_min_size
+    && max_freq > samples.length () * tessedit_cluster_accept_fraction)
+    ch = (char) max_index;
+}
+
+
+void CHAR_SAMPLES::print(FILE *f) { 
+  CHAR_SAMPLE_IT sample_it = &samples;
+
+  fprintf (f, "Collected " INT32FORMAT " samples\n", samples.length ());
+
+  #ifndef SECURE_NAMES
+  if (tessedit_cluster_debug)
+    for (sample_it.mark_cycle_pt ();
+    !sample_it.cycled_list (); sample_it.forward ())
+  sample_it.data ()->print (f);
+
+  if (ch == '\0')
+    fprintf (f, "\nCluster not used for adaption\n");
+  else
+    fprintf (f, "\nCluster used to adapt to '%c's\n", ch);
+  #endif
+}
+
+
+CHAR_PROTO::CHAR_PROTO() { 
+  xsize = 0;
+  ysize = 0;
+  ch = '\0';
+  nsamples = 0;
+  proto_data = NULL;
+  proto = NULL;
+}
+
+
+CHAR_PROTO::CHAR_PROTO(INT32 x_size,
+                       INT32 y_size,
+                       INT32 n_samples,
+                       float initial_value,
+                       char c) {
+  INT32 x;
+  INT32 y;
+
+  xsize = x_size;
+  ysize = y_size;
+  ch = c;
+  nsamples = n_samples;
+
+  ALLOC_2D_ARRAY(xsize, ysize, proto_data, proto, float); 
+
+  for (y = 0; y < ysize; y++)
+    for (x = 0; x < xsize; x++)
+      proto[x][y] = initial_value;
+}
+
+
+CHAR_PROTO::CHAR_PROTO(CHAR_SAMPLE *sample) { 
+  INT32 x;
+  INT32 y;
+  IMAGELINE imline_s;
+
+  if (sample->image () == NULL) {
+    xsize = 0;
+    ysize = 0;
+    ch = '\0';
+    nsamples = 0;
+    proto_data = NULL;
+    proto = NULL;
+  }
+  else {
+    ch = sample->character ();
+    xsize = sample->image ()->get_xsize ();
+    ysize = sample->image ()->get_ysize ();
+    nsamples = 1;
+
+    ALLOC_2D_ARRAY(xsize, ysize, proto_data, proto, float); 
+
+    for (y = 0; y < ysize; y++) {
+      sample->image ()->fast_get_line (0, y, xsize, &imline_s);
+      for (x = 0; x < xsize; x++)
+        if (imline_s.pixels[x] == BINIM_WHITE)
+          proto[x][y] = 1.0;
+      else
+        proto[x][y] = -1.0;
+    }
+  }
+}
+
+
+CHAR_PROTO::~CHAR_PROTO () {
+  if (proto_data != NULL)
+    FREE_2D_ARRAY(proto_data, proto); 
+}
+
+
+float CHAR_PROTO::match_sample(CHAR_SAMPLE *test_sample) { 
+  CHAR_PROTO *test_proto;
+  float score;
+
+  if (test_sample->image () != NULL) {
+    test_proto = new CHAR_PROTO (test_sample);
+    if (xsize > test_proto->x_size ())
+      score = this->match (test_proto);
+    else {
+      demo_word = -demo_word;    // Flag different call
+      score = test_proto->match (this);
+    }
+  }
+  else
+    return BAD_SCORE;
+
+  delete test_proto;
+
+  return score;
+}
+
+
+float CHAR_PROTO::match(CHAR_PROTO *test_proto) { 
+  INT32 xsize2 = test_proto->x_size ();
+  INT32 y_size;
+  INT32 y_size2;
+  INT32 x_offset;
+  INT32 y_offset;
+  INT32 x;
+  INT32 y;
+  CHAR_PROTO *match_proto;
+  float score;
+  float sum = 0.0;
+
+  ASSERT_HOST (xsize >= xsize2);
+
+  x_offset = (xsize - xsize2) / 2;
+
+  if (ysize < test_proto->y_size ()) {
+    y_size = test_proto->y_size ();
+    y_size2 = ysize;
+    y_offset = (y_size - y_size2) / 2;
+
+    match_proto = new CHAR_PROTO (xsize,
+      y_size,
+      nsamples * test_proto->n_samples (),
+      0, '\0');
+
+    for (y = 0; y < y_offset; y++) {
+      for (x = 0; x < xsize2; x++) {
+        match_proto->data ()[x + x_offset][y] =
+          test_proto->data ()[x][y] * nsamples;
+        sum += match_proto->data ()[x + x_offset][y];
+      }
+    }
+
+    for (y = y_offset + y_size2; y < y_size; y++) {
+      for (x = 0; x < xsize2; x++) {
+        match_proto->data ()[x + x_offset][y] =
+          test_proto->data ()[x][y] * nsamples;
+        sum += match_proto->data ()[x + x_offset][y];
+      }
+    }
+
+    for (y = y_offset; y < y_offset + y_size2; y++) {
+      for (x = 0; x < x_offset; x++) {
+        match_proto->data ()[x][y] = proto[x][y - y_offset] *
+          test_proto->n_samples ();
+        sum += match_proto->data ()[x][y];
+      }
+
+      for (x = x_offset + xsize2; x < xsize; x++) {
+        match_proto->data ()[x][y] = proto[x][y - y_offset] *
+          test_proto->n_samples ();
+        sum += match_proto->data ()[x][y];
+      }
+
+      for (x = x_offset; x < x_offset + xsize2; x++) {
+        match_proto->data ()[x][y] =
+          proto[x][y - y_offset] * test_proto->data ()[x - x_offset][y];
+        sum += match_proto->data ()[x][y];
+      }
+    }
+  }
+  else {
+    y_size = ysize;
+    y_size2 = test_proto->y_size ();
+    y_offset = (y_size - y_size2) / 2;
+
+    match_proto = new CHAR_PROTO (xsize,
+      y_size,
+      nsamples * test_proto->n_samples (),
+      0, '\0');
+
+    for (y = 0; y < y_offset; y++)
+    for (x = 0; x < xsize; x++) {
+      match_proto->data ()[x][y] =
+        proto[x][y] * test_proto->n_samples ();
+      sum += match_proto->data ()[x][y];
+    }
+
+    for (y = y_offset + y_size2; y < y_size; y++)
+    for (x = 0; x < xsize; x++) {
+      match_proto->data ()[x][y] =
+        proto[x][y] * test_proto->n_samples ();
+      sum += match_proto->data ()[x][y];
+    }
+
+    for (y = y_offset; y < y_offset + y_size2; y++) {
+      for (x = 0; x < x_offset; x++) {
+        match_proto->data ()[x][y] =
+          proto[x][y] * test_proto->n_samples ();
+        sum += match_proto->data ()[x][y];
+      }
+
+      for (x = x_offset + xsize2; x < xsize; x++) {
+        match_proto->data ()[x][y] =
+          proto[x][y] * test_proto->n_samples ();
+        sum += match_proto->data ()[x][y];
+      }
+
+      for (x = x_offset; x < x_offset + xsize2; x++) {
+        match_proto->data ()[x][y] = proto[x][y] *
+          test_proto->data ()[x - x_offset][y - y_offset];
+        sum += match_proto->data ()[x][y];
+      }
+    }
+  }
+
+  score = (1.0 - sum /
+    (xsize * y_size * nsamples * test_proto->n_samples ()));
+
+  if (tessedit_mm_debug) {
+    if (score < 0) {
+      tprintf ("Match score %f\n", score);
+      tprintf ("x: %d, y: %d, ns: %d, nt: %d, dx %d, dy: %d\n",
+        xsize, y_size, nsamples, test_proto->n_samples (),
+        x_offset, y_offset);
+      for (y = 0; y < y_size; y++) {
+        tprintf ("\n%d", y);
+        for (x = 0; x < xsize; x++)
+          tprintf ("\t%d", match_proto->data ()[x][y]);
+
+      }
+      tprintf ("\n");
+      fflush(debug_fp); 
+    }
+  }
+
+#ifndef GRAPHICS_DISABLED
+  if (tessedit_display_mm) {
+    tprintf ("Match score %f\n", score);
+    display_images (this->make_image (),
+      test_proto->make_image (), match_proto->make_image ());
+  }
+  else if (demo_word != 0) {
+    if (demo_word > 0)
+      display_image (test_proto->make_image (), "Test sample",
+        300, 400, FALSE);
+    else
+      display_image (this->make_image (), "Test sample", 300, 400, FALSE);
+
+    display_image (match_proto->make_image (), "Best match",
+      700, 400, TRUE);
+  }
+#endif
+
+  delete match_proto;
+
+  return score;
+}
+
+
+void CHAR_PROTO::enlarge_prototype(INT32 new_xsize, INT32 new_ysize) { 
+  float *old_proto_data = proto_data;
+  float **old_proto = proto;
+  INT32 old_xsize = xsize;
+  INT32 old_ysize = ysize;
+  INT32 x_offset;
+  INT32 y_offset;
+  INT32 x;
+  INT32 y;
+
+  ASSERT_HOST (new_xsize >= xsize && new_ysize >= ysize);
+
+  xsize = new_xsize;
+  ysize = new_ysize;
+  ALLOC_2D_ARRAY(xsize, ysize, proto_data, proto, float); 
+  x_offset = (xsize - old_xsize) / 2;
+  y_offset = (ysize - old_ysize) / 2;
+
+  for (y = 0; y < y_offset; y++)
+    for (x = 0; x < xsize; x++)
+      proto[x][y] = nsamples;
+
+  for (y = y_offset + old_ysize; y < ysize; y++)
+    for (x = 0; x < xsize; x++)
+      proto[x][y] = nsamples;
+
+  for (y = y_offset; y < y_offset + old_ysize; y++) {
+    for (x = 0; x < x_offset; x++)
+      proto[x][y] = nsamples;
+
+    for (x = x_offset + old_xsize; x < xsize; x++)
+      proto[x][y] = nsamples;
+
+    for (x = x_offset; x < x_offset + old_xsize; x++)
+      proto[x][y] = old_proto[x - x_offset][y - y_offset];
+  }
+
+  FREE_2D_ARRAY(old_proto_data, old_proto); 
+}
+
+
+void CHAR_PROTO::add_sample(CHAR_SAMPLE *sample) { 
+  INT32 x_offset;
+  INT32 y_offset;
+  INT32 x;
+  INT32 y;
+  IMAGELINE imline_s;
+  INT32 sample_xsize = sample->image ()->get_xsize ();
+  INT32 sample_ysize = sample->image ()->get_ysize ();
+
+  x_offset = (xsize - sample_xsize) / 2;
+  y_offset = (ysize - sample_ysize) / 2;
+
+  ASSERT_HOST (x_offset >= 0 && y_offset >= 0);
+
+  for (y = 0; y < y_offset; y++)
+    for (x = 0; x < xsize; x++)
+      proto[x][y]++;             // Treat pixels outside the
+  // range as white
+  for (y = y_offset + sample_ysize; y < ysize; y++)
+    for (x = 0; x < xsize; x++)
+      proto[x][y]++;
+
+  for (y = y_offset; y < y_offset + sample_ysize; y++) {
+    sample->image ()->fast_get_line (0,
+      y - y_offset, sample_xsize, &imline_s);
+    for (x = x_offset; x < x_offset + sample_xsize; x++) {
+      if (imline_s.pixels[x - x_offset] == BINIM_WHITE)
+        proto[x][y]++;
+      else
+        proto[x][y]--;
+    }
+
+    for (x = 0; x < x_offset; x++)
+      proto[x][y]++;
+
+    for (x = x_offset + sample_xsize; x < xsize; x++)
+      proto[x][y]++;
+  }
+
+  nsamples++;
+}
+
+
+IMAGE *CHAR_PROTO::make_image() { 
+  IMAGE *image;
+  IMAGELINE imline_p;
+  INT32 x;
+  INT32 y;
+
+  ASSERT_HOST (nsamples != 0);
+
+  image = new (IMAGE);
+  image->create (xsize, ysize, 8);
+
+  for (y = 0; y < ysize; y++) {
+    image->fast_get_line (0, y, xsize, &imline_p);
+
+    for (x = 0; x < xsize; x++) {
+      imline_p.pixels[x] = 128 +
+        (UINT8) ((proto[x][y] * 128.0) / (0.00001 + nsamples));
+    }
+
+    image->fast_put_line (0, y, xsize, &imline_p);
+  }
+  return image;
+}

ccmain/control.h

@@ -0,0 +1,193 @@
+/**********************************************************************
+ * File:        control.h  (Formerly control.h)
+ * Description: Module-independent matcher controller.
+ * Author:		Ray Smith
+ * Created:		Thu Apr 23 11:09:58 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           CONTROL_H
+#define           CONTROL_H
+
+#include          "varable.h"
+#include          "ocrblock.h"
+//#include                                      "epapdest.h"
+#include          "ratngs.h"
+#include          "statistc.h"
+//#include                                      "epapconv.h"
+#include          "ocrshell.h"
+#include          "pageres.h"
+#include          "charsample.h"
+#include          "notdll.h"
+
+enum ACCEPTABLE_WERD_TYPE
+{
+  AC_UNACCEPTABLE,               //Unacceptable word
+  AC_LOWER_CASE,                 //ALL lower case
+  AC_UPPER_CASE,                 //ALL upper case
+  AC_INITIAL_CAP,                //ALL but initial lc
+  AC_LC_ABBREV,                  //a.b.c.
+  AC_UC_ABBREV                   //A.B.C.
+};
+
+typedef BOOL8 (*BLOB_REJECTOR) (PBLOB *, BLOB_CHOICE_IT *, void *);
+
+extern INT_VAR_H (tessedit_single_match, FALSE, "Top choice only from CP");
+//extern BOOL_VAR_H(tessedit_small_match,FALSE,"Use small matrix matcher");
+extern BOOL_VAR_H (tessedit_print_text, FALSE, "Write text to stdout");
+extern BOOL_VAR_H (tessedit_draw_words, FALSE, "Draw source words");
+extern BOOL_VAR_H (tessedit_draw_outwords, FALSE, "Draw output words");
+extern BOOL_VAR_H (tessedit_training_wiseowl, FALSE,
+"Call WO to learn blobs");
+extern BOOL_VAR_H (tessedit_training_tess, FALSE, "Call Tess to learn blobs");
+extern BOOL_VAR_H (tessedit_matcher_is_wiseowl, FALSE, "Call WO to classify");
+extern BOOL_VAR_H (tessedit_dump_choices, FALSE, "Dump char choices");
+extern BOOL_VAR_H (tessedit_fix_fuzzy_spaces, TRUE,
+"Try to improve fuzzy spaces");
+extern BOOL_VAR_H (tessedit_unrej_any_wd, FALSE,
+"Dont bother with word plausibility");
+extern BOOL_VAR_H (tessedit_fix_hyphens, TRUE, "Crunch double hyphens?");
+extern BOOL_VAR_H (tessedit_reject_fullstops, FALSE, "Reject all fullstops");
+extern BOOL_VAR_H (tessedit_reject_suspect_fullstops, FALSE,
+"Reject suspect fullstops");
+extern BOOL_VAR_H (tessedit_redo_xheight, TRUE, "Check/Correct x-height");
+extern BOOL_VAR_H (tessedit_cluster_adaption_on, TRUE,
+"Do our own adaption - ems only");
+extern BOOL_VAR_H (tessedit_enable_doc_dict, TRUE,
+"Add words to the document dictionary");
+extern BOOL_VAR_H (word_occ_first, FALSE, "Do word occ before re-est xht");
+extern BOOL_VAR_H (tessedit_xht_fiddles_on_done_wds, TRUE,
+"Apply xht fix up even if done");
+extern BOOL_VAR_H (tessedit_xht_fiddles_on_no_rej_wds, TRUE,
+"Apply xht fix up even in no rejects");
+extern INT_VAR_H (x_ht_check_word_occ, 2, "Check Char Block occupancy");
+extern INT_VAR_H (x_ht_stringency, 1, "How many confirmed a/n to accept?");
+extern BOOL_VAR_H (x_ht_quality_check, TRUE, "Dont allow worse quality");
+extern BOOL_VAR_H (tessedit_debug_block_rejection, FALSE,
+"Block and Row stats");
+extern INT_VAR_H (debug_x_ht_level, 0, "Reestimate debug");
+extern BOOL_VAR_H (rej_use_xht, TRUE, "Individual rejection control");
+extern BOOL_VAR_H (debug_acceptable_wds, FALSE, "Dump word pass/fail chk");
+extern STRING_VAR_H (chs_leading_punct, "('`\"", "Leading punctuation");
+extern
+STRING_VAR_H (chs_trailing_punct1, ").,;:?!", "1st Trailing punctuation");
+extern STRING_VAR_H (chs_trailing_punct2, ")'`\"",
+"2nd Trailing punctuation");
+extern double_VAR_H (quality_rej_pc, 0.08,
+"good_quality_doc lte rejection limit");
+extern double_VAR_H (quality_blob_pc, 0.0,
+"good_quality_doc gte good blobs limit");
+extern double_VAR_H (quality_outline_pc, 1.0,
+"good_quality_doc lte outline error limit");
+extern double_VAR_H (quality_char_pc, 0.95,
+"good_quality_doc gte good char limit");
+extern INT_VAR_H (quality_min_initial_alphas_reqd, 2,
+"alphas in a good word");
+extern BOOL_VAR_H (tessedit_tess_adapt_to_rejmap, FALSE,
+"Use reject map to control Tesseract adaption");
+extern INT_VAR_H (tessedit_tess_adaption_mode, 3,
+"Adaptation decision algorithm for tess");
+extern INT_VAR_H (tessedit_em_adaption_mode, 62,
+"Adaptation decision algorithm for ems matrix matcher");
+extern BOOL_VAR_H (tessedit_cluster_adapt_after_pass1, FALSE,
+"Adapt using clusterer after pass 1");
+extern BOOL_VAR_H (tessedit_cluster_adapt_after_pass2, FALSE,
+"Adapt using clusterer after pass 1");
+extern BOOL_VAR_H (tessedit_cluster_adapt_after_pass3, FALSE,
+"Adapt using clusterer after pass 1");
+extern BOOL_VAR_H (tessedit_cluster_adapt_before_pass1, FALSE,
+"Adapt using clusterer before Tess adaping during pass 1");
+extern INT_VAR_H (tessedit_cluster_adaption_mode, 0,
+"Adaptation decision algorithm for matrix matcher");
+extern BOOL_VAR_H (tessedit_adaption_debug, FALSE,
+"Generate and print debug information for adaption");
+extern BOOL_VAR_H (tessedit_minimal_rej_pass1, FALSE,
+"Do minimal rejection on pass 1 output");
+extern BOOL_VAR_H (tessedit_test_adaption, FALSE,
+"Test adaption criteria");
+extern BOOL_VAR_H (tessedit_global_adaption, FALSE,
+"Adapt to all docs over time");
+extern BOOL_VAR_H (tessedit_matcher_log, FALSE, "Log matcher activity");
+extern INT_VAR_H (tessedit_test_adaption_mode, 3,
+"Adaptation decision algorithm for tess");
+extern BOOL_VAR_H (test_pt, FALSE, "Test for point");
+extern double_VAR_H (test_pt_x, 99999.99, "xcoord");
+extern double_VAR_H (test_pt_y, 99999.99, "ycoord");
+void recog_pseudo_word(                         //recognize blobs
+                       BLOCK_LIST *block_list,  //blocks to check
+                       BOX &selection_box);
+BOOL8 recog_interactive(            //recognize blobs
+                        BLOCK *,    //block
+                        ROW *row,   //row of word
+                        WERD *word  //word to recognize
+                       );
+void recog_all_words(                              //process words
+                     PAGE_RES *page_res,           //page structure
+                     volatile ETEXT_DESC *monitor  //progress monitor
+                    );
+void classify_word_pass1(                 //recog one word
+                         WERD_RES *word,  //word to do
+                         ROW *row,
+                         BOOL8 cluster_adapt,
+                         CHAR_SAMPLES_LIST *char_clusters,
+                         CHAR_SAMPLE_LIST *chars_waiting);
+                                 //word to do
+void classify_word_pass2(WERD_RES *word, ROW *row); 
+void match_word_pass2(                 //recog one word
+                      WERD_RES *word,  //word to do
+                      ROW *row,
+                      float x_height);
+void fix_rep_char(                //Repeated char word
+                  WERD_RES *word  //word to do
+                 );
+void fix_quotes(               //make double quotes
+                char *string,  //string to fix
+                WERD *word,    //word to do //char choices
+                BLOB_CHOICE_LIST_CLIST *blob_choices);
+void fix_hyphens(               //crunch double hyphens
+                 char *string,  //string to fix
+                 WERD *word,    //word to do //char choices
+                 BLOB_CHOICE_LIST_CLIST *blob_choices);
+void merge_blobs(               //combine 2 blobs
+                 PBLOB *blob1,  //dest blob
+                 PBLOB *blob2   //source blob
+                );
+void choice_dump_tester(                           //dump chars in word
+                        PBLOB *,                   //blob
+                        DENORM *,                  //de-normaliser
+                        BOOL8 correct,             //ly segmented
+                        char *text,                //correct text
+                        INT32 count,               //chars in text
+                        BLOB_CHOICE_LIST *ratings  //list of results
+                       );
+WERD *make_bln_copy(WERD *src_word, ROW *row, float x_height, DENORM *denorm); 
+ACCEPTABLE_WERD_TYPE acceptable_word_string(const char *s); 
+BOOL8 check_debug_pt(WERD_RES *word, int location); 
+void set_word_fonts(                 //good chars in word
+                    WERD_RES *word,  //word to adapt to //detailed results
+                    BLOB_CHOICE_LIST_CLIST *blob_choices);
+void font_recognition_pass(  //good chars in word
+                           PAGE_RES_IT &page_res_it);
+void add_in_one_row(               //good chars in word
+                    ROW_RES *row,  //current row
+                    STATS *fonts,  //font stats
+                    INT8 *italic,  //output count
+                    INT8 *bold     //output count
+                   );
+void find_modal_font(                  //good chars in word
+                     STATS *fonts,     //font stats
+                     INT8 *font_out,   //output font
+                     INT8 *font_count  //output count
+                    );
+#endif

ccmain/docqual.h

@@ -0,0 +1,155 @@
+/******************************************************************
+ * File:        docqual.h  (Formerly docqual.h)
+ * Description: Document Quality Metrics
+ * Author:		Phil Cheatle
+ * Created:		Mon May  9 11:27:28 BST 1994
+ *
+ * (C) Copyright 1994, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           DOCQUAL_H
+#define           DOCQUAL_H
+
+#include          "control.h"
+#include          "notdll.h"
+
+enum GARBAGE_LEVEL
+{
+  G_NEVER_CRUNCH,
+  G_OK,
+  G_DODGY,
+  G_TERRIBLE
+};
+
+extern STRING_VAR_H (outlines_odd, "%| ", "Non standard number of outlines");
+extern STRING_VAR_H (outlines_2, "ij!?%\":;",
+"Non standard number of outlines");
+extern BOOL_VAR_H (docqual_excuse_outline_errs, FALSE,
+"Allow outline errs in unrejection?");
+extern BOOL_VAR_H (tessedit_good_quality_unrej, TRUE,
+"Reduce rejection on good docs");
+extern BOOL_VAR_H (tessedit_use_reject_spaces, TRUE, "Reject spaces?");
+extern double_VAR_H (tessedit_reject_doc_percent, 65.00,
+"%rej allowed before rej whole doc");
+extern double_VAR_H (tessedit_reject_block_percent, 45.00,
+"%rej allowed before rej whole block");
+extern double_VAR_H (tessedit_reject_row_percent, 40.00,
+"%rej allowed before rej whole row");
+extern double_VAR_H (tessedit_whole_wd_rej_row_percent, 70.00,
+"%of row rejects in whole word rejects which prevents whole row rejection");
+extern BOOL_VAR_H (tessedit_preserve_blk_rej_perfect_wds, TRUE,
+"Only rej partially rejected words in block rejection");
+extern BOOL_VAR_H (tessedit_preserve_row_rej_perfect_wds, TRUE,
+"Only rej partially rejected words in row rejection");
+extern BOOL_VAR_H (tessedit_dont_blkrej_good_wds, FALSE,
+"Use word segmentation quality metric");
+extern BOOL_VAR_H (tessedit_dont_rowrej_good_wds, FALSE,
+"Use word segmentation quality metric");
+extern INT_VAR_H (tessedit_preserve_min_wd_len, 2,
+"Only preserve wds longer than this");
+extern BOOL_VAR_H (tessedit_row_rej_good_docs, TRUE,
+"Apply row rejection to good docs");
+extern double_VAR_H (tessedit_good_doc_still_rowrej_wd, 1.1,
+"rej good doc wd if more than this fraction rejected");
+extern BOOL_VAR_H (tessedit_reject_bad_qual_wds, TRUE,
+"Reject all bad quality wds");
+extern BOOL_VAR_H (tessedit_debug_doc_rejection, FALSE, "Page stats");
+extern BOOL_VAR_H (tessedit_debug_quality_metrics, FALSE,
+"Output data to debug file");
+extern BOOL_VAR_H (bland_unrej, FALSE, "unrej potential with no chekcs");
+extern double_VAR_H (quality_rowrej_pc, 1.1,
+"good_quality_doc gte good char limit");
+extern BOOL_VAR_H (unlv_tilde_crunching, TRUE,
+"Mark v.bad words for tilde crunch");
+extern BOOL_VAR_H (crunch_early_merge_tess_fails, TRUE,
+"Before word crunch?");
+extern BOOL_VAR_H (crunch_early_convert_bad_unlv_chs, FALSE,
+"Take out ~^ early?");
+extern double_VAR_H (crunch_terrible_rating, 80.0, "crunch rating lt this");
+extern BOOL_VAR_H (crunch_terrible_garbage, TRUE, "As it says");
+extern double_VAR_H (crunch_poor_garbage_cert, -9.0,
+"crunch garbage cert lt this");
+extern double_VAR_H (crunch_poor_garbage_rate, 60,
+"crunch garbage rating lt this");
+extern double_VAR_H (crunch_pot_poor_rate, 40,
+"POTENTIAL crunch rating lt this");
+extern double_VAR_H (crunch_pot_poor_cert, -8.0,
+"POTENTIAL crunch cert lt this");
+extern BOOL_VAR_H (crunch_pot_garbage, TRUE, "POTENTIAL crunch garbage");
+extern double_VAR_H (crunch_del_rating, 60,
+"POTENTIAL crunch rating lt this");
+extern double_VAR_H (crunch_del_cert, -10.0, "POTENTIAL crunch cert lt this");
+extern double_VAR_H (crunch_del_min_ht, 0.7, "Del if word ht lt xht x this");
+extern double_VAR_H (crunch_del_max_ht, 3.0, "Del if word ht gt xht x this");
+extern double_VAR_H (crunch_del_min_width, 3.0,
+"Del if word width lt xht x this");
+extern double_VAR_H (crunch_del_high_word, 1.5,
+"Del if word gt xht x this above bl");
+extern double_VAR_H (crunch_del_low_word, 0.5,
+"Del if word gt xht x this below bl");
+extern double_VAR_H (crunch_small_outlines_size, 0.6,
+"Small if lt xht x this");
+extern INT_VAR_H (crunch_rating_max, 10, "For adj length in rating per ch");
+extern INT_VAR_H (crunch_pot_indicators, 1,
+"How many potential indicators needed");
+extern BOOL_VAR_H (crunch_leave_ok_strings, TRUE,
+"Dont touch sensible strings");
+extern BOOL_VAR_H (crunch_accept_ok, TRUE, "Use acceptability in okstring");
+extern BOOL_VAR_H (crunch_leave_accept_strings, FALSE,
+"Dont pot crunch sensible strings");
+extern BOOL_VAR_H (crunch_include_numerals, FALSE, "Fiddle alpha figures");
+extern INT_VAR_H (crunch_leave_lc_strings, 4,
+"Dont crunch words with long lower case strings");
+extern INT_VAR_H (crunch_leave_uc_strings, 4,
+"Dont crunch words with long lower case strings");
+extern INT_VAR_H (crunch_long_repetitions, 3,
+"Crunch words with long repetitions");
+extern INT_VAR_H (crunch_debug, 0, "As it says");
+INT16 word_blob_quality(  //Blob seg changes
+                        WERD_RES *word,
+                        ROW *row);
+BOOL8 crude_match_blobs(PBLOB *blob1, PBLOB *blob2); 
+INT16 word_outline_errs(  //Outline count errs
+                        WERD_RES *word);
+void word_char_quality(  //Blob seg changes
+                       WERD_RES *word,
+                       ROW *row,
+                       INT16 *match_count,
+                       INT16 *accepted_match_count);
+void unrej_good_chs(WERD_RES *word, ROW *row); 
+void print_boxes(WERD *word); 
+INT16 count_outline_errs(char c, INT16 outline_count); 
+void quality_based_rejection(PAGE_RES_IT &page_res_it, BOOL8 good_quality_doc); 
+void unrej_good_quality_words(  //unreject potential
+                              PAGE_RES_IT &page_res_it);
+void doc_and_block_rejection(  //reject big chunks
+                             PAGE_RES_IT &page_res_it,
+                             BOOL8 good_quality_doc);
+void reject_whole_page(PAGE_RES_IT &page_res_it); 
+void tilde_crunch(PAGE_RES_IT &page_res_it); 
+BOOL8 terrible_word_crunch(WERD_RES *word, GARBAGE_LEVEL garbage_level); 
+BOOL8 potential_word_crunch(WERD_RES *word,
+                            GARBAGE_LEVEL garbage_level,
+                            BOOL8 ok_dict_word);
+void tilde_delete(PAGE_RES_IT &page_res_it); 
+                                 //word to do
+void convert_bad_unlv_chs(WERD_RES *word_res); 
+                                 //word to do
+void merge_tess_fails(WERD_RES *word_res); 
+GARBAGE_LEVEL garbage_word(WERD_RES *word, BOOL8 ok_dict_word); 
+CRUNCH_MODE word_deletable(WERD_RES *word, INT16 &delete_mode); 
+INT16 failure_count(WERD_RES *word); 
+BOOL8 noise_outlines(WERD *word); 
+                                 //word to do
+void insert_rej_cblobs(WERD_RES *word); 
+#endif

ccmain/expandblob.cpp

@@ -0,0 +1,82 @@
+/**************************************************************************
+ * Revision 5.1  89/07/27  11:46:53  11:46:53  ray ()
+ * (C) Copyright 1989, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+**************************************************************************/
+#include "mfcpch.h"
+#include "expandblob.h"
+#include "tessclas.h"
+#include "const.h"
+#include "structures.h"
+#include "freelist.h"
+
+/***********************************************************************
+free_blob(blob) frees the blob and everything it is connected to,
+i.e. outlines, nodes, edgepts, bytevecs, ratings etc
+*************************************************************************/
+void free_blob(  /*blob to free */
+               register TBLOB *blob) {
+  if (blob == NULL)
+    return;                      /*duff blob */
+  free_tree (blob->outlines);    /*do the tree of outlines */
+  oldblob(blob);  /*free the actual blob */
+}
+
+
+/***************************************************************************
+free_tree(outline) frees the current outline
+and then its sub-tree
+*****************************************************************************/
+void free_tree(  /*outline to draw */
+               register TESSLINE *outline) {
+  if (outline == NULL)
+    return;                      /*duff outline */
+  if (outline->next != NULL)
+    free_tree (outline->next);
+  if (outline->child != NULL)
+    free_tree (outline->child);  /*and sub-tree */
+  free_outline(outline);  /*free the outline */
+}
+
+
+/*******************************************************************************
+free_outline(outline) frees an outline and anything connected to it
+*********************************************************************************/
+void free_outline(  /*outline to free */
+                  register TESSLINE *outline) {
+  if (outline->compactloop != NULL)
+                                 /*no compact loop */
+      memfree (outline->compactloop);
+
+  if (outline->loop != NULL)
+    free_loop (outline->loop);
+
+  oldoutline(outline);
+}
+
+
+/*********************************************************************************
+free_loop(startpt) frees all the elements of the closed loop
+starting at startpt
+***********************************************************************************/
+void free_loop(  /*outline to free */
+               register EDGEPT *startpt) {
+  register EDGEPT *edgept;       /*current point */
+
+  if (startpt == NULL)
+    return;
+  edgept = startpt;
+  do {
+    edgept = oldedgept (edgept); /*free it and move on */
+  }
+  while (edgept != startpt);
+}

ccmain/expandblob.h

@@ -0,0 +1,13 @@
+#ifndef EXPANDBLOB_H
+#define EXPANDBLOB_H
+
+#include "tessclas.h"
+
+void free_blob(register TBLOB *blob); 
+
+void free_tree(register TESSLINE *outline); 
+
+void free_outline(register TESSLINE *outline); 
+
+void free_loop(register EDGEPT *startpt); 
+#endif

ccmain/fixspace.cpp

@@ -0,0 +1,974 @@
+/******************************************************************
+ * File:        fixspace.cpp  (Formerly fixspace.c)
+ * Description: Implements a pass over the page res, exploring the alternative
+ *					spacing possibilities, trying to use context to improve the
+          word spacing
+* Author:		Phil Cheatle
+* Created:		Thu Oct 21 11:38:43 BST 1993
+*
+* (C) Copyright 1993, Hewlett-Packard Ltd.
+** Licensed under the Apache License, Version 2.0 (the "License");
+** you may not use this file except in compliance with the License.
+** You may obtain a copy of the License at
+** http://www.apache.org/licenses/LICENSE-2.0
+** Unless required by applicable law or agreed to in writing, software
+** distributed under the License is distributed on an "AS IS" BASIS,
+** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+** See the License for the specific language governing permissions and
+** limitations under the License.
+*
+**********************************************************************/
+
+#include "mfcpch.h"
+#include          <ctype.h>
+#include          "reject.h"
+#include          "statistc.h"
+#include          "genblob.h"
+#include          "control.h"
+#include          "fixspace.h"
+#include          "tessvars.h"
+#include          "tessbox.h"
+#include          "secname.h"
+
+#define EXTERN
+
+EXTERN BOOL_VAR (fixsp_check_for_fp_noise_space, TRUE,
+"Try turning noise to space in fixed pitch");
+EXTERN BOOL_VAR (fixsp_fp_eval, TRUE, "Use alternate evaluation for fp");
+EXTERN BOOL_VAR (fixsp_noise_score_fixing, TRUE, "More sophisticated?");
+EXTERN INT_VAR (fixsp_non_noise_limit, 1,
+"How many non-noise blbs either side?");
+EXTERN double_VAR (fixsp_small_outlines_size, 0.28, "Small if lt xht x this");
+
+EXTERN BOOL_VAR (fixsp_ignore_punct, TRUE, "In uniform spacing calc");
+EXTERN BOOL_VAR (fixsp_numeric_fix, TRUE, "Try to deal with numeric punct");
+EXTERN BOOL_VAR (fixsp_prefer_joined_1s, TRUE, "Arbitrary boost");
+EXTERN BOOL_VAR (tessedit_test_uniform_wd_spacing, FALSE,
+"Limit context word spacing");
+EXTERN BOOL_VAR (tessedit_prefer_joined_punct, FALSE,
+"Reward punctation joins");
+EXTERN INT_VAR (fixsp_done_mode, 1, "What constitues done for spacing");
+EXTERN INT_VAR (debug_fix_space_level, 0, "Contextual fixspace debug");
+EXTERN STRING_VAR (numeric_punctuation, ".,",
+"Punct. chs expected WITHIN numbers");
+
+#define PERFECT_WERDS   999
+#define MAXSPACING      128      /*max expected spacing in pix */
+
+/*************************************************************************
+ * fix_fuzzy_spaces()
+ * Walk over the page finding sequences of words joined by fuzzy spaces. Extract
+ * them as a sublist, process the sublist to find the optimal arrangement of
+ * spaces then replace the sublist in the ROW_RES.
+ *************************************************************************/
+
+void fix_fuzzy_spaces(                               //find fuzzy words
+                      volatile ETEXT_DESC *monitor,  //progress monitor
+                      INT32 word_count,              //count of words in doc
+                      PAGE_RES *page_res) {
+  BLOCK_RES_IT block_res_it;     //iterators
+  ROW_RES_IT row_res_it;
+  WERD_RES_IT word_res_it_from;
+  WERD_RES_IT word_res_it_to;
+  WERD_RES *word_res;
+  WERD_RES_LIST fuzzy_space_words;
+  INT16 new_length;
+  BOOL8 prevent_null_wd_fixsp;   //DONT process blobless wds
+  INT32 word_index;              //current word
+
+  block_res_it.set_to_list (&page_res->block_res_list);
+  word_index = 0;
+  for (block_res_it.mark_cycle_pt ();
+  !block_res_it.cycled_list (); block_res_it.forward ()) {
+    row_res_it.set_to_list (&block_res_it.data ()->row_res_list);
+    for (row_res_it.mark_cycle_pt ();
+    !row_res_it.cycled_list (); row_res_it.forward ()) {
+      word_res_it_from.set_to_list (&row_res_it.data ()->word_res_list);
+      while (!word_res_it_from.at_last ()) {
+        word_res = word_res_it_from.data ();
+        while (!word_res_it_from.at_last () &&
+          !(word_res->combination ||
+          word_res_it_from.data_relative (1)->
+          word->flag (W_FUZZY_NON) ||
+          word_res_it_from.data_relative (1)->
+        word->flag (W_FUZZY_SP))) {
+          fix_sp_fp_word (word_res_it_from, row_res_it.data ()->row);
+          word_res = word_res_it_from.forward ();
+          word_index++;
+          if (monitor != NULL) {
+            monitor->ocr_alive = TRUE;
+            monitor->progress = 90 + 5 * word_index / word_count;
+          }
+        }
+
+        if (!word_res_it_from.at_last ()) {
+          word_res_it_to = word_res_it_from;
+          prevent_null_wd_fixsp =
+            word_res->word->gblob_list ()->empty ();
+          if (check_debug_pt (word_res, 60))
+            debug_fix_space_level.set_value (10);
+          word_res_it_to.forward ();
+          word_index++;
+          if (monitor != NULL) {
+            monitor->ocr_alive = TRUE;
+            monitor->progress = 90 + 5 * word_index / word_count;
+          }
+          while (!word_res_it_to.at_last () &&
+            (word_res_it_to.data_relative (1)->
+            word->flag (W_FUZZY_NON) ||
+            word_res_it_to.data_relative (1)->
+          word->flag (W_FUZZY_SP))) {
+            if (check_debug_pt (word_res, 60))
+              debug_fix_space_level.set_value (10);
+            if (word_res->word->gblob_list ()->empty ())
+              prevent_null_wd_fixsp = TRUE;
+            word_res = word_res_it_to.forward ();
+          }
+          if (check_debug_pt (word_res, 60))
+            debug_fix_space_level.set_value (10);
+          if (word_res->word->gblob_list ()->empty ())
+            prevent_null_wd_fixsp = TRUE;
+          if (prevent_null_wd_fixsp)
+            word_res_it_from = word_res_it_to;
+          else {
+            fuzzy_space_words.assign_to_sublist (&word_res_it_from,
+              &word_res_it_to);
+            fix_fuzzy_space_list (fuzzy_space_words,
+              row_res_it.data ()->row);
+            new_length = fuzzy_space_words.length ();
+            word_res_it_from.add_list_before (&fuzzy_space_words);
+            for (;
+              (!word_res_it_from.at_last () &&
+            (new_length > 0)); new_length--) {
+              word_res_it_from.forward ();
+            }
+          }
+          if (test_pt)
+            debug_fix_space_level.set_value (0);
+        }
+        fix_sp_fp_word (word_res_it_from, row_res_it.data ()->row);
+        //Last word in row
+      }
+    }
+  }
+}
+
+
+void fix_fuzzy_space_list(  //space explorer
+                          WERD_RES_LIST &best_perm,
+                          ROW *row) {
+  INT16 best_score;
+  WERD_RES_LIST current_perm;
+  INT16 current_score;
+  BOOL8 improved = FALSE;
+
+                                 //default score
+  best_score = eval_word_spacing (best_perm);
+
+  dump_words (best_perm, best_score, 1, improved);
+
+  if (best_score != PERFECT_WERDS)
+    initialise_search(best_perm, current_perm); 
+
+  while ((best_score != PERFECT_WERDS) && !current_perm.empty ()) {
+    match_current_words(current_perm, row); 
+    current_score = eval_word_spacing (current_perm);
+    dump_words (current_perm, current_score, 2, improved);
+    if (current_score > best_score) {
+      best_perm.clear ();
+      best_perm.deep_copy (&current_perm);
+      best_score = current_score;
+      improved = TRUE;
+    }
+    if (current_score < PERFECT_WERDS)
+      transform_to_next_perm(current_perm); 
+  }
+  dump_words (best_perm, best_score, 3, improved);
+}
+
+
+void initialise_search(WERD_RES_LIST &src_list, WERD_RES_LIST &new_list) { 
+  WERD_RES_IT src_it(&src_list); 
+  WERD_RES_IT new_it(&new_list); 
+  WERD_RES *src_wd;
+  WERD_RES *new_wd;
+
+  for (src_it.mark_cycle_pt (); !src_it.cycled_list (); src_it.forward ()) {
+    src_wd = src_it.data ();
+    if (!src_wd->combination) {
+      new_wd = new WERD_RES (*src_wd);
+      new_wd->combination = FALSE;
+      new_wd->part_of_combo = FALSE;
+      new_it.add_after_then_move (new_wd);
+    }
+  }
+}
+
+
+void match_current_words(WERD_RES_LIST &words, ROW *row) { 
+  WERD_RES_IT word_it(&words); 
+  WERD_RES *word;
+
+  for (word_it.mark_cycle_pt (); !word_it.cycled_list (); word_it.forward ()) {
+    word = word_it.data ();
+    if ((!word->part_of_combo) && (word->outword == NULL))
+      classify_word_pass2(word, row); 
+  }
+}
+
+
+/*************************************************************************
+ * eval_word_spacing()
+ * The basic measure is the number of characters in contextually confirmed
+ * words. (I.e the word is done)
+ * If all words are contextually confirmed the evaluation is deemed perfect.
+ *
+ * Some fiddles are done to handle "1"s as these are VERY frequent causes of
+ * fuzzy spaces. The problem with the basic measure is that "561 63" would score
+ * the same as "56163", though given our knowledge that the space is fuzzy, and
+ * that there is a "1" next to the fuzzy space, we need to ensure that "56163"
+ * is prefered.
+ *
+ * The solution is to NOT COUNT the score of any word which has a digit at one
+ * end and a "1Il" as the character the other side of the space.
+ *
+ * Conversly, any character next to a "1" within a word is counted as a positive
+ * score. Thus "561 63" would score 4 (3 chars in a numeric word plus 1 side of
+ * the "1" joined).  "56163" would score 7 - all chars in a numeric word + 2
+ * sides of a "1" joined.
+ *
+ * The joined 1 rule is applied to any word REGARDLESS of contextual
+ * confirmation.  Thus "PS7a71 3/7a" scores 1 (neither word is contexutally
+ * confirmed. The only score is from the joined 1. "PS7a713/7a" scores 2.
+ *
+ *************************************************************************/
+INT16 eval_word_spacing(WERD_RES_LIST &word_res_list) { 
+  WERD_RES_IT word_res_it(&word_res_list); 
+  INT16 total_score = 0;
+  INT16 word_count = 0;
+  INT16 done_word_count = 0;
+  INT16 word_len;
+  INT16 i;
+  WERD_RES *word;                //current word
+  INT16 prev_word_score = 0;
+  BOOL8 prev_word_done = FALSE;
+  BOOL8 prev_char_1 = FALSE;     //prev ch a "1/I/l"?
+  BOOL8 prev_char_digit = FALSE; //prev ch 2..9 or 0
+  BOOL8 current_char_1 = FALSE;
+  BOOL8 current_word_ok_so_far;
+  STRING punct_chars = "!\"`',.:;";
+  BOOL8 prev_char_punct = FALSE;
+  BOOL8 current_char_punct = FALSE;
+  BOOL8 word_done = FALSE;
+
+  do {
+    word = word_res_it.data ();
+    word_done = fixspace_thinks_word_done (word);
+    word_count++;
+    if (word->tess_failed) {
+      total_score += prev_word_score;
+      if (prev_word_done)
+        done_word_count++;
+      prev_word_score = 0;
+      prev_char_1 = FALSE;
+      prev_char_digit = FALSE;
+      prev_word_done = FALSE;
+    }
+    else {
+      /*
+        Can we add the prev word score and potentially count this word?
+        Yes IF it didnt end in a 1 when the first char of this word is a digit
+          AND it didnt end in a digit when the first char of this word is a 1
+      */
+      word_len = word->reject_map.length ();
+      current_word_ok_so_far = FALSE;
+      if (!((prev_char_1 &&
+        digit_or_numeric_punct (word,
+        word->best_choice->string ()[0])) ||
+        (prev_char_digit &&
+        ((word_done &&
+        (word->best_choice->string ()[0] == '1')) ||
+        (!word_done &&
+        STRING (conflict_set_I_l_1).contains (word->best_choice->
+      string ()[0])))))) {
+        total_score += prev_word_score;
+        if (prev_word_done)
+          done_word_count++;
+        current_word_ok_so_far = word_done;
+      }
+
+      if ((current_word_ok_so_far) &&
+        (!tessedit_test_uniform_wd_spacing ||
+        ((word->best_choice->permuter () == NUMBER_PERM) ||
+      uniformly_spaced (word)))) {
+        prev_word_done = TRUE;
+        prev_word_score = word_len;
+      }
+      else {
+        prev_word_done = FALSE;
+        prev_word_score = 0;
+      }
+
+      if (fixsp_prefer_joined_1s) {
+        /* Add 1 to total score for every joined 1 regardless of context and rejtn */
+
+        for (i = 0, prev_char_1 = FALSE; i < word_len; i++) {
+          current_char_1 = word->best_choice->string ()[i] == '1';
+          if (prev_char_1 || (current_char_1 && (i > 0)))
+            total_score++;
+          prev_char_1 = current_char_1;
+        }
+      }
+
+      /* Add 1 to total score for every joined punctuation regardless of context
+        and rejtn */
+      if (tessedit_prefer_joined_punct) {
+        for (i = 0, prev_char_punct = FALSE; i < word_len; i++) {
+          current_char_punct =
+            punct_chars.contains (word->best_choice->string ()[i]);
+          if (prev_char_punct || (current_char_punct && (i > 0)))
+            total_score++;
+          prev_char_punct = current_char_punct;
+        }
+      }
+      prev_char_digit = digit_or_numeric_punct (word,
+        word->best_choice->
+        string ()[word_len - 1]);
+      prev_char_1 =
+        ((word_done
+        && (word->best_choice->string ()[word_len - 1] == '1'))
+        || (!word_done
+        && STRING (conflict_set_I_l_1).contains (word->best_choice->
+        string ()[word_len -
+        1])));
+    }
+    /* Find next word */
+    do
+    word_res_it.forward ();
+    while (word_res_it.data ()->part_of_combo);
+  }
+  while (!word_res_it.at_first ());
+  total_score += prev_word_score;
+  if (prev_word_done)
+    done_word_count++;
+  if (done_word_count == word_count)
+    return PERFECT_WERDS;
+  else
+    return total_score;
+}
+
+
+BOOL8 digit_or_numeric_punct(WERD_RES *word, char ch) { 
+  return (isdigit (ch) ||
+    (fixsp_numeric_fix &&
+    (word->best_choice->permuter () == NUMBER_PERM) &&
+    STRING (numeric_punctuation).contains (ch)));
+}
+
+
+/*************************************************************************
+ * transform_to_next_perm()
+ * Examines the current word list to find the smallest word gap size. Then walks
+ * the word list closing any gaps of this size by either inserted new
+ * combination words, or extending existing ones.
+ *
+ * The routine COULD be limited to stop it building words longer than N blobs.
+ *
+ * If there are no more gaps then it DELETES the entire list and returns the
+ * empty list to cause termination.
+ *************************************************************************/
+void transform_to_next_perm(WERD_RES_LIST &words) { 
+  WERD_RES_IT word_it(&words); 
+  WERD_RES_IT prev_word_it(&words); 
+  WERD_RES *word;
+  WERD_RES *prev_word;
+  WERD_RES *combo;
+  WERD *copy_word;
+  INT16 prev_right = -1;
+  BOX box;
+  INT16 gap;
+  INT16 min_gap = MAX_INT16;
+
+  for (word_it.mark_cycle_pt (); !word_it.cycled_list (); word_it.forward ()) {
+    word = word_it.data ();
+    if (!word->part_of_combo) {
+      box = word->word->bounding_box ();
+      if (prev_right >= 0) {
+        gap = box.left () - prev_right;
+        if (gap < min_gap)
+          min_gap = gap;
+      }
+      prev_right = box.right ();
+    }
+  }
+  if (min_gap < MAX_INT16) {
+    prev_right = -1;             //back to start
+    word_it.set_to_list (&words);
+    for (;                       //cant use cycle pt due to inserted combos at start of list
+    (prev_right < 0) || !word_it.at_first (); word_it.forward ()) {
+      word = word_it.data ();
+      if (!word->part_of_combo) {
+        box = word->word->bounding_box ();
+        if (prev_right >= 0) {
+          gap = box.left () - prev_right;
+          if (gap <= min_gap) {
+            prev_word = prev_word_it.data ();
+            if (prev_word->combination)
+              combo = prev_word;
+            else {
+              /* Make a new combination and insert before the first word being joined */
+              copy_word = new WERD;
+              *copy_word = *(prev_word->word);
+              //deep copy
+              combo = new WERD_RES (copy_word);
+              combo->combination = TRUE;
+              prev_word->part_of_combo = TRUE;
+              prev_word_it.add_before_then_move (combo);
+            }
+            combo->word->set_flag (W_EOL, word->word->flag (W_EOL));
+            if (word->combination) {
+              combo->word->join_on (word->word);
+              //Move blbs to combo
+                                 //old combo no longer needed
+              delete word_it.extract ();
+            }
+            else {
+                                 //Cpy current wd to combo
+              combo->copy_on (word);
+              word->part_of_combo = TRUE;
+            }
+            combo->done = FALSE;
+            if (combo->outword != NULL) {
+              delete combo->outword;
+              delete combo->best_choice;
+              delete combo->raw_choice;
+              combo->outword = NULL;
+              combo->best_choice = NULL;
+              combo->raw_choice = NULL;
+            }
+          }
+          else
+                                 //catch up
+              prev_word_it = word_it;
+        }
+        prev_right = box.right ();
+      }
+    }
+  }
+  else
+    words.clear ();              //signal termination
+}
+
+
+void dump_words(WERD_RES_LIST &perm, INT16 score, INT16 mode, BOOL8 improved) { 
+  WERD_RES_IT word_res_it(&perm); 
+  static STRING initial_str;
+
+  if (debug_fix_space_level > 0) {
+    if (mode == 1) {
+      initial_str = "";
+      for (word_res_it.mark_cycle_pt ();
+      !word_res_it.cycled_list (); word_res_it.forward ()) {
+        if (!word_res_it.data ()->part_of_combo) {
+          initial_str += word_res_it.data ()->best_choice->string ();
+          initial_str += ' ';
+        }
+      }
+    }
+
+    #ifndef SECURE_NAMES
+    if (debug_fix_space_level > 1) {
+      switch (mode) {
+        case 1:
+          tprintf ("EXTRACTED (%d): \"", score);
+          break;
+        case 2:
+          tprintf ("TESTED (%d): \"", score);
+          break;
+        case 3:
+          tprintf ("RETURNED (%d): \"", score);
+          break;
+      }
+
+      for (word_res_it.mark_cycle_pt ();
+      !word_res_it.cycled_list (); word_res_it.forward ()) {
+        if (!word_res_it.data ()->part_of_combo)
+          tprintf ("%s/%1d ",
+            word_res_it.data ()->best_choice->string ().
+            string (),
+            (int) word_res_it.data ()->best_choice->permuter ());
+      }
+      tprintf ("\"\n");
+    }
+    else if (improved) {
+      tprintf ("FIX SPACING \"%s\" => \"", initial_str.string ());
+      for (word_res_it.mark_cycle_pt ();
+      !word_res_it.cycled_list (); word_res_it.forward ()) {
+        if (!word_res_it.data ()->part_of_combo)
+          tprintf ("%s/%1d ",
+            word_res_it.data ()->best_choice->string ().
+            string (),
+            (int) word_res_it.data ()->best_choice->permuter ());
+      }
+      tprintf ("\"\n");
+    }
+    #endif
+  }
+}
+
+
+/*************************************************************************
+ * uniformly_spaced()
+ * Return true if one of the following are true:
+ *    - All inter-char gaps are the same width
+ *	- The largest gap is no larger than twice the mean/median of the others
+ *	- The largest gap is < 64/5 = 13 and all others are <= 0
+ * **** REMEMBER - WE'RE NOW WORKING WITH A BLN WERD !!!
+ *************************************************************************/
+BOOL8 uniformly_spaced(  //sensible word
+                       WERD_RES *word) {
+  PBLOB_IT blob_it;
+  BOX box;
+  INT16 prev_right = -MAX_INT16;
+  INT16 gap;
+  INT16 max_gap = -MAX_INT16;
+  INT16 max_gap_count = 0;
+  STATS gap_stats (0, MAXSPACING);
+  BOOL8 result;
+  const ROW *row = word->denorm.row ();
+  float max_non_space;
+  float normalised_max_nonspace;
+  INT16 i = 0;
+  STRING punct_chars = "\"`',.:;";
+
+  blob_it.set_to_list (word->outword->blob_list ());
+
+  for (blob_it.mark_cycle_pt (); !blob_it.cycled_list (); blob_it.forward ()) {
+    box = blob_it.data ()->bounding_box ();
+    if ((prev_right > -MAX_INT16) &&
+      (!fixsp_ignore_punct ||
+      (!punct_chars.contains (word->best_choice->string ()[i - 1]) &&
+    !punct_chars.contains (word->best_choice->string ()[i])))) {
+      gap = box.left () - prev_right;
+      if (gap < max_gap)
+        gap_stats.add (gap, 1);
+      else if (gap == max_gap)
+        max_gap_count++;
+      else {
+        if (max_gap_count > 0)
+          gap_stats.add (max_gap, max_gap_count);
+        max_gap = gap;
+        max_gap_count = 1;
+      }
+    }
+    prev_right = box.right ();
+    i++;
+  }
+
+  max_non_space = (row->space () + 3 * row->kern ()) / 4;
+  normalised_max_nonspace = max_non_space * bln_x_height / row->x_height ();
+
+  result = ((gap_stats.get_total () == 0) ||
+    (max_gap <= normalised_max_nonspace) ||
+    ((gap_stats.get_total () > 2) &&
+    (max_gap <= 2 * gap_stats.median ())) ||
+    ((gap_stats.get_total () <= 2) &&
+    (max_gap <= 2 * gap_stats.mean ())));
+  #ifndef SECURE_NAMES
+  if ((debug_fix_space_level > 1)) {
+    if (result)
+      tprintf
+        ("ACCEPT SPACING FOR: \"%s\" norm_maxnon = %f max=%d maxcount=%d total=%d mean=%f median=%f\n",
+        word->best_choice->string ().string (), normalised_max_nonspace,
+        max_gap, max_gap_count, gap_stats.get_total (), gap_stats.mean (),
+        gap_stats.median ());
+    else
+      tprintf
+        ("REJECT SPACING FOR: \"%s\" norm_maxnon = %f max=%d maxcount=%d total=%d mean=%f median=%f\n",
+        word->best_choice->string ().string (), normalised_max_nonspace,
+        max_gap, max_gap_count, gap_stats.get_total (), gap_stats.mean (),
+        gap_stats.median ());
+  }
+  #endif
+
+  return result;
+}
+
+
+BOOL8 fixspace_thinks_word_done(WERD_RES *word) { 
+  if (word->done)
+    return TRUE;
+
+  /*
+    Use all the standard pass 2 conditions for mode 5 in set_done() in
+    reject.c BUT DONT REJECT IF THE WERD IS AMBIGUOUS - FOR SPACING WE DONT
+    CARE WHETHER WE HAVE of/at on/an etc.
+  */
+  if ((fixsp_done_mode > 0) &&
+    (word->tess_accepted ||
+    ((fixsp_done_mode == 2) &&
+    (word->reject_map.reject_count () == 0)) ||
+    (fixsp_done_mode == 3)) &&
+    (strchr (word->best_choice->string ().string (), ' ') == NULL) &&
+    ((word->best_choice->permuter () == SYSTEM_DAWG_PERM) ||
+    (word->best_choice->permuter () == FREQ_DAWG_PERM) ||
+    (word->best_choice->permuter () == USER_DAWG_PERM) ||
+    (word->best_choice->permuter () == NUMBER_PERM)))
+    return TRUE;
+  else
+    return FALSE;
+}
+
+
+/*************************************************************************
+ * fix_sp_fp_word()
+ * Test the current word to see if it can be split by deleting noise blobs. If
+ * so, do the buisiness.
+ * Return with the iterator pointing to the same place if the word is unchanged,
+ * or the last of the replacement words.
+ *************************************************************************/
+void fix_sp_fp_word(WERD_RES_IT &word_res_it, ROW *row) { 
+  WERD_RES *word_res;
+  WERD_RES_LIST sub_word_list;
+  WERD_RES_IT sub_word_list_it(&sub_word_list); 
+  INT16 blob_index;
+  INT16 new_length;
+  float junk;
+
+  word_res = word_res_it.data ();
+  if (!fixsp_check_for_fp_noise_space ||
+    word_res->word->flag (W_REP_CHAR) ||
+    word_res->combination ||
+    word_res->part_of_combo || !word_res->word->flag (W_DONT_CHOP))
+    return;
+
+  blob_index = worst_noise_blob (word_res, &junk);
+  if (blob_index < 0)
+    return;
+
+  #ifndef SECURE_NAMES
+  if (debug_fix_space_level > 1) {
+    tprintf ("FP fixspace working on \"%s\"\n",
+      word_res->best_choice->string ().string ());
+  }
+  #endif
+  gblob_sort_list ((PBLOB_LIST *) word_res->word->rej_cblob_list (), FALSE);
+  sub_word_list_it.add_after_stay_put (word_res_it.extract ());
+  fix_noisy_space_list(sub_word_list, row); 
+  new_length = sub_word_list.length ();
+  word_res_it.add_list_before (&sub_word_list);
+  for (; (!word_res_it.at_last () && (new_length > 1)); new_length--) {
+    word_res_it.forward ();
+  }
+}
+
+
+void fix_noisy_space_list(WERD_RES_LIST &best_perm, ROW *row) { 
+  INT16 best_score;
+  WERD_RES_IT best_perm_it(&best_perm); 
+  WERD_RES_LIST current_perm;
+  WERD_RES_IT current_perm_it(&current_perm); 
+  WERD_RES *old_word_res;
+  WERD_RES *new_word_res;
+  INT16 current_score;
+  BOOL8 improved = FALSE;
+
+                                 //default score
+  best_score = fp_eval_word_spacing (best_perm);
+
+  dump_words (best_perm, best_score, 1, improved);
+
+  new_word_res = new WERD_RES;
+  old_word_res = best_perm_it.data ();
+                                 //Kludge to force deep copy
+  old_word_res->combination = TRUE;
+  *new_word_res = *old_word_res; //deep copy
+                                 //Undo kludge
+  old_word_res->combination = FALSE;
+                                 //Undo kludge
+  new_word_res->combination = FALSE;
+  current_perm_it.add_to_end (new_word_res);
+
+  break_noisiest_blob_word(current_perm); 
+
+  while ((best_score != PERFECT_WERDS) && !current_perm.empty ()) {
+    match_current_words(current_perm, row); 
+    current_score = fp_eval_word_spacing (current_perm);
+    dump_words (current_perm, current_score, 2, improved);
+    if (current_score > best_score) {
+      best_perm.clear ();
+      best_perm.deep_copy (&current_perm);
+      best_score = current_score;
+      improved = TRUE;
+    }
+    if (current_score < PERFECT_WERDS)
+      break_noisiest_blob_word(current_perm); 
+  }
+  dump_words (best_perm, best_score, 3, improved);
+}
+
+
+/*************************************************************************
+ * break_noisiest_blob_word()
+ * Find the word with the blob which looks like the worst noise.
+ * Break the word into two, deleting the noise blob.
+ *************************************************************************/
+void break_noisiest_blob_word(WERD_RES_LIST &words) { 
+  WERD_RES_IT word_it(&words); 
+  WERD_RES_IT worst_word_it;
+  float worst_noise_score = 9999;
+  int worst_blob_index = -1;     //noisiest blb of noisiest wd
+  int blob_index;                //of wds noisiest blb
+  float noise_score;             //of wds noisiest blb
+  WERD_RES *word_res;
+  C_BLOB_IT blob_it;
+  C_BLOB_IT rej_cblob_it;
+  C_BLOB_LIST new_blob_list;
+  C_BLOB_IT new_blob_it;
+  C_BLOB_IT new_rej_cblob_it;
+  WERD *new_word;
+  INT16 start_of_noise_blob;
+  INT16 i;
+
+  for (word_it.mark_cycle_pt (); !word_it.cycled_list (); word_it.forward ()) {
+    blob_index = worst_noise_blob (word_it.data (), &noise_score);
+    if ((blob_index > -1) && (worst_noise_score > noise_score)) {
+      worst_noise_score = noise_score;
+      worst_blob_index = blob_index;
+      worst_word_it = word_it;
+    }
+  }
+  if (worst_blob_index < 0) {
+    words.clear ();              //signal termination
+    return;
+  }
+
+  /* Now split the worst_word_it */
+
+  word_res = worst_word_it.data ();
+
+  /* Move blobs before noise blob to a new bloblist */
+
+  new_blob_it.set_to_list (&new_blob_list);
+  blob_it.set_to_list (word_res->word->cblob_list ());
+  for (i = 0; i < worst_blob_index; i++, blob_it.forward ()) {
+    new_blob_it.add_after_then_move (blob_it.extract ());
+  }
+  start_of_noise_blob = blob_it.data ()->bounding_box ().left ();
+  delete blob_it.extract ();     //throw out noise blb
+
+  new_word = new WERD (&new_blob_list, word_res->word);
+  new_word->set_flag (W_EOL, FALSE);
+  word_res->word->set_flag (W_BOL, FALSE);
+  word_res->word->set_blanks (1);//After break
+
+  new_rej_cblob_it.set_to_list (new_word->rej_cblob_list ());
+  rej_cblob_it.set_to_list (word_res->word->rej_cblob_list ());
+  for (;
+    (!rej_cblob_it.empty () &&
+    (rej_cblob_it.data ()->bounding_box ().left () <
+  start_of_noise_blob)); rej_cblob_it.forward ()) {
+    new_rej_cblob_it.add_after_then_move (rej_cblob_it.extract ());
+  }
+
+  worst_word_it.add_before_then_move (new WERD_RES (new_word));
+
+  word_res->done = FALSE;
+  if (word_res->outword != NULL) {
+    delete word_res->outword;
+    delete word_res->best_choice;
+    delete word_res->raw_choice;
+    word_res->outword = NULL;
+    word_res->best_choice = NULL;
+    word_res->raw_choice = NULL;
+  }
+}
+
+
+INT16 worst_noise_blob(WERD_RES *word_res, float *worst_noise_score) { 
+  PBLOB_IT blob_it;
+  INT16 blob_count;
+  float noise_score[512];
+  int i;
+  int min_noise_blob;            //1st contender
+  int max_noise_blob;            //last contender
+  int non_noise_count;
+  int worst_noise_blob;          //Worst blob
+  float small_limit = bln_x_height * fixsp_small_outlines_size;
+  float non_noise_limit = bln_x_height * 0.8;
+
+  blob_it.set_to_list (word_res->outword->blob_list ());
+  //normalised
+  blob_count = blob_it.length ();
+  ASSERT_HOST (blob_count <= 512);
+  if (blob_count < 5)
+    return -1;                   //too short to split
+  /* Get the noise scores for all blobs */
+
+  #ifndef SECURE_NAMES
+  if (debug_fix_space_level > 5)
+    tprintf ("FP fixspace Noise metrics for \"%s\": ",
+      word_res->best_choice->string ().string ());
+  #endif
+
+  for (i = 0; i < blob_count; i++, blob_it.forward ()) {
+    if (word_res->reject_map[i].accepted ())
+      noise_score[i] = non_noise_limit;
+    else
+      noise_score[i] = blob_noise_score (blob_it.data ());
+
+    if (debug_fix_space_level > 5)
+      tprintf ("%1.1f ", noise_score[i]);
+  }
+  if (debug_fix_space_level > 5)
+    tprintf ("\n");
+
+  /* Now find the worst one which is far enough away from the end of the word */
+
+  non_noise_count = 0;
+  for (i = 0;
+  (i < blob_count) && (non_noise_count < fixsp_non_noise_limit); i++) {
+    if (noise_score[i] >= non_noise_limit)
+      non_noise_count++;
+  }
+  if (non_noise_count < fixsp_non_noise_limit)
+    return -1;
+  min_noise_blob = i;
+
+  non_noise_count = 0;
+  for (i = blob_count - 1;
+  (i >= 0) && (non_noise_count < fixsp_non_noise_limit); i--) {
+    if (noise_score[i] >= non_noise_limit)
+      non_noise_count++;
+  }
+  if (non_noise_count < fixsp_non_noise_limit)
+    return -1;
+  max_noise_blob = i;
+
+  if (min_noise_blob > max_noise_blob)
+    return -1;
+
+  *worst_noise_score = small_limit;
+  worst_noise_blob = -1;
+  for (i = min_noise_blob; i <= max_noise_blob; i++) {
+    if (noise_score[i] < *worst_noise_score) {
+      worst_noise_blob = i;
+      *worst_noise_score = noise_score[i];
+    }
+  }
+  return worst_noise_blob;
+}
+
+
+float blob_noise_score(PBLOB *blob) { 
+  OUTLINE_IT outline_it;
+  BOX box;                       //BB of outline
+  INT16 outline_count = 0;
+  INT16 max_dimension;
+  INT16 largest_outline_dimension = 0;
+
+  outline_it.set_to_list (blob->out_list ());
+  for (outline_it.mark_cycle_pt ();
+  !outline_it.cycled_list (); outline_it.forward ()) {
+    outline_count++;
+    box = outline_it.data ()->bounding_box ();
+    if (box.height () > box.width ())
+      max_dimension = box.height ();
+    else
+      max_dimension = box.width ();
+
+    if (largest_outline_dimension < max_dimension)
+      largest_outline_dimension = max_dimension;
+  }
+
+  if (fixsp_noise_score_fixing) {
+    if (outline_count > 5)
+                                 //penalise LOTS of blobs
+      largest_outline_dimension *= 2;
+
+    box = blob->bounding_box ();
+
+    if ((box.bottom () > bln_baseline_offset * 4) ||
+      (box.top () < bln_baseline_offset / 2))
+                                 //Lax blob is if high or low
+      largest_outline_dimension /= 2;
+  }
+  return largest_outline_dimension;
+}
+
+
+void fixspace_dbg(WERD_RES *word) { 
+  BOX box = word->word->bounding_box ();
+  BOOL8 show_map_detail = FALSE;
+  INT16 i;
+
+  box.print ();
+  #ifndef SECURE_NAMES
+  tprintf (" \"%s\" ", word->best_choice->string ().string ());
+  tprintf ("Blob count: %d (word); %d/%d (outword)\n",
+    word->word->gblob_list ()->length (),
+    word->outword->gblob_list ()->length (),
+    word->outword->rej_blob_list ()->length ());
+  word->reject_map.print (debug_fp);
+  tprintf ("\n");
+  if (show_map_detail) {
+    tprintf ("\"%s\"\n", word->best_choice->string ().string ());
+    for (i = 0; word->best_choice->string ()[i] != '\0'; i++) {
+      tprintf ("**** \"%c\" ****\n", word->best_choice->string ()[i]);
+      word->reject_map[i].full_print (debug_fp);
+    }
+  }
+
+  tprintf ("Tess Accepted: %s\n", word->tess_accepted ? "TRUE" : "FALSE");
+  tprintf ("Done flag: %s\n\n", word->done ? "TRUE" : "FALSE");
+  #endif
+}
+
+
+/*************************************************************************
+ * fp_eval_word_spacing()
+ * Evaluation function for fixed pitch word lists.
+ *
+ * Basically, count the number of "nice" characters - those which are in tess
+ * acceptable words or in dict words and are not rejected.
+ * Penalise any potential noise chars
+ *************************************************************************/
+
+INT16 fp_eval_word_spacing(WERD_RES_LIST &word_res_list) { 
+  WERD_RES_IT word_it(&word_res_list); 
+  WERD_RES *word;
+  PBLOB_IT blob_it;
+  INT16 word_length;
+  INT16 score = 0;
+  INT16 i;
+  const char *chs;
+  float small_limit = bln_x_height * fixsp_small_outlines_size;
+
+  if (!fixsp_fp_eval)
+    return (eval_word_spacing (word_res_list));
+
+  for (word_it.mark_cycle_pt (); !word_it.cycled_list (); word_it.forward ()) {
+    word = word_it.data ();
+    word_length = word->reject_map.length ();
+    chs = word->best_choice->string ().string ();
+    if ((word->done ||
+      word->tess_accepted) ||
+      (word->best_choice->permuter () == SYSTEM_DAWG_PERM) ||
+      (word->best_choice->permuter () == FREQ_DAWG_PERM) ||
+      (word->best_choice->permuter () == USER_DAWG_PERM) ||
+    (safe_dict_word (chs) > 0)) {
+      blob_it.set_to_list (word->outword->blob_list ());
+      for (i = 0; i < word_length; i++, blob_it.forward ()) {
+        if ((chs[i] == ' ') ||
+          (blob_noise_score (blob_it.data ()) < small_limit))
+          score -= 1;            //penalise possibly erroneous non-space
+
+        else if (word->reject_map[i].accepted ())
+          score++;
+      }
+    }
+  }
+  if (score < 0)
+    score = 0;
+  return score;
+}

ccmain/fixspace.h

@@ -0,0 +1,72 @@
+/******************************************************************
+ * File:        fixspace.h  (Formerly fixspace.h)
+ * Description: Implements a pass over the page res, exploring the alternative
+ *					spacing possibilities, trying to use context to improve the
+          word spacing
+* Author:		Phil Cheatle
+* Created:		Thu Oct 21 11:38:43 BST 1993
+*
+* (C) Copyright 1993, Hewlett-Packard Ltd.
+** Licensed under the Apache License, Version 2.0 (the "License");
+** you may not use this file except in compliance with the License.
+** You may obtain a copy of the License at
+** http://www.apache.org/licenses/LICENSE-2.0
+** Unless required by applicable law or agreed to in writing, software
+** distributed under the License is distributed on an "AS IS" BASIS,
+** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+** See the License for the specific language governing permissions and
+** limitations under the License.
+*
+**********************************************************************/
+
+#ifndef           FIXSPACE_H
+#define           FIXSPACE_H
+
+#include          "pageres.h"
+#include          "varable.h"
+#include          "ocrclass.h"
+#include          "notdll.h"
+
+extern BOOL_VAR_H (fixsp_check_for_fp_noise_space, TRUE,
+"Try turning noise to space in fixed pitch");
+extern BOOL_VAR_H (fixsp_fp_eval, TRUE, "Use alternate evaluation for fp");
+extern BOOL_VAR_H (fixsp_noise_score_fixing, TRUE, "More sophisticated?");
+extern INT_VAR_H (fixsp_non_noise_limit, 1,
+"How many non-noise blbs either side?");
+extern double_VAR_H (fixsp_small_outlines_size, 0.28,
+"Small if lt xht x this");
+extern BOOL_VAR_H (fixsp_ignore_punct, TRUE, "In uniform spacing calc");
+extern BOOL_VAR_H (fixsp_numeric_fix, TRUE, "Try to deal with numeric punct");
+extern BOOL_VAR_H (fixsp_prefer_joined_1s, TRUE, "Arbitrary boost");
+extern BOOL_VAR_H (tessedit_test_uniform_wd_spacing, FALSE,
+"Limit context word spacing");
+extern BOOL_VAR_H (tessedit_prefer_joined_punct, FALSE,
+"Reward punctation joins");
+extern INT_VAR_H (fixsp_done_mode, 1, "What constitues done for spacing");
+extern INT_VAR_H (debug_fix_space_level, 0, "Contextual fixspace debug");
+extern STRING_VAR_H (numeric_punctuation, ".,",
+"Punct. chs expected WITHIN numbers");
+void fix_fuzzy_spaces(                               //find fuzzy words
+                      volatile ETEXT_DESC *monitor,  //progress monitor
+                      INT32 word_count,              //count of words in doc
+                      PAGE_RES *page_res);
+void fix_fuzzy_space_list(  //space explorer
+                          WERD_RES_LIST &best_perm,
+                          ROW *row);
+void initialise_search(WERD_RES_LIST &src_list, WERD_RES_LIST &new_list); 
+void match_current_words(WERD_RES_LIST &words, ROW *row); 
+INT16 eval_word_spacing(WERD_RES_LIST &word_res_list); 
+BOOL8 digit_or_numeric_punct(WERD_RES *word, char ch); 
+void transform_to_next_perm(WERD_RES_LIST &words); 
+void dump_words(WERD_RES_LIST &perm, INT16 score, INT16 mode, BOOL8 improved); 
+BOOL8 uniformly_spaced(  //sensible word
+                       WERD_RES *word);
+BOOL8 fixspace_thinks_word_done(WERD_RES *word); 
+void fix_sp_fp_word(WERD_RES_IT &word_res_it, ROW *row); 
+void fix_noisy_space_list(WERD_RES_LIST &best_perm, ROW *row); 
+void break_noisiest_blob_word(WERD_RES_LIST &words); 
+INT16 worst_noise_blob(WERD_RES *word_res, float *worst_noise_score); 
+float blob_noise_score(PBLOB *blob); 
+void fixspace_dbg(WERD_RES *word); 
+INT16 fp_eval_word_spacing(WERD_RES_LIST &word_res_list); 
+#endif

ccmain/fixxht.cpp

@@ -0,0 +1,790 @@
+/**********************************************************************
+ * File:        fixxht.cpp  (Formerly fixxht.c)
+ * Description: Improve x_ht and look out for case inconsistencies
+ * Author:		Phil Cheatle
+ * Created:		Thu Aug  5 14:11:08 BST 1993
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          <string.h>
+#include          <ctype.h>
+#include          "varable.h"
+#include          "tessvars.h"
+#include          "control.h"
+#include          "reject.h"
+#include          "fixxht.h"
+#include          "secname.h"
+
+#define EXTERN
+
+EXTERN double_VAR (x_ht_fraction_of_caps_ht, 0.7,
+"Fract of cps ht est of xht");
+EXTERN double_VAR (x_ht_variation, 0.35,
+"Err band as fract of caps/xht dist");
+EXTERN double_VAR (x_ht_sub_variation, 0.5,
+"Err band as fract of caps/xht dist");
+EXTERN BOOL_VAR (rej_trial_ambigs, TRUE,
+"reject x-ht ambigs when under trial");
+EXTERN BOOL_VAR (x_ht_conservative_ambigs, FALSE,
+"Dont rely on ambigs + maxht");
+EXTERN BOOL_VAR (x_ht_check_est, TRUE, "Cross check estimates");
+EXTERN BOOL_VAR (x_ht_case_flip, FALSE, "Flip or reject suspect case");
+EXTERN BOOL_VAR (x_ht_include_dodgy_blobs, TRUE,
+"Include blobs with possible noise?");
+EXTERN BOOL_VAR (x_ht_limit_flip_trials, TRUE,
+"Dont do trial flips when ambigs are close to xht?");
+EXTERN BOOL_VAR (rej_use_check_block_occ, TRUE,
+"Analyse rejection behaviour");
+
+EXTERN STRING_VAR (chs_non_ambig_caps_ht,
+"!#$%&()/12346789?ABDEFGHIKLNQRT[]\\bdfhkl",
+"Reliable ascenders");
+EXTERN STRING_VAR (chs_x_ht, "acegmnopqrsuvwxyz", "X height chars");
+EXTERN STRING_VAR (chs_non_ambig_x_ht, "aenqr", "reliable X height chars");
+EXTERN STRING_VAR (chs_ambig_caps_x, "cCmMoO05sSuUvVwWxXzZ",
+"X ht or caps ht chars");
+EXTERN STRING_VAR (chs_bl_ambig_caps_x, "pPyY", " Caps or descender ambigs");
+
+/* The following arent used in this module but are used in applybox.c */
+EXTERN STRING_VAR (chs_caps_ht,
+"!#$%&()/0123456789?ABCDEFGHIJKLMNOPQRSTUVWXYZ[]\\bdfhkl{|}",
+"Ascender chars");
+EXTERN STRING_VAR (chs_desc, "gjpqy", "Descender chars");
+EXTERN STRING_VAR (chs_non_ambig_bl,
+"!#$%&01246789?ABCDEFGHIKLMNORSTUVWXYZabcdehiklmnorstuvwxz",
+"Reliable baseline chars");
+EXTERN STRING_VAR (chs_odd_top, "ijt", "Chars with funny ascender region");
+EXTERN STRING_VAR (chs_odd_bot, "()35JQ[]\\/{}|", "Chars with funny base");
+
+/* The following arent used but are defined for completeness */
+EXTERN STRING_VAR (chs_bl,
+"!#$%&()/01246789?ABCDEFGHIJKLMNOPRSTUVWXYZ[]\\abcdefhiklmnorstuvwxz{}",
+"Baseline chars");
+EXTERN STRING_VAR (chs_non_ambig_desc, "gq", "Reliable descender chars");
+
+/*************************************************************************
+ * re_estimate_x_ht()
+ *
+ * Walk the blobs in the word together with the text string and reject map.
+ * NOTE: All evaluation is done on the baseline normalised word. This is so that
+ * the BOX class can be used (integer). The reasons for this are:
+ *   a) We must use the outword - ie the Tess result
+ *   b) The outword is always converted to integer representation as that is how
+ *      Tess works
+ *   c) We would like to use the BOX class, cos its there - this is integer
+ *      precision.
+ *   d) If we de-normed the outword we would get rounding errors and would find
+ *      that integers are too imprecise (x-height around 15 pixels instead of a
+ *      scale of 128 in bln form.
+ *   CONVINCED?
+ *
+ * A) Try to re-estimatate x-ht and caps ht from confirmed pts in word.
+ *
+ *    FOR each non reject blob
+ *       IF char is baseline posn ambiguous
+ *			Remove ambiguity by comparing its posn with respect to baseline.
+ *		IF char is a confirmed x-ht char
+ *			Add x-ht posn to confirmed_x_ht pts for word
+ *    IF char is a confirmed caps-ht char
+ *			Add blob_ht to caps ht pts for word
+ *
+ *    IF Std Dev of caps hts < 2  (AND # samples > 0)
+ *		Use mean as caps ht estimate (Dont use median as we can expect a
+ *			fair variation between the heights of the NON_AMBIG_CAPS_HT_CHS)
+ *    IF Std Dev of caps hts >= 2  (AND # samples > 0)
+ *			Suspect small caps font.
+ *			Look for 2 clusters,	each with Std Dev < 2.
+ *			IF 2 clusters found
+ *			Pick the smaller median as the caps ht estimate of the smallcaps.
+ *
+ *    IF failed to estimate a caps ht
+ *       Use the median caps ht if there is one,
+ *		ELSE use the caps ht estimate of the previous word. NO!!!
+ *
+ *
+ *    IF there are confirmed x-height chars
+ *			Estimate confirmed x-height as the median value
+ *    ELSE IF there is a confirmed caps ht
+ *			Estimate confirmed x-height as a fraction of confirmed caps ht value
+ *		ELSE
+ *			Use the value for the previous word or the row value if this is the
+ *			first word in the block. NO!!!
+ *
+ * B) Add in case ambiguous blobs based on confirmed x-ht/caps ht, changing case
+ *    as necessary. Reestimate caps ht and x-ht as in A, using the extended
+ *    clusters.
+ *
+ * C) If word contains rejects, and x-ht estimate significantly differs from
+ *    original estimate, return TRUE so that the word can be rematched
+ *************************************************************************/
+
+void re_estimate_x_ht(                     //improve for 1 word
+                      WERD_RES *word_res,  //word to do
+                      float *trial_x_ht    //new match value
+                     ) {
+  PBLOB_IT blob_it;
+  INT16 blob_ht_above_baseline;
+
+  const char *word_str;
+  INT16 i;
+
+  STATS all_blobs_ht (0, 300);   //every blob in word
+  STATS x_ht (0, 300);           //confirmed pts in wd
+  STATS caps_ht (0, 300);        //confirmed pts in wd
+  STATS case_ambig (0, 300);     //lower case ambigs
+
+  INT16 rej_blobs_count = 0;
+  INT16 rej_blobs_max_height = 0;
+  INT32 rej_blobs_max_area = 0;
+  float x_ht_ok_variation;
+  float max_blob_ht;
+  float marginally_above_x_ht;
+
+  BOX blob_box;                  //blob bounding box
+  float est_x_ht = 0.0;          //word estimate
+  float est_caps_ht = 0.0;       //word estimate
+                                 //based on hard data?
+  BOOL8 est_caps_ht_certain = FALSE;
+  BOOL8 est_x_ht_certain = FALSE;//based on hard data?
+  BOOL8 trial = FALSE;           //Sepeculative values?
+  BOOL8 no_comment = FALSE;      //No change in xht
+  float ambig_lc_x_est;
+  float ambig_uc_caps_est;
+  INT16 x_ht_ambigs = 0;
+  INT16 caps_ht_ambigs = 0;
+
+  /* Calculate default variation of blob x_ht from bln x_ht for bln word */
+  x_ht_ok_variation =
+    (bln_x_height / x_ht_fraction_of_caps_ht - bln_x_height) * x_ht_variation;
+
+  word_str = word_res->best_choice->string ().string ();
+  /*
+    Cycle blobs, allocating to one of the stats sets when possible.
+  */
+  blob_it.set_to_list (word_res->outword->blob_list ());
+  for (blob_it.mark_cycle_pt (), i = 0;
+  !blob_it.cycled_list (); blob_it.forward (), i++) {
+    if (!dodgy_blob (blob_it.data ())) {
+      blob_box = blob_it.data ()->bounding_box ();
+      blob_ht_above_baseline = blob_box.top () - bln_baseline_offset;
+      all_blobs_ht.add (blob_ht_above_baseline, 1);
+
+      if (word_res->reject_map[i].rejected ()) {
+        rej_blobs_count++;
+        if (blob_box.height () > rej_blobs_max_height)
+          rej_blobs_max_height = blob_box.height ();
+        if (blob_box.area () > rej_blobs_max_area)
+          rej_blobs_max_area = blob_box.area ();
+      }
+      else {
+        if (STRING (chs_non_ambig_x_ht).contains (word_str[i]))
+          x_ht.add (blob_ht_above_baseline, 1);
+
+        if (STRING (chs_non_ambig_caps_ht).contains (word_str[i]))
+          caps_ht.add (blob_ht_above_baseline, 1);
+
+        if (STRING (chs_ambig_caps_x).contains (word_str[i])) {
+          case_ambig.add (blob_ht_above_baseline, 1);
+          if (STRING (chs_x_ht).contains (word_str[i]))
+            x_ht_ambigs++;
+          else
+            caps_ht_ambigs++;
+        }
+
+        if (STRING (chs_bl_ambig_caps_x).contains (word_str[i])) {
+          if (STRING (chs_x_ht).contains (word_str[i])) {
+            /* confirm x_height provided > 15% total height below baseline */
+            if ((bln_baseline_offset - blob_box.bottom ()) /
+              (float) blob_box.height () > 0.15)
+              x_ht.add (blob_ht_above_baseline, 1);
+          }
+          else {
+            /* confirm caps_height provided < 5% total height below baseline */
+            if ((bln_baseline_offset - blob_box.bottom ()) /
+              (float) blob_box.height () < 0.05)
+              caps_ht.add (blob_ht_above_baseline, 1);
+          }
+        }
+      }
+    }
+  }
+  est_caps_ht = estimate_from_stats (caps_ht);
+  est_x_ht = estimate_from_stats (x_ht);
+  est_ambigs(word_res, case_ambig, &ambig_lc_x_est, &ambig_uc_caps_est); 
+  max_blob_ht = all_blobs_ht.ile (0.9999);
+
+  #ifndef SECURE_NAMES
+  if (debug_x_ht_level >= 20) {
+    tprintf ("Mode20:A: %s ", word_str);
+    word_res->reject_map.print (debug_fp);
+    tprintf (" XHT:%f CAP:%f MAX:%f AMBIG X:%f CAP:%f\n",
+      est_x_ht, est_caps_ht, max_blob_ht,
+      ambig_lc_x_est, ambig_uc_caps_est);
+  }
+  #endif
+  if (!x_ht_conservative_ambigs &&
+    (ambig_lc_x_est > 0) &&
+    (ambig_lc_x_est == ambig_uc_caps_est) &&
+  (max_blob_ht > ambig_lc_x_est + x_ht_ok_variation)) {
+                                 //may be zero but believe xht
+    ambig_uc_caps_est = est_caps_ht;
+    #ifndef SECURE_NAMES
+    if (debug_x_ht_level >= 20)
+      tprintf ("Mode20:B: Fiddle ambig_uc_caps_est to %f\n",
+        ambig_lc_x_est);
+    #endif
+  }
+
+  /* Now make some estimates */
+
+  if ((est_x_ht > 0) ||
+    (est_caps_ht > 0) ||
+  ((ambig_lc_x_est > 0) && (ambig_lc_x_est != ambig_uc_caps_est))) {
+    /* There is some sensible data to go on so make the most of it. */
+    if (debug_x_ht_level >= 20)
+      tprintf ("Mode20:C: Sensible Data\n", ambig_lc_x_est);
+    if (est_x_ht > 0) {
+      est_x_ht_certain = TRUE;
+      if (est_caps_ht == 0) {
+        if ((ambig_uc_caps_est > ambig_lc_x_est) &&
+          (ambig_uc_caps_est > est_x_ht + x_ht_ok_variation))
+          est_caps_ht = ambig_uc_caps_est;
+        else
+          est_caps_ht = est_x_ht / x_ht_fraction_of_caps_ht;
+      }
+      if (case_ambig.get_total () > 0)
+        improve_estimate(word_res, est_x_ht, est_caps_ht, x_ht, caps_ht); 
+      est_caps_ht_certain = caps_ht.get_total () > 0;
+      #ifndef SECURE_NAMES
+      if (debug_x_ht_level >= 20)
+        tprintf ("Mode20:D: Est from xht XHT:%f CAP:%f\n",
+          est_x_ht, est_caps_ht);
+      #endif
+    }
+    else if (est_caps_ht > 0) {
+      est_caps_ht_certain = TRUE;
+      if ((ambig_lc_x_est > 0) &&
+        (ambig_lc_x_est < est_caps_ht - x_ht_ok_variation))
+        est_x_ht = ambig_lc_x_est;
+      else
+        est_x_ht = est_caps_ht * x_ht_fraction_of_caps_ht;
+      if (ambig_lc_x_est + ambig_uc_caps_est > 0)
+        improve_estimate(word_res, est_x_ht, est_caps_ht, x_ht, caps_ht); 
+      est_x_ht_certain = x_ht.get_total () > 0;
+      #ifndef SECURE_NAMES
+      if (debug_x_ht_level >= 20)
+        tprintf ("Mode20:E: Est from caps XHT:%f CAP:%f\n",
+          est_x_ht, est_caps_ht);
+      #endif
+    }
+    else {
+      /* Do something based on case ambig chars alone - we have guessed that the
+        ambigs are lower case. */
+      est_x_ht = ambig_lc_x_est;
+      est_x_ht_certain = TRUE;
+      if (ambig_uc_caps_est > ambig_lc_x_est) {
+        est_caps_ht = ambig_uc_caps_est;
+        est_caps_ht_certain = TRUE;
+      }
+      else
+        est_caps_ht = est_x_ht / x_ht_fraction_of_caps_ht;
+
+      #ifndef SECURE_NAMES
+      if (debug_x_ht_level >= 20)
+        tprintf ("Mode20:F: Est from ambigs XHT:%f CAP:%f\n",
+          est_x_ht, est_caps_ht);
+      #endif
+    }
+    /* Check for sane interpretation of evidence:
+      Try shifting caps ht if min certain caps ht is not significantly greater
+      than the estimated x ht or the max certain x ht is not significantly less
+      than the estimated caps ht. */
+    if (x_ht_check_est) {
+      if ((caps_ht.get_total () > 0) &&
+      (est_x_ht + x_ht_ok_variation >= caps_ht.ile (0.0001))) {
+        trial = TRUE;
+        est_caps_ht = est_x_ht;
+        est_x_ht = x_ht_fraction_of_caps_ht * est_caps_ht;
+
+        #ifndef SECURE_NAMES
+        if (debug_x_ht_level >= 20)
+          tprintf ("Mode20:G: Trial XHT:%f CAP:%f\n",
+            est_x_ht, est_caps_ht);
+        #endif
+      }
+      else if ((x_ht.get_total () > 0) &&
+      (est_caps_ht - x_ht_ok_variation <= x_ht.ile (0.9999))) {
+        trial = TRUE;
+        est_x_ht = est_caps_ht;
+        est_caps_ht = est_x_ht / x_ht_fraction_of_caps_ht;
+        #ifndef SECURE_NAMES
+        if (debug_x_ht_level >= 20)
+          tprintf ("Mode20:H: Trial XHT:%f CAP:%f\n",
+            est_x_ht, est_caps_ht);
+        #endif
+      }
+    }
+  }
+
+  else {
+    /* There is no sensible data so we're in the dark. */
+
+    marginally_above_x_ht = bln_x_height +
+      x_ht_ok_variation * x_ht_sub_variation;
+    /*
+      If there are no rejects, or the only rejects have a narrow height, or have
+      a small area compared to a normal char, then estimate the x-height as the
+      original one. (I.e dont fiddle about if the only rejects look like
+      punctuation) - we use max height as mean or median will be too low if
+      there are only two blobs - Eg "F."
+    */
+
+    if (debug_x_ht_level >= 20)
+      tprintf ("Mode20:I: In the dark\n");
+
+    if ((rej_blobs_count == 0) ||
+      (rej_blobs_max_height < 0.3 * max_blob_ht) ||
+    (rej_blobs_max_area < 0.3 * max_blob_ht * max_blob_ht)) {
+      no_comment = TRUE;
+      if (debug_x_ht_level >= 20)
+        tprintf ("Mode20:J: No comment due to no rejects\n");
+    }
+    else if (x_ht_limit_flip_trials &&
+      ((max_blob_ht < marginally_above_x_ht) ||
+      ((ambig_lc_x_est > 0) &&
+      (ambig_lc_x_est == ambig_uc_caps_est) &&
+    (ambig_lc_x_est < marginally_above_x_ht)))) {
+      no_comment = TRUE;
+      if (debug_x_ht_level >= 20)
+        tprintf ("Mode20:K: No comment as close to xht %f < %f\n",
+          ambig_lc_x_est, marginally_above_x_ht);
+    }
+    else if (x_ht_conservative_ambigs && (ambig_uc_caps_est > 0)) {
+      trial = TRUE;
+      est_caps_ht = ambig_lc_x_est;
+      est_x_ht = x_ht_fraction_of_caps_ht * est_caps_ht;
+
+      #ifndef SECURE_NAMES
+      if (debug_x_ht_level >= 20)
+        tprintf ("Mode20:L: Trial XHT:%f CAP:%f\n",
+          est_x_ht, est_caps_ht);
+      #endif
+    }
+    /*
+      If the top of the word is nowhere near where we expect ascenders to be
+      (less than half the x_ht -> caps_ht distance) - suspect an all caps word
+      at the x-ht. Estimate x-ht accordingly - but only as a TRIAL!
+      NOTE we do NOT check location of baseline. Commas can descend as much as
+      real descenders so we would need to do something to make sure that any
+      disqualifying descenders were not at the end.
+    */
+    else {
+      if (max_blob_ht <
+      (bln_x_height + bln_x_height / x_ht_fraction_of_caps_ht) / 2.0) {
+        trial = TRUE;
+        est_x_ht = x_ht_fraction_of_caps_ht * max_blob_ht;
+        est_caps_ht = max_blob_ht;
+
+        #ifndef SECURE_NAMES
+        if (debug_x_ht_level >= 20)
+          tprintf ("Mode20:M: Trial XHT:%f CAP:%f\n",
+            est_x_ht, est_caps_ht);
+        #endif
+      }
+      else {
+        no_comment = TRUE;
+        if (debug_x_ht_level >= 20)
+          tprintf ("Mode20:N: No comment as nothing else matched\n");
+      }
+    }
+  }
+
+  /* Sanity check - reject word if fails */
+
+  if (!no_comment &&
+    ((est_x_ht > 2 * bln_x_height) ||
+    (est_x_ht / word_res->denorm.scale () <= min_sane_x_ht_pixels) ||
+  (est_caps_ht <= est_x_ht) || (est_caps_ht >= 2.5 * est_x_ht))) {
+    no_comment = TRUE;
+    if (!trial && rej_use_xht) {
+      if (debug_x_ht_level >= 2) {
+        tprintf ("Sanity check rejecting %s ", word_str);
+        word_res->reject_map.print (debug_fp);
+        tprintf ("\n");
+      }
+      word_res->reject_map.rej_word_xht_fixup ();
+
+    }
+    if (debug_x_ht_level >= 20)
+      tprintf ("Mode20:O: No comment as nothing else matched\n");
+  }
+
+  if (no_comment || trial) {
+    word_res->x_height = bln_x_height / word_res->denorm.scale ();
+    word_res->guessed_x_ht = TRUE;
+    word_res->caps_height = (bln_x_height / x_ht_fraction_of_caps_ht) /
+      word_res->denorm.scale ();
+    word_res->guessed_caps_ht = TRUE;
+    /*
+    Reject ambigs in the current word if we are uncertain and:
+        there are rejects OR
+        there is only one char which is an ambig OR
+        there is conflict between the case of the ambigs even though there is
+        no height separation Eg "Ms" recognised from "MS"
+    */
+    if (rej_trial_ambigs &&
+      ((word_res->reject_map.reject_count () > 0) ||
+      (word_res->reject_map.length () == 1) ||
+    ((x_ht_ambigs > 0) && (caps_ht_ambigs > 0)))) {
+      #ifndef SECURE_NAMES
+      if (debug_x_ht_level >= 2) {
+        tprintf ("TRIAL Rej Ambigs %s ", word_str);
+        word_res->reject_map.print (debug_fp);
+      }
+      #endif
+      reject_ambigs(word_res); 
+      if (debug_x_ht_level >= 2) {
+        tprintf (" ");
+        word_res->reject_map.print (debug_fp);
+        tprintf ("\n");
+      }
+    }
+  }
+  else {
+    word_res->x_height = est_x_ht / word_res->denorm.scale ();
+    word_res->guessed_x_ht = !est_x_ht_certain;
+    word_res->caps_height = est_caps_ht / word_res->denorm.scale ();
+    word_res->guessed_caps_ht = !est_caps_ht_certain;
+  }
+
+  if (!no_comment && (fabs (est_x_ht - bln_x_height) > x_ht_ok_variation))
+    *trial_x_ht = est_x_ht / word_res->denorm.scale ();
+  else
+    *trial_x_ht = 0.0;
+
+  #ifndef SECURE_NAMES
+  if (((*trial_x_ht > 0) && (debug_x_ht_level >= 3)) ||
+  (debug_x_ht_level >= 5)) {
+    tprintf ("%s ", word_str);
+    word_res->reject_map.print (debug_fp);
+    tprintf
+      (" X:%0.2f Cps:%0.2f Mxht:%0.2f RJ MxHt:%d MxAr:%d Rematch:%c\n",
+      est_x_ht, est_caps_ht, max_blob_ht, rej_blobs_max_height,
+      rej_blobs_max_area, *trial_x_ht > 0 ? '*' : ' ');
+  }
+  #endif
+
+}
+
+
+/*************************************************************************
+ * check_block_occ()
+ * Checks word for coarse block occupancy, rejecting more chars and flipping
+ * case of case ambiguous chars as required.
+ *************************************************************************/
+
+void check_block_occ(WERD_RES *word_res) { 
+  PBLOB_IT blob_it;
+  STRING new_string;
+  REJMAP new_map = word_res->reject_map;
+  WERD_CHOICE *new_choice;
+
+  const char *word_str = word_res->best_choice->string ().string ();
+  INT16 i;
+  INT16 reject_count = 0;
+  char confirmed_char;
+  float x_ht;
+  float caps_ht;
+
+  if (word_res->x_height > 0)
+    x_ht = word_res->x_height * word_res->denorm.scale ();
+  else
+    x_ht = bln_x_height;
+
+  if (word_res->caps_height > 0)
+    caps_ht = word_res->caps_height * word_res->denorm.scale ();
+  else
+    caps_ht = x_ht / x_ht_fraction_of_caps_ht;
+
+  blob_it.set_to_list (word_res->outword->blob_list ());
+
+  for (blob_it.mark_cycle_pt (), i = 0;
+  !blob_it.cycled_list (); blob_it.forward (), i++) {
+    new_string += word_str[i];   //default copy
+    if (word_res->reject_map[i].accepted ()) {
+      confirmed_char = check_blob_occ (word_str[i],
+        blob_it.data ()->bounding_box ().
+        top () - bln_baseline_offset, x_ht,
+        caps_ht);
+
+      if (confirmed_char == '\0') {
+        if (rej_use_check_block_occ) {
+          new_map[i].setrej_xht_fixup ();
+          reject_count++;
+        }
+      }
+      else
+        new_string[i] = confirmed_char;
+    }
+  }
+  if ((reject_count > 0) || (new_string != word_str)) {
+    if (debug_x_ht_level >= 2) {
+      tprintf ("Shape Verification: %s ", word_str);
+      word_res->reject_map.print (debug_fp);
+      tprintf (" -> %s ", new_string.string ());
+      new_map.print (debug_fp);
+      tprintf ("\n");
+    }
+    new_choice = new WERD_CHOICE (new_string.string (),
+      word_res->best_choice->rating (),
+      word_res->best_choice->certainty (),
+      word_res->best_choice->permuter ());
+    delete word_res->best_choice;
+    word_res->best_choice = new_choice;
+    word_res->reject_map = new_map;
+  }
+}
+
+
+/*************************************************************************
+ * check_blob_occ()
+ *
+ * Checks blob for position relative to position above baseline
+ * Returns 0 for reject, or (possibly case shifted) confirmed char
+ *************************************************************************/
+
+char check_blob_occ(char proposed_char,
+                    INT16 blob_ht_above_baseline,
+                    float x_ht,
+                    float caps_ht) {
+  BOOL8 blob_definite_x_ht;
+  BOOL8 blob_definite_caps_ht;
+  float acceptable_variation;
+
+  acceptable_variation = (caps_ht - x_ht) * x_ht_variation;
+  /* ??? REJECT if expected descender and nothing significantly below BL */
+
+  /* ??? REJECT if expected ascender and nothing significantly above x-ht */
+
+  /*
+    IF AMBIG_CAPS_X_CHS
+      IF blob is definitely an ascender ( > xht + xht err )AND
+        char is an x-ht char
+      THEN
+        flip case
+      IF blob is defintiely an x-ht ( <= xht + xht err ) AND
+        char is an ascender char
+      THEN
+        flip case
+  */
+  blob_definite_x_ht = blob_ht_above_baseline <= x_ht + acceptable_variation;
+  blob_definite_caps_ht = blob_ht_above_baseline >=
+    caps_ht - acceptable_variation;
+
+  if (STRING (chs_ambig_caps_x).contains (proposed_char)) {
+    if ((!blob_definite_x_ht && !blob_definite_caps_ht) ||
+      (proposed_char == '0' && !blob_definite_caps_ht) ||
+      (proposed_char == 'o' && !blob_definite_x_ht))
+      return '\0';
+
+    else if (blob_definite_caps_ht &&
+    STRING (chs_x_ht).contains (proposed_char)) {
+      if (x_ht_case_flip)
+                                 //flip to upper case
+        return (char) toupper (proposed_char);
+      else
+        return '\0';
+    }
+
+    else if (blob_definite_x_ht &&
+    !STRING (chs_x_ht).contains (proposed_char)) {
+      if (x_ht_case_flip)
+                                 //flip to lower case
+        return (char) tolower (proposed_char);
+      else
+        return '\0';
+    }
+  }
+  else
+  if ((STRING (chs_non_ambig_x_ht).contains (proposed_char)
+    && !blob_definite_x_ht)
+    || (STRING (chs_non_ambig_caps_ht).contains (proposed_char)
+    && !blob_definite_caps_ht))
+    return '\0';
+  return proposed_char;
+}
+
+
+float estimate_from_stats(STATS &stats) { 
+  if (stats.get_total () <= 0)
+    return 0.0;
+  else if (stats.get_total () >= 3)
+    return stats.ile (0.5);      //median
+  else
+    return stats.mean ();
+}
+
+
+void improve_estimate(WERD_RES *word_res,
+                      float &est_x_ht,
+                      float &est_caps_ht,
+                      STATS &x_ht,
+                      STATS &caps_ht) {
+  PBLOB_IT blob_it;
+  INT16 blob_ht_above_baseline;
+
+  const char *word_str;
+  INT16 i;
+  BOX blob_box;                  //blob bounding box
+  char confirmed_char;
+  float new_val;
+
+  /* IMPROVE estimates here - if good estimates, and case ambig chars,
+    rescan blobs to fix case ambig blobs, re-estimate hts  ??? maybe always do
+    it after deciding x-height
+  */
+
+  blob_it.set_to_list (word_res->outword->blob_list ());
+  word_str = word_res->best_choice->string ().string ();
+  for (blob_it.mark_cycle_pt (), i = 0;
+  !blob_it.cycled_list (); blob_it.forward (), i++) {
+    if ((STRING (chs_ambig_caps_x).contains (word_str[i])) &&
+    (!dodgy_blob (blob_it.data ()))) {
+      blob_box = blob_it.data ()->bounding_box ();
+      blob_ht_above_baseline = blob_box.top () - bln_baseline_offset;
+      confirmed_char = check_blob_occ (word_str[i],
+        blob_ht_above_baseline,
+        est_x_ht, est_caps_ht);
+      if (confirmed_char != '\0')
+        if (STRING (chs_x_ht).contains (confirmed_char))
+          x_ht.add (blob_ht_above_baseline, 1);
+      else
+        caps_ht.add (blob_ht_above_baseline, 1);
+    }
+  }
+  new_val = estimate_from_stats (x_ht);
+  if (new_val > 0)
+    est_x_ht = new_val;
+  new_val = estimate_from_stats (caps_ht);
+  if (new_val > 0)
+    est_caps_ht = new_val;
+}
+
+
+void reject_ambigs(  //rej any accepted xht ambig chars
+                   WERD_RES *word) {
+  const char *word_str;
+  int i = 0;
+
+  word_str = word->best_choice->string ().string ();
+  while (*word_str != '\0') {
+    if (STRING (chs_ambig_caps_x).contains (*word_str))
+      word->reject_map[i].setrej_xht_fixup ();
+    word_str++;
+    i++;
+  }
+}
+
+
+void est_ambigs(                          //xht ambig ht stats
+                WERD_RES *word_res,
+                STATS &stats,
+                float *ambig_lc_x_est,    //xht est
+                float *ambig_uc_caps_est  //caps est
+               ) {
+  float x_ht_ok_variation;
+  STATS short_ambigs (0, 300);
+  STATS tall_ambigs (0, 300);
+  PBLOB_IT blob_it;
+  BOX blob_box;                  //blob bounding box
+  INT16 blob_ht_above_baseline;
+
+  const char *word_str;
+  INT16 i;
+  float min;                     //min ambig ch ht
+  float max;                     //max ambig ch ht
+  float short_limit;             // for lower case
+  float tall_limit;              // for upper case
+
+  x_ht_ok_variation =
+    (bln_x_height / x_ht_fraction_of_caps_ht - bln_x_height) * x_ht_variation;
+
+  if (stats.get_total () == 0) {
+    *ambig_lc_x_est = 0;
+    *ambig_uc_caps_est = 0;
+  }
+  else {
+    min = stats.ile (0.0);
+    max = stats.ile (0.99999);
+    if ((max - min) < x_ht_ok_variation) {
+      *ambig_lc_x_est = *ambig_uc_caps_est = stats.mean ();
+      //close enough
+    }
+    else {
+    /* Try reclustering into lower and upper case chars */
+      short_limit = min + (max - min) * x_ht_variation;
+      tall_limit = max - (max - min) * x_ht_variation;
+      word_str = word_res->best_choice->string ().string ();
+      blob_it.set_to_list (word_res->outword->blob_list ());
+      for (blob_it.mark_cycle_pt (), i = 0;
+      !blob_it.cycled_list (); blob_it.forward (), i++) {
+        if (word_res->reject_map[i].accepted () &&
+          STRING (chs_ambig_caps_x).contains (word_str[i]) &&
+        (!dodgy_blob (blob_it.data ()))) {
+          blob_box = blob_it.data ()->bounding_box ();
+          blob_ht_above_baseline =
+            blob_box.top () - bln_baseline_offset;
+          if (blob_ht_above_baseline <= short_limit)
+            short_ambigs.add (blob_ht_above_baseline, 1);
+          else if (blob_ht_above_baseline >= tall_limit)
+            tall_ambigs.add (blob_ht_above_baseline, 1);
+        }
+      }
+      *ambig_lc_x_est = short_ambigs.mean ();
+      *ambig_uc_caps_est = tall_ambigs.mean ();
+      /* Cop out if we havent got sensible clusters. */
+      if (*ambig_uc_caps_est - *ambig_lc_x_est <= x_ht_ok_variation)
+        *ambig_lc_x_est = *ambig_uc_caps_est = stats.mean ();
+      //close enough
+    }
+  }
+}
+
+
+/*************************************************************************
+ * dodgy_blob()
+ * Returns true if the blob has more than one outline, one above the other.
+ * These are dodgy as the top blob could be noise, causing the bounding box xht
+ * to be misleading
+ *************************************************************************/
+
+BOOL8 dodgy_blob(PBLOB *blob) { 
+  OUTLINE_IT outline_it = blob->out_list ();
+  INT16 highest_bottom = -MAX_INT16;
+  INT16 lowest_top = MAX_INT16;
+  BOX outline_box;
+
+  if (x_ht_include_dodgy_blobs)
+    return FALSE;                //no blob is ever dodgy
+  for (outline_it.mark_cycle_pt ();
+  !outline_it.cycled_list (); outline_it.forward ()) {
+    outline_box = outline_it.data ()->bounding_box ();
+    if (lowest_top > outline_box.top ())
+      lowest_top = outline_box.top ();
+    if (highest_bottom < outline_box.bottom ())
+      highest_bottom = outline_box.bottom ();
+  }
+  return highest_bottom >= lowest_top;
+}

ccmain/fixxht.h

@@ -0,0 +1,92 @@
+/**********************************************************************
+ * File:        fixxht.h  (Formerly fixxht.h)
+ * Description: Improve x_ht and look out for case inconsistencies
+ * Author:		Phil Cheatle
+ * Created:		Thu Aug  5 14:11:08 BST 1993
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           FIXXHT_H
+#define           FIXXHT_H
+
+#include          "varable.h"
+#include          "statistc.h"
+#include          "pageres.h"
+#include          "notdll.h"
+
+extern double_VAR_H (x_ht_fraction_of_caps_ht, 0.7,
+"Fract of cps ht est of xht");
+extern double_VAR_H (x_ht_variation, 0.35,
+"Err band as fract of caps/xht dist");
+extern double_VAR_H (x_ht_sub_variation, 0.5,
+"Err band as fract of caps/xht dist");
+extern BOOL_VAR_H (rej_trial_ambigs, TRUE,
+"reject x-ht ambigs when under trial");
+extern BOOL_VAR_H (x_ht_conservative_ambigs, FALSE,
+"Dont rely on ambigs + maxht");
+extern BOOL_VAR_H (x_ht_check_est, TRUE, "Cross check estimates");
+extern BOOL_VAR_H (x_ht_case_flip, FALSE, "Flip or reject suspect case");
+extern BOOL_VAR_H (x_ht_include_dodgy_blobs, TRUE,
+"Include blobs with possible noise?");
+extern BOOL_VAR_H (x_ht_limit_flip_trials, TRUE,
+"Dont do trial flips when ambigs are close to xht?");
+extern BOOL_VAR_H (rej_use_check_block_occ, TRUE,
+"Analyse rejection behaviour");
+extern STRING_VAR_H (chs_non_ambig_caps_ht,
+"!#$%&()/12346789?ABDEFGHIKLNQRT[]\\bdfhkl",
+"Reliable ascenders");
+extern STRING_VAR_H (chs_x_ht, "acegmnopqrsuvwxyz", "X height chars");
+extern STRING_VAR_H (chs_non_ambig_x_ht, "aenqr", "reliable X height chars");
+extern STRING_VAR_H (chs_ambig_caps_x, "cCmMoO05sSuUvVwWxXzZ",
+"X ht or caps ht chars");
+extern STRING_VAR_H (chs_bl_ambig_caps_x, "pPyY",
+" Caps or descender ambigs");
+extern STRING_VAR_H (chs_caps_ht,
+"!#$%&()/0123456789?ABCDEFGHIJKLMNOPQRSTUVWXYZ[]\\bdfhkl{|}",
+"Ascender chars");
+extern STRING_VAR_H (chs_desc, "gjpqy", "Descender chars");
+extern STRING_VAR_H (chs_non_ambig_bl,
+"!#$%&01246789?ABCDEFGHIKLMNORSTUVWXYZabcdehiklmnorstuvwxz",
+"Reliable baseline chars");
+extern STRING_VAR_H (chs_odd_top, "ijt", "Chars with funny ascender region");
+extern STRING_VAR_H (chs_odd_bot, "()35JQ[]\\/{}|", "Chars with funny base");
+extern STRING_VAR_H (chs_bl,
+"!#$%&()/01246789?ABCDEFGHIJKLMNOPRSTUVWXYZ[]\\abcdefhiklmnorstuvwxz{}",
+"Baseline chars");
+extern STRING_VAR_H (chs_non_ambig_desc, "gq", "Reliable descender chars");
+void re_estimate_x_ht(                     //improve for 1 word
+                      WERD_RES *word_res,  //word to do
+                      float *trial_x_ht    //new match value
+                     );
+void check_block_occ(WERD_RES *word_res); 
+char check_blob_occ(char proposed_char,
+                    INT16 blob_ht_above_baseline,
+                    float x_ht,
+                    float caps_ht);
+float estimate_from_stats(STATS &stats); 
+void improve_estimate(WERD_RES *word_res,
+                      float &est_x_ht,
+                      float &est_caps_ht,
+                      STATS &x_ht,
+                      STATS &caps_ht);
+void reject_ambigs(  //rej any accepted xht ambig chars
+                   WERD_RES *word);
+                                 //xht ambig ht stats
+void est_ambigs(WERD_RES *word_res,
+                STATS &stats,
+                float *ambig_lc_x_est,    //xht est
+                float *ambig_uc_caps_est  //caps est
+               );
+BOOL8 dodgy_blob(PBLOB *blob); 
+#endif

ccmain/imgscale.cpp

@@ -0,0 +1,154 @@
+/**********************************************************************
+ * File:        imgscale.cpp  (Formerly dyn_prog.c)
+ * Description: Dynamic programming for smart scaling of images.
+ * Author:      Phil Cheatle
+ * Created:     Wed Nov 18 16:12:03 GMT 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+/*************************************************************************
+ * This is really Sheelagh's code that I've hacked into a more usable form.
+ * It is used by scaleim.c  All I did to it was to change "factor" from int to
+ * float.
+ *************************************************************************/
+/************************************************************************
+ *  This version uses the result of the previous row to influence the
+ *  current row's calculation.
+ ************************************************************************/
+
+#include "mfcpch.h"
+#include <stdio.h>
+#include <stdlib.h>
+#include "errcode.h"
+
+#define f(xc, yc) ((xc - factor*yc)*(xc - factor*yc))
+
+#define g(oldyc, yc, oldxc, xc) (factor*factor*(oldyc - yc)*(oldyc - yc)/(abs(oldxc - xc) + 1))
+
+void
+dyn_exit (const char s[]) {
+  fprintf (stderr, "%s", s);
+  err_exit(); 
+}
+
+
+void dyn_prog(  //The clever bit
+              int n,
+              int *x,
+              int *y,
+              int ymax,
+              int *oldx,
+              int *oldy,
+              int oldn,
+              float factor) {
+  int i, z, j, matchflag;
+  int **ymin;
+  float **F, fz;
+
+  /* F[i][z] gives minimum over y <= z */
+
+  F = (float **) calloc (n, sizeof (float *));
+  ymin = (int **) calloc (n, sizeof (int *));
+  if ((F == NULL) || (ymin == NULL))
+    dyn_exit ("Error in calloc\n");
+
+  for (i = 0; i < n; i++) {
+    F[i] = (float *) calloc (ymax - n + i + 1, sizeof (float));
+    ymin[i] = (int *) calloc (ymax - n + i + 1, sizeof (int));
+    if ((F[i] == NULL) || (ymin[i] == NULL))
+      dyn_exit ("Error in calloc\n");
+  }
+
+  F[0][0] = f (x[0], 0);
+  /* find nearest transition of same sign (white to black) */
+  j = 0;
+  while ((j < oldn) && (oldx[j] < x[0]))
+    j += 2;
+  if (j >= oldn)
+    j -= 2;
+  else if ((j - 2 >= 0) && ((x[0] - oldx[j - 2]) < (oldx[j] - x[0])))
+    j -= 2;
+  if (abs (oldx[j] - x[0]) < factor) {
+    matchflag = 1;
+    F[0][0] += g (oldy[j], 0, oldx[j], x[0]);
+  }
+  else
+    matchflag = 0;
+  ymin[0][0] = 0;
+
+  for (z = 1; z < ymax - n + 1; z++) {
+    fz = f (x[0], z);
+    /* add penalty for deviating from previous row if necessary */
+
+    if (matchflag)
+      fz += g (oldy[j], z, oldx[j], x[0]);
+    if (fz < F[0][z - 1]) {
+      F[0][z] = fz;
+      ymin[0][z] = z;
+    }
+    else {
+      F[0][z] = F[0][z - 1];
+      ymin[0][z] = ymin[0][z - 1];
+    }
+  }
+
+  for (i = 1; i < n; i++) {
+    F[i][i] = f (x[i], i) + F[i - 1][i - 1];
+    /* add penalty for deviating from previous row if necessary */
+    if (j > 0)
+      j--;
+    else
+      j++;
+    while ((j < oldn) && (oldx[j] < x[i]))
+      j += 2;
+    if (j >= oldn)
+      j -= 2;
+    else if ((j - 2 >= 0) && ((x[i] - oldx[j - 2]) < (oldx[j] - x[i])))
+      j -= 2;
+    if (abs (oldx[j] - x[i]) < factor) {
+      matchflag = 1;
+      F[i][i] += g (oldy[j], i, oldx[j], x[i]);
+    }
+    else
+      matchflag = 0;
+    ymin[i][i] = i;
+    for (z = i + 1; z < ymax - n + i + 1; z++) {
+      fz = f (x[i], z) + F[i - 1][z - 1];
+      /* add penalty for deviating from previous row if necessary */
+      if (matchflag)
+        fz += g (oldy[j], z, oldx[j], x[i]);
+      if (fz < F[i][z - 1]) {
+        F[i][z] = fz;
+        ymin[i][z] = z;
+      }
+      else {
+        F[i][z] = F[i][z - 1];
+        ymin[i][z] = ymin[i][z - 1];
+      }
+    }
+  }
+
+  y[n - 1] = ymin[n - 1][ymax - 1];
+  for (i = n - 2; i >= 0; i--)
+    y[i] = ymin[i][y[i + 1] - 1];
+
+  for (i = 0; i < n; i++) {
+    free (F[i]);
+    free (ymin[i]);
+  }
+  free(F); 
+  free(ymin); 
+
+  return;
+}

ccmain/imgscale.h

@@ -0,0 +1,32 @@
+/**********************************************************************
+ * File:        imgscale.h  (Formerly dyn_prog.h)
+ * Description: Dynamic programming for smart scaling of images.
+ * Author:      Phil Cheatle
+ * Created:     Wed Nov 18 16:12:03 GMT 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           IMGSCALE_H
+#define           IMGSCALE_H
+
+void dyn_prog(  //The clever bit
+              int n,
+              int *x,
+              int *y,
+              int ymax,
+              int *oldx,
+              int *oldy,
+              int oldn,
+              float factor);
+#endif

ccmain/matmatch.cpp

@@ -0,0 +1,404 @@
+/**********************************************************************
+ * File:        matmatch.cpp  (Formerly matrix_match.c)
+ * Description: matrix matching routines for Tessedit
+ * Author:      Chris Newton
+ * Created:     Wed Nov 24 15:57:41 GMT 1993
+ *
+ * (C) Copyright 1993, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          <stdlib.h>
+#include          <math.h>
+#include          <string.h>
+#include          <ctype.h>
+#ifdef __UNIX__
+#include          <assert.h>
+#endif
+#include          "tessvars.h"
+#include          "stderr.h"
+#include          "img.h"
+#include          "evnts.h"
+#include          "showim.h"
+#include          "hosthplb.h"
+#include          "grphics.h"
+#include          "evnts.h"
+#include          "adaptions.h"
+#include          "matmatch.h"
+#include          "secname.h"
+
+#define EXTERN
+
+EXTERN BOOL_VAR (tessedit_display_mm, FALSE, "Display matrix matches");
+EXTERN BOOL_VAR (tessedit_mm_debug, FALSE,
+"Print debug information for matrix matcher");
+EXTERN INT_VAR (tessedit_mm_prototype_min_size, 3,
+"Smallest number of samples in a cluster for a prototype to be used");
+
+// Colours for displaying the match
+#define BB_COLOUR 0
+#define BW_COLOUR 1
+#define WB_COLOUR 3
+#define UB_COLOUR 5
+#define BU_COLOUR 7
+#define UU_COLOUR 9
+#define WU_COLOUR 11
+#define UW_COLOUR 13
+#define WW_COLOUR 15
+
+#define BINIM_BLACK 0
+#define BINIM_WHITE 1
+
+float matrix_match(  // returns match score
+                   IMAGE *image1,
+                   IMAGE *image2) {
+  ASSERT_HOST (image1->get_bpp () == 1 && image2->get_bpp () == 1);
+
+  if (image1->get_xsize () >= image2->get_xsize ())
+    return match1 (image1, image2);
+  else
+    return match1 (image2, image1);
+}
+
+
+float match1(  /* returns match score */
+             IMAGE *image_w,
+             IMAGE *image_n) {
+  INT32 x_offset;
+  INT32 y_offset;
+  INT32 x_size = image_w->get_xsize ();
+  INT32 y_size;
+  INT32 x_size2 = image_n->get_xsize ();
+  INT32 y_size2;
+  IMAGE match_image;
+  IMAGELINE imline_w;
+  IMAGELINE imline_n;
+  IMAGELINE match_imline;
+  INT32 x;
+  INT32 y;
+  float sum = 0.0;
+
+  x_offset = (image_w->get_xsize () - image_n->get_xsize ()) / 2;
+
+  ASSERT_HOST (x_offset >= 0);
+  match_imline.init (x_size);
+
+  sum = 0;
+
+  if (image_w->get_ysize () < image_n->get_ysize ()) {
+    y_size = image_n->get_ysize ();
+    y_size2 = image_w->get_ysize ();
+    y_offset = (y_size - y_size2) / 2;
+
+    if (tessedit_display_mm && !tessedit_mm_use_prototypes)
+      tprintf ("I1 (%d, %d), I2 (%d, %d), MI (%d, %d)\n", x_size,
+        image_w->get_ysize (), x_size2, image_n->get_ysize (),
+        x_size, y_size);
+
+    match_image.create (x_size, y_size, 4);
+
+    for (y = 0; y < y_offset; y++) {
+      image_n->fast_get_line (0, y, x_size2, &imline_n);
+      for (x = 0; x < x_size2; x++) {
+        if (imline_n.pixels[x] == BINIM_BLACK) {
+          sum += -1;
+          match_imline.pixels[x] = UB_COLOUR;
+        }
+        else {
+          match_imline.pixels[x] = UW_COLOUR;
+        }
+      }
+      match_image.fast_put_line (x_offset, y, x_size2, &match_imline);
+    }
+
+    for (y = y_offset + y_size2; y < y_size; y++) {
+      image_n->fast_get_line (0, y, x_size2, &imline_n);
+      for (x = 0; x < x_size2; x++) {
+        if (imline_n.pixels[x] == BINIM_BLACK) {
+          sum += -1.0;
+          match_imline.pixels[x] = UB_COLOUR;
+        }
+        else {
+          match_imline.pixels[x] = UW_COLOUR;
+        }
+      }
+      match_image.fast_put_line (x_offset, y, x_size2, &match_imline);
+    }
+
+    for (y = y_offset; y < y_offset + y_size2; y++) {
+      image_w->fast_get_line (0, y - y_offset, x_size, &imline_w);
+      image_n->fast_get_line (0, y, x_size2, &imline_n);
+      for (x = 0; x < x_offset; x++) {
+        if (imline_w.pixels[x] == BINIM_BLACK) {
+          sum += -1.0;
+          match_imline.pixels[x] = BU_COLOUR;
+        }
+        else {
+          match_imline.pixels[x] = WU_COLOUR;
+        }
+      }
+
+      for (x = x_offset + x_size2; x < x_size; x++) {
+        if (imline_w.pixels[x] == BINIM_BLACK) {
+          sum += -1.0;
+          match_imline.pixels[x] = BU_COLOUR;
+        }
+        else {
+          match_imline.pixels[x] = WU_COLOUR;
+        }
+      }
+
+      for (x = x_offset; x < x_offset + x_size2; x++) {
+        if (imline_n.pixels[x - x_offset] == imline_w.pixels[x]) {
+          sum += 1.0;
+          if (imline_w.pixels[x] == BINIM_BLACK)
+            match_imline.pixels[x] = BB_COLOUR;
+          else
+            match_imline.pixels[x] = WW_COLOUR;
+        }
+        else {
+          sum += -1.0;
+          if (imline_w.pixels[x] == BINIM_BLACK)
+            match_imline.pixels[x] = BW_COLOUR;
+          else
+            match_imline.pixels[x] = WB_COLOUR;
+        }
+      }
+
+      match_image.fast_put_line (0, y, x_size, &match_imline);
+    }
+  }
+  else {
+    y_size = image_w->get_ysize ();
+    y_size2 = image_n->get_ysize ();
+    y_offset = (y_size - y_size2) / 2;
+
+    if (tessedit_display_mm && !tessedit_mm_use_prototypes)
+      tprintf ("I1 (%d, %d), I2 (%d, %d), MI (%d, %d)\n", x_size,
+        image_w->get_ysize (), x_size2, image_n->get_ysize (),
+        x_size, y_size);
+
+    match_image.create (x_size, y_size, 4);
+
+    for (y = 0; y < y_offset; y++) {
+      image_w->fast_get_line (0, y, x_size, &imline_w);
+      for (x = 0; x < x_size; x++) {
+        if (imline_w.pixels[x] == BINIM_BLACK) {
+          sum += -1;
+          match_imline.pixels[x] = BU_COLOUR;
+        }
+        else {
+          match_imline.pixels[x] = WU_COLOUR;
+        }
+      }
+      match_image.fast_put_line (0, y, x_size, &match_imline);
+    }
+
+    for (y = y_offset + y_size2; y < y_size; y++) {
+      image_w->fast_get_line (0, y, x_size, &imline_w);
+      for (x = 0; x < x_size; x++) {
+        if (imline_w.pixels[x] == BINIM_BLACK) {
+          sum += -1;
+          match_imline.pixels[x] = BU_COLOUR;
+        }
+        else {
+          match_imline.pixels[x] = WU_COLOUR;
+        }
+      }
+      match_image.fast_put_line (0, y, x_size, &match_imline);
+    }
+
+    for (y = y_offset; y < y_offset + y_size2; y++) {
+      image_w->fast_get_line (0, y, x_size, &imline_w);
+      image_n->fast_get_line (0, y - y_offset, x_size2, &imline_n);
+      for (x = 0; x < x_offset; x++) {
+        if (imline_w.pixels[x] == BINIM_BLACK) {
+          sum += -1.0;
+          match_imline.pixels[x] = BU_COLOUR;
+        }
+        else {
+          match_imline.pixels[x] = WU_COLOUR;
+        }
+      }
+
+      for (x = x_offset + x_size2; x < x_size; x++) {
+        if (imline_w.pixels[x] == BINIM_BLACK) {
+          sum += -1.0;
+          match_imline.pixels[x] = BU_COLOUR;
+        }
+        else {
+          match_imline.pixels[x] = WU_COLOUR;
+        }
+      }
+
+      for (x = x_offset; x < x_offset + x_size2; x++) {
+        if (imline_n.pixels[x - x_offset] == imline_w.pixels[x]) {
+          sum += 1.0;
+          if (imline_w.pixels[x] == BINIM_BLACK)
+            match_imline.pixels[x] = BB_COLOUR;
+          else
+            match_imline.pixels[x] = WW_COLOUR;
+        }
+        else {
+          sum += -1.0;
+          if (imline_w.pixels[x] == BINIM_BLACK)
+            match_imline.pixels[x] = BW_COLOUR;
+          else
+            match_imline.pixels[x] = WB_COLOUR;
+        }
+      }
+
+      match_image.fast_put_line (0, y, x_size, &match_imline);
+    }
+  }
+
+#ifndef GRAPHICS_DISABLED
+  if (tessedit_display_mm && !tessedit_mm_use_prototypes) {
+    tprintf ("Match score %f\n", 1.0 - sum / (x_size * y_size));
+    display_images(image_w, image_n, &match_image); 
+  }
+#endif
+
+  if (tessedit_mm_debug)
+    tprintf ("Match score %f\n", 1.0 - sum / (x_size * y_size));
+
+  return (1.0 - sum / (x_size * y_size));
+}
+
+
+/*************************************************************************
+ * display_images()
+ *
+ * Show a pair of images, plus the match image
+ *
+ *************************************************************************/
+
+#ifndef GRAPHICS_DISABLED
+void display_images(IMAGE *image_w, IMAGE *image_n, IMAGE *match_image) { 
+  WINDOW w_im_window;
+  WINDOW n_im_window;
+  WINDOW match_window;
+  GRAPHICS_EVENT event;          //output event
+  INT16 i;
+
+                                 // xmin xmax ymin ymax
+  w_im_window = create_window ("Image 1", SCROLLINGWIN, 20, 100, 10 * image_w->get_xsize (), 10 * image_w->get_ysize (), 0, image_w->get_xsize (), 0, image_w->get_ysize (),
+    TRUE, FALSE, FALSE, TRUE);   // down event & key only
+
+  clear_view_surface(w_im_window); 
+  show_sub_image (image_w,
+    0, 0,
+    image_w->get_xsize (), image_w->get_ysize (),
+    w_im_window, 0, 0);
+
+  line_color_index(w_im_window, RED); 
+  for (i = 1; i < image_w->get_xsize (); i++) {
+    move2d (w_im_window, i, 0);
+    draw2d (w_im_window, i, image_w->get_ysize ());
+  }
+  for (i = 1; i < image_w->get_ysize (); i++) {
+    move2d (w_im_window, 0, i);
+    draw2d (w_im_window, image_w->get_xsize (), i);
+  }
+
+                                 // xmin xmax ymin ymax
+  n_im_window = create_window ("Image 2", SCROLLINGWIN, 240, 100, 10 * image_n->get_xsize (), 10 * image_n->get_ysize (), 0, image_n->get_xsize (), 0, image_n->get_ysize (),
+    TRUE, FALSE, FALSE, TRUE);   // down event & key only
+
+  clear_view_surface(n_im_window); 
+  show_sub_image (image_n,
+    0, 0,
+    image_n->get_xsize (), image_n->get_ysize (),
+    n_im_window, 0, 0);
+
+  line_color_index(n_im_window, RED); 
+  for (i = 1; i < image_n->get_xsize (); i++) {
+    move2d (n_im_window, i, 0);
+    draw2d (n_im_window, i, image_n->get_ysize ());
+  }
+  for (i = 1; i < image_n->get_ysize (); i++) {
+    move2d (n_im_window, 0, i);
+    draw2d (n_im_window, image_n->get_xsize (), i);
+  }
+  overlap_picture_ops(TRUE); 
+
+                                 // xmin xmax ymin ymax
+  match_window = create_window ("Match Result", SCROLLINGWIN, 460, 100, 10 * match_image->get_xsize (), 10 * match_image->get_ysize (), 0, match_image->get_xsize (), 0, match_image->get_ysize (),
+    TRUE, FALSE, FALSE, TRUE);   // down event & key only
+
+  clear_view_surface(match_window); 
+  show_sub_image (match_image,
+    0, 0,
+    match_image->get_xsize (), match_image->get_ysize (),
+    match_window, 0, 0);
+
+  line_color_index(match_window, RED); 
+  for (i = 1; i < match_image->get_xsize (); i++) {
+    move2d (match_window, i, 0);
+    draw2d (match_window, i, match_image->get_ysize ());
+  }
+  for (i = 1; i < match_image->get_ysize (); i++) {
+    move2d (match_window, 0, i);
+    draw2d (match_window, match_image->get_xsize (), i);
+  }
+  overlap_picture_ops(TRUE); 
+
+  await_event(match_window, TRUE, ANY_EVENT, &event); 
+  destroy_window(w_im_window); 
+  destroy_window(n_im_window); 
+  destroy_window(match_window); 
+}
+
+
+/*************************************************************************
+ * display_image()
+ *
+ * Show a single image
+ *
+ *************************************************************************/
+
+WINDOW display_image(IMAGE *image,
+                     const char *title,
+                     INT32 x,
+                     INT32 y,
+                     BOOL8 wait) {
+  WINDOW im_window;
+  INT16 i;
+  GRAPHICS_EVENT event;          //output event
+
+                                 // xmin xmax ymin ymax
+  im_window = create_window (title, SCROLLINGWIN, x, y, 10 * image->get_xsize (), 10 * image->get_ysize (), 0, image->get_xsize (), 0, image->get_ysize (),
+    TRUE, FALSE, FALSE, TRUE);   // down event & key only
+
+  clear_view_surface(im_window); 
+  show_sub_image (image,
+    0, 0,
+    image->get_xsize (), image->get_ysize (), im_window, 0, 0);
+
+  line_color_index(im_window, RED); 
+  for (i = 1; i < image->get_xsize (); i++) {
+    move2d (im_window, i, 0);
+    draw2d (im_window, i, image->get_ysize ());
+  }
+  for (i = 1; i < image->get_ysize (); i++) {
+    move2d (im_window, 0, i);
+    draw2d (im_window, image->get_xsize (), i);
+  }
+  overlap_picture_ops(TRUE); 
+
+  if (wait)
+    await_event(im_window, TRUE, ANY_EVENT, &event); 
+
+  return im_window;
+}
+#endif

ccmain/matmatch.h

@@ -0,0 +1,48 @@
+/**********************************************************************
+ * File:        matmatch.h  (Formerly matrix_match.h)
+ * Description: matrix matching routines for Tessedit
+ * Author:      Chris Newton
+ * Created:     Wed Nov 24 15:57:41 GMT 1993
+ *
+ * (C) Copyright 1993, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           MATMATCH_H
+#define           MATMATCH_H
+
+#include          "img.h"
+#include          "hosthplb.h"
+#include          "notdll.h"
+
+#define BINIM_BLACK 0
+#define BINIM_WHITE 1
+#define BAD_MATCH 9999.0
+
+extern BOOL_VAR_H (tessedit_display_mm, FALSE, "Display matrix matches");
+extern BOOL_VAR_H (tessedit_mm_debug, FALSE,
+"Print debug information for matrix matcher");
+extern INT_VAR_H (tessedit_mm_prototype_min_size, 3,
+"Smallest number of samples in a cluster for a prototype to be used");
+float matrix_match(  // returns match score
+                   IMAGE *image1,
+                   IMAGE *image2);
+float match1(  /* returns match score */
+             IMAGE *image_w,
+             IMAGE *image_n);
+void display_images(IMAGE *image_w, IMAGE *image_n, IMAGE *match_image); 
+WINDOW display_image(IMAGE *image,
+                     const char *title,
+                     INT32 x,
+                     INT32 y,
+                     BOOL8 wait);
+#endif

ccmain/output.h

@@ -0,0 +1,112 @@
+/******************************************************************
+ * File:        output.h  (Formerly output.h)
+ * Description: Output pass
+ * Author:		Phil Cheatle
+ * Created:		Thu Aug  4 10:56:08 BST 1994
+ *
+ * (C) Copyright 1994, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           OUTPUT_H
+#define           OUTPUT_H
+
+#include          "varable.h"
+//#include                                      "epapconv.h"
+#include          "pageres.h"
+#include          "notdll.h"
+
+extern BOOL_EVAR_H (tessedit_write_block_separators, TRUE,
+"Write block separators in output");
+extern BOOL_VAR_H (tessedit_write_raw_output, FALSE,
+"Write raw stuff to name.raw");
+extern BOOL_EVAR_H (tessedit_write_output, TRUE, "Write text to name.txt");
+extern BOOL_EVAR_H (tessedit_write_txt_map, TRUE,
+"Write .txt to .etx map file");
+extern BOOL_EVAR_H (tessedit_write_rep_codes, TRUE,
+"Write repetition char code");
+extern BOOL_EVAR_H (tessedit_write_unlv, FALSE, "Write .unlv output file");
+extern STRING_EVAR_H (unrecognised_char, "|",
+"Output char for unidentified blobs");
+extern INT_EVAR_H (suspect_level, 99, "Suspect marker level");
+extern INT_VAR_H (suspect_space_level, 100,
+"Min suspect level for rejecting spaces");
+extern INT_VAR_H (suspect_short_words, 2,
+"Dont Suspect dict wds longer than this");
+extern BOOL_VAR_H (suspect_constrain_1Il, FALSE,
+"UNLV keep 1Il chars rejected");
+extern double_VAR_H (suspect_rating_per_ch, 999.9,
+"Dont touch bad rating limit");
+extern double_VAR_H (suspect_accept_rating, -999.9,
+"Accept good rating limit");
+extern BOOL_EVAR_H (tessedit_minimal_rejection, FALSE,
+"Only reject tess failures");
+extern BOOL_VAR_H (tessedit_zero_rejection, FALSE, "Dont reject ANYTHING");
+extern BOOL_VAR_H (tessedit_word_for_word, FALSE,
+"Make output have exactly one word per WERD");
+extern BOOL_VAR_H (tessedit_consistent_reps, TRUE,
+"Force all rep chars the same");
+void output_pass(  //Tess output pass //send to api
+                 PAGE_RES_IT &page_res_it,
+                 BOOL8 write_to_shm);
+void write_results(                           //output a word
+                   PAGE_RES_IT &page_res_it,  //full info
+                   char newline_type,         //type of newline
+                   BOOL8 force_eol,           //override tilde crunch?
+                   BOOL8 write_to_shm         //send to api
+                  );
+WERD_CHOICE *make_epaper_choice(                   //convert one word
+                                WERD_RES *word,    //word to do
+                                char newline_type  //type of newline
+                               );
+INT16 make_reject (              //make reject code
+BOX * inset_box,                 //bounding box
+INT16 prevright,                 //previous char
+INT16 nextleft,                  //next char
+DENORM * denorm,                 //de-normalizer
+char word_string[]               //output string
+);
+char determine_newline_type(                   //test line ends
+                            WERD *word,        //word to do
+                            BLOCK *block,      //current block
+                            WERD *next_word,   //next word
+                            BLOCK *next_block  //block of next word
+                           );
+void write_cooked_text(                     //write output
+                       WERD *word,          //word to do
+                       const STRING &text,  //text to write
+                       BOOL8 acceptable,    //good stuff
+                       BOOL8 pass2,         //done on pass2
+                       FILE *fp             //file to write
+                      );
+void write_shm_text(                    //write output
+                    WERD_RES *word,     //word to do
+                    BLOCK *block,       //block it is from
+                    ROW_RES *row,       //row it is from
+                    const STRING &text  //text to write
+                   );
+void write_map(                //output a map file
+               FILE *mapfile,  //mapfile to write to
+               WERD_RES *word);
+FILE *open_outfile(  //open .map & .unlv file
+                   const char *extension);
+void write_unlv_text(WERD_RES *word); 
+char get_rep_char(  // what char is repeated?
+                  WERD_RES *word);
+void ensure_rep_chars_are_consistent(WERD_RES *word); 
+void set_unlv_suspects(WERD_RES *word); 
+INT16 count_alphas(  //how many alphas
+                   const char *s);
+INT16 count_alphanums(  //how many alphanums
+                      const char *s);
+BOOL8 acceptable_number_string(const char *s); 
+#endif

ccmain/paircmp.cpp

@@ -0,0 +1,107 @@
+/**********************************************************************
+ * File:        paircmp.cpp  (Formerly paircmp.c)
+ * Description: Code to compare two blobs using the adaptive matcher
+ * Author:		Ray Smith
+ * Created:		Wed Apr 21 09:31:02 BST 1993
+ *
+ * (C) Copyright 1993, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          "blobcmp.h"
+#include          "tfacep.h"
+#include          "paircmp.h"
+
+#define EXTERN
+
+/**********************************************************************
+ * compare_blob_pairs
+ *
+ * A blob processor to compare pairs of selected blobs.
+ **********************************************************************/
+
+BOOL8 compare_blob_pairs(             //blob processor
+                         BLOCK *,
+                         ROW *row,    //row it came from
+                         WERD *,
+                         PBLOB *blob  //blob to compare
+                        ) {
+  static ROW *prev_row = NULL;   //other in pair
+  static PBLOB *prev_blob = NULL;
+  float rating;                  //from matcher
+
+  if (prev_row == NULL || prev_blob == NULL) {
+    prev_row = row;
+    prev_blob = blob;
+  }
+  else {
+    rating = compare_blobs (prev_blob, prev_row, blob, row);
+    tprintf ("Rating=%g\n", rating);
+    prev_row = NULL;
+    prev_blob = NULL;
+  }
+  return TRUE;
+}
+
+
+/**********************************************************************
+ * compare_blobs
+ *
+ * Compare 2 blobs and return the rating.
+ **********************************************************************/
+
+float compare_blobs(               //match 2 blobs
+                    PBLOB *blob1,  //first blob
+                    ROW *row1,     //row it came from
+                    PBLOB *blob2,  //other blob
+                    ROW *row2) {
+  PBLOB *bn_blob1;               //baseline norm
+  PBLOB *bn_blob2;
+  DENORM denorm1, denorm2;
+  float rating;                  //match result
+
+  bn_blob1 = blob1->baseline_normalise (row1, &denorm1);
+  bn_blob2 = blob2->baseline_normalise (row2, &denorm2);
+  rating = compare_bln_blobs (bn_blob1, &denorm1, bn_blob2, &denorm2);
+  delete bn_blob1;
+  delete bn_blob2;
+  return rating;
+}
+
+
+/**********************************************************************
+ * compare_bln_blobs
+ *
+ * Compare 2 baseline normalised blobs and return the rating.
+ **********************************************************************/
+
+float compare_bln_blobs(               //match 2 blobs
+                        PBLOB *blob1,  //first blob
+                        DENORM *denorm1,
+                        PBLOB *blob2,  //other blob
+                        DENORM *denorm2) {
+  TBLOB *tblob1;                 //tessblobs
+  TBLOB *tblob2;
+  TEXTROW tessrow1, tessrow2;    //tess rows
+  float rating;                  //match result
+
+  tblob1 = make_tess_blob (blob1, TRUE);
+  make_tess_row(denorm1, &tessrow1); 
+  tblob2 = make_tess_blob (blob2, TRUE);
+  make_tess_row(denorm2, &tessrow2); 
+  rating = compare_tess_blobs (tblob1, &tessrow1, tblob2, &tessrow2);
+  free_blob(tblob1); 
+  free_blob(tblob2); 
+
+  return rating;
+}

ccmain/paircmp.h

@@ -0,0 +1,43 @@
+/**********************************************************************
+ * File:        paircmp.h  (Formerly paircmp.h)
+ * Description: Code to compare two blobs using the adaptive matcher
+ * Author:		Ray Smith
+ * Created:		Wed Apr 21 09:31:02 BST 1993
+ *
+ * (C) Copyright 1993, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           PAIRCMP_H
+#define           PAIRCMP_H
+
+#include          "ocrblock.h"
+#include          "varable.h"
+#include          "notdll.h"
+
+BOOL8 compare_blob_pairs(             //blob processor
+                         BLOCK *,
+                         ROW *row,    //row it came from
+                         WERD *,
+                         PBLOB *blob  //blob to compare
+                        );
+float compare_blobs(               //match 2 blobs
+                    PBLOB *blob1,  //first blob
+                    ROW *row1,     //row it came from
+                    PBLOB *blob2,  //other blob
+                    ROW *row2);
+float compare_bln_blobs(               //match 2 blobs
+                        PBLOB *blob1,  //first blob
+                        DENORM *denorm1,
+                        PBLOB *blob2,  //other blob
+                        DENORM *denorm2);
+#endif

ccmain/reject.h

@@ -0,0 +1,175 @@
+/**********************************************************************
+ * File:        reject.h  (Formerly reject.h)
+ * Description: Rejection functions used in tessedit
+ * Author:		Phil Cheatle
+ * Created:		Wed Sep 23 16:50:21 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           REJECT_H
+#define           REJECT_H
+
+#include          "varable.h"
+#include          "pageres.h"
+#include          "notdll.h"
+
+extern INT_VAR_H (tessedit_reject_mode, 5, "Rejection algorithm");
+extern INT_VAR_H (tessedit_ok_mode, 5, "Acceptance decision algorithm");
+extern BOOL_VAR_H (tessedit_use_nn, TRUE, "");
+extern BOOL_VAR_H (tessedit_rejection_debug, FALSE, "Adaption debug");
+extern BOOL_VAR_H (tessedit_rejection_stats, FALSE, "Show NN stats");
+extern BOOL_VAR_H (tessedit_flip_0O, TRUE, "Contextual 0O O0 flips");
+extern double_VAR_H (tessedit_lower_flip_hyphen, 1.5,
+"Aspect ratio dot/hyphen test");
+extern double_VAR_H (tessedit_upper_flip_hyphen, 1.8,
+"Aspect ratio dot/hyphen test");
+extern BOOL_VAR_H (rej_trust_doc_dawg, FALSE,
+"Use DOC dawg in 11l conf. detector");
+extern BOOL_VAR_H (rej_1Il_use_dict_word, FALSE, "Use dictword test");
+extern BOOL_VAR_H (rej_1Il_trust_permuter_type, TRUE, "Dont double check");
+extern BOOL_VAR_H (one_ell_conflict_default, TRUE,
+"one_ell_conflict default");
+extern BOOL_VAR_H (show_char_clipping, FALSE, "Show clip image window?");
+extern BOOL_VAR_H (nn_debug, FALSE, "NN DEBUGGING?");
+extern BOOL_VAR_H (nn_reject_debug, FALSE, "NN DEBUG each char?");
+extern BOOL_VAR_H (nn_lax, FALSE, "Use 2nd rate matches");
+extern BOOL_VAR_H (nn_double_check_dict, FALSE, "Double check");
+extern BOOL_VAR_H (nn_conf_double_check_dict, TRUE,
+"Double check for confusions");
+extern BOOL_VAR_H (nn_conf_1Il, TRUE, "NN use 1Il conflicts");
+extern BOOL_VAR_H (nn_conf_Ss, TRUE, "NN use Ss conflicts");
+extern BOOL_VAR_H (nn_conf_hyphen, TRUE, "NN hyphen conflicts");
+extern BOOL_VAR_H (nn_conf_test_good_qual, FALSE, "NN dodgy 1Il cross check");
+extern BOOL_VAR_H (nn_conf_test_dict, TRUE, "NN dodgy 1Il cross check");
+extern BOOL_VAR_H (nn_conf_test_sensible, TRUE, "NN dodgy 1Il cross check");
+extern BOOL_VAR_H (nn_conf_strict_on_dodgy_chs, TRUE,
+"Require stronger NN match");
+extern double_VAR_H (nn_dodgy_char_threshold, 0.99, "min accept score");
+extern INT_VAR_H (nn_conf_accept_level, 4, "NN accept dodgy 1Il matches? ");
+extern INT_VAR_H (nn_conf_initial_i_level, 3,
+"NN accept initial Ii match level ");
+extern BOOL_VAR_H (no_unrej_dubious_chars, TRUE,
+"Dubious chars next to reject?");
+extern BOOL_VAR_H (no_unrej_no_alphanum_wds, TRUE,
+"Stop unrej of non A/N wds?");
+extern BOOL_VAR_H (no_unrej_1Il, FALSE, "Stop unrej of 1Ilchars?");
+extern BOOL_VAR_H (rej_use_tess_accepted, TRUE,
+"Individual rejection control");
+extern BOOL_VAR_H (rej_use_tess_blanks, TRUE, "Individual rejection control");
+extern BOOL_VAR_H (rej_use_good_perm, TRUE, "Individual rejection control");
+extern BOOL_VAR_H (rej_use_sensible_wd, FALSE, "Extend permuter check");
+extern BOOL_VAR_H (rej_alphas_in_number_perm, FALSE, "Extend permuter check");
+extern double_VAR_H (rej_whole_of_mostly_reject_word_fract, 0.85,
+"if >this fract");
+extern INT_VAR_H (rej_mostly_reject_mode, 1,
+"0-never, 1-afterNN, 2-after new xht");
+extern double_VAR_H (tessed_fullstop_aspect_ratio, 1.2,
+"if >this fract then reject");
+extern INT_VAR_H (net_image_width, 40, "NN input image width");
+extern INT_VAR_H (net_image_height, 36, "NN input image height");
+extern INT_VAR_H (net_image_x_height, 22, "NN input image x_height");
+extern INT_VAR_H (tessedit_image_border, 2, "Rej blbs near image edge limit");
+extern INT_VAR_H (net_bl_nodes, 20, "Number of baseline nodes");
+extern double_VAR_H (nn_reject_threshold, 0.5, "NN min accept score");
+extern double_VAR_H (nn_reject_head_and_shoulders, 0.6,
+"top scores sep factor");
+extern STRING_VAR_H (ok_single_ch_non_alphanum_wds, "-?\075",
+"Allow NN to unrej");
+extern STRING_VAR_H (ok_repeated_ch_non_alphanum_wds, "-?*\075",
+"Allow NN to unrej");
+extern STRING_VAR_H (conflict_set_I_l_1, "Il1[]", "Il1 conflict set");
+extern STRING_VAR_H (conflict_set_S_s, "Ss$", "Ss conflict set");
+extern STRING_VAR_H (conflict_set_hyphen, "-_~", "hyphen conflict set");
+extern STRING_VAR_H (dubious_chars_left_of_reject, "!'+`()-./\\<>;:^_,~\"",
+"Unreliable chars");
+extern STRING_VAR_H (dubious_chars_right_of_reject, "!'+`()-./\\<>;:^_,~\"",
+"Unreliable chars");
+extern INT_VAR_H (min_sane_x_ht_pixels, 8,
+"Reject any x-ht lt or eq than this");
+void set_done(  //set done flag
+              WERD_RES *word,
+              INT16 pass);
+void make_reject_map(            //make rej map for wd //detailed results
+                     WERD_RES *word,
+                     BLOB_CHOICE_LIST_CLIST *blob_choices,
+                     ROW *row,
+                     INT16 pass  //1st or 2nd?
+                    );
+void reject_blanks(WERD_RES *word); 
+void reject_I_1_L(WERD_RES *word); 
+                                 //detailed results
+void reject_poor_matches(WERD_RES *word, BLOB_CHOICE_LIST_CLIST *blob_choices); 
+float compute_reject_threshold(  //compute threshold //detailed results
+                               BLOB_CHOICE_LIST_CLIST *blob_choices);
+int sort_floats(                   //qsort function
+                const void *arg1,  //ptrs to floats
+                const void *arg2);
+void reject_edge_blobs(WERD_RES *word); 
+BOOL8 one_ell_conflict(WERD_RES *word_res, BOOL8 update_map); 
+INT16 first_alphanum_pos(const char *word); 
+INT16 alpha_count(const char *word); 
+BOOL8 word_contains_non_1_digit(const char *word); 
+BOOL8 test_ambig_word(  //test for ambiguity
+                      WERD_RES *word);
+                                 //original word
+BOOL8 ambig_word(const char *start_word,
+                 char *temp_word,     //alterable copy
+                 INT16 test_char_pos  //idx to char to alter
+                );
+const char *char_ambiguities(char c); 
+
+#ifndef EMBEDDED
+void test_ambigs(const char *word);
+#endif
+ 
+void nn_recover_rejects(WERD_RES *word, ROW *row); 
+void nn_match_word(  //Match a word
+                   WERD_RES *word,
+                   ROW *row);
+                                 //of character
+INT16 nn_match_char(IMAGE &scaled_image,
+                    float baseline_pos,       //rel to scaled_image
+                    BOOL8 dict_word,          //part of dict wd?
+                    BOOL8 checked_dict_word,  //part of dict wd?
+                    BOOL8 sensible_word,      //part acceptable str?
+                    BOOL8 centre,             //not at word ends?
+                    BOOL8 good_quality_word,  //initial segmentation
+                    char tess_ch              //confirm this?
+                   );
+INT16 evaluate_net_match(char top,
+                         float top_score,
+                         char next,
+                         float next_score,
+                         char tess_ch,
+                         BOOL8 dict_word,
+                         BOOL8 checked_dict_word,
+                         BOOL8 sensible_word,
+                         BOOL8 centre,
+                         BOOL8 good_quality_word);
+void dont_allow_dubious_chars(WERD_RES *word); 
+
+void dont_allow_1Il(WERD_RES *word); 
+
+INT16 count_alphanums(  //how many alphanums
+                      WERD_RES *word);
+void reject_mostly_rejects(  //rej all if most rejectd
+                           WERD_RES *word);
+BOOL8 repeated_nonalphanum_wd(WERD_RES *word, ROW *row); 
+BOOL8 repeated_ch_string(const char *rep_ch_str); 
+INT16 safe_dict_word(const char *s); 
+void flip_hyphens(WERD_RES *word); 
+void flip_0O(WERD_RES *word); 
+BOOL8 non_O_upper(char c); 
+BOOL8 non_0_digit(char c); 
+#endif

ccmain/scaleimg.cpp

@@ -0,0 +1,366 @@
+/**********************************************************************
+ * File:        scaleimg.cpp  (Formerly scaleim.c)
+ * Description: Smart scaling of images.
+ * Author:      Phil Cheatle
+ * Created:     Wed Nov 18 16:12:03 GMT 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+/*************************************************************************
+ * This is really Sheelagh's code that I've hacked into a more usable form.
+ * You simply call scale_image() passing in source and target images. The target
+ * image should be empty, but created - in order to define the destination
+ * size.
+ *************************************************************************/
+
+#include "mfcpch.h"
+#include          <stdlib.h>
+#include          <string.h>
+#include          "fileerr.h"
+#include          "tprintf.h"
+#include          "grphics.h"
+#include          "img.h"
+//#include                                      "basefile.h"
+#include          "imgscale.h"
+#include          "scaleimg.h"
+
+void scale_image(                     //scale an image
+                 IMAGE &image,        //source image
+                 IMAGE &target_image  //target image
+                ) {
+  INT32 xsize, ysize, new_xsize, new_ysize;
+  IMAGELINE line, new_line;
+  int *hires, *lores, *oldhires, *oldlores;
+  int i, j, n, oldn, row, col;
+  int offset = 0;                //not used here
+  float factor;
+  UINT8 curr_colour, new_colour;
+  int dummy = -1;
+  IMAGE image2;                  //horiz scaled image
+
+  xsize = image.get_xsize ();
+  ysize = image.get_ysize ();
+  new_xsize = target_image.get_xsize ();
+  new_ysize = target_image.get_ysize ();
+  if (new_ysize > new_xsize)
+    new_line.init (new_ysize);
+  else
+    new_line.init (new_xsize);
+
+  factor = (float) xsize / (float) new_xsize;
+
+  hires = (int *) calloc (xsize, sizeof (int));
+  lores = (int *) calloc (new_xsize, sizeof (int));
+  oldhires = (int *) calloc (xsize, sizeof (int));
+  oldlores = (int *) calloc (new_xsize, sizeof (int));
+  if ((hires == NULL) || (lores == NULL) || (oldhires == NULL)
+  || (oldlores == NULL)) {
+    fprintf (stderr, "Calloc error in scale_image\n");
+    err_exit(); 
+  }
+
+  image2.create (new_xsize, ysize, image.get_bpp ());
+
+  oldn = 0;
+  /* do first row separately because hires[col-1] doesn't make sense here */
+  image.fast_get_line (0, 0, xsize, &line);
+  /* each line nominally begins with white */
+  curr_colour = 1;
+  n = 0;
+  for (i = 0; i < xsize; i++) {
+    new_colour = *(line.pixels + i);
+    if (new_colour != curr_colour) {
+      hires[n] = i;
+      n++;
+      curr_colour = new_colour;
+    }
+  }
+  if (offset != 0)
+    for (i = 0; i < n; i++)
+      hires[i] += offset;
+
+  if (n > new_xsize) {
+    tprintf ("Too many transitions (%d) on line 0\n", n);
+    scale_image_cop_out(image,
+                        target_image,
+                        factor,
+                        hires,
+                        lores,
+                        oldhires,
+                        oldlores);
+    return;
+  }
+  else if (n > 0)
+    dyn_prog (n, hires, lores, new_xsize, &dummy, &dummy, 0, factor);
+  else
+    lores[0] = new_xsize;
+
+  curr_colour = 1;
+  j = 0;
+  for (i = 0; i < new_xsize; i++) {
+    if (lores[j] == i) {
+      curr_colour = 1 - curr_colour;
+      j++;
+    }
+    *(new_line.pixels + i) = curr_colour;
+  }
+  image2.put_line (0, 0, new_xsize, &new_line, 0);
+
+  for (i = 0; i < n; i++) {
+    oldhires[i] = hires[i];
+    oldlores[i] = lores[i];
+  }
+
+  for (i = n; i < oldn; i++) {
+    oldhires[i] = 0;
+    oldlores[i] = 0;
+  }
+  oldn = n;
+
+  for (row = 1; row < ysize; row++) {
+    image.fast_get_line (0, row, xsize, &line);
+    /* each line nominally begins with white */
+    curr_colour = 1;
+    n = 0;
+    for (i = 0; i < xsize; i++) {
+      new_colour = *(line.pixels + i);
+      if (new_colour != curr_colour) {
+        hires[n] = i;
+        n++;
+        curr_colour = new_colour;
+      }
+    }
+    for (i = n; i < oldn; i++) {
+      hires[i] = 0;
+      lores[i] = 0;
+    }
+    if (offset != 0)
+      for (i = 0; i < n; i++)
+        hires[i] += offset;
+
+    if (n > new_xsize) {
+      tprintf ("Too many transitions (%d) on line %d\n", n, row);
+      scale_image_cop_out(image,
+                          target_image,
+                          factor,
+                          hires,
+                          lores,
+                          oldhires,
+                          oldlores);
+      return;
+    }
+    else if (n > 0)
+      dyn_prog(n, hires, lores, new_xsize, oldhires, oldlores, oldn, factor); 
+    else
+      lores[0] = new_xsize;
+
+    curr_colour = 1;
+    j = 0;
+    for (i = 0; i < new_xsize; i++) {
+      if (lores[j] == i) {
+        curr_colour = 1 - curr_colour;
+        j++;
+      }
+      *(new_line.pixels + i) = curr_colour;
+    }
+    image2.put_line (0, row, new_xsize, &new_line, 0);
+
+    for (i = 0; i < n; i++) {
+      oldhires[i] = hires[i];
+      oldlores[i] = lores[i];
+    }
+    for (i = n; i < oldn; i++) {
+      oldhires[i] = 0;
+      oldlores[i] = 0;
+    }
+    oldn = n;
+  }
+
+  free(hires); 
+  free(lores); 
+  free(oldhires); 
+  free(oldlores); 
+
+  /* NOW DO THE VERTICAL SCALING from image2 to target_image*/
+
+  xsize = new_xsize;
+  factor = (float) ysize / (float) new_ysize;
+  offset = 0;
+
+  hires = (int *) calloc (ysize, sizeof (int));
+  lores = (int *) calloc (new_ysize, sizeof (int));
+  oldhires = (int *) calloc (ysize, sizeof (int));
+  oldlores = (int *) calloc (new_ysize, sizeof (int));
+  if ((hires == NULL) || (lores == NULL) || (oldhires == NULL)
+  || (oldlores == NULL)) {
+    fprintf (stderr, "Calloc error in scale_image (vert)\n");
+    err_exit(); 
+  }
+
+  oldn = 0;
+  /* do first col separately because hires[col-1] doesn't make sense here */
+  image2.get_column (0, 0, ysize, &line, 0);
+  /* each line nominally begins with white */
+  curr_colour = 1;
+  n = 0;
+  for (i = 0; i < ysize; i++) {
+    new_colour = *(line.pixels + i);
+    if (new_colour != curr_colour) {
+      hires[n] = i;
+      n++;
+      curr_colour = new_colour;
+    }
+  }
+
+  if (offset != 0)
+    for (i = 0; i < n; i++)
+      hires[i] += offset;
+
+  if (n > new_ysize) {
+    tprintf ("Too many transitions (%d) on column 0\n", n);
+    scale_image_cop_out(image,
+                        target_image,
+                        factor,
+                        hires,
+                        lores,
+                        oldhires,
+                        oldlores);
+    return;
+  }
+  else if (n > 0)
+    dyn_prog (n, hires, lores, new_ysize, &dummy, &dummy, 0, factor);
+  else
+    lores[0] = new_ysize;
+
+  curr_colour = 1;
+  j = 0;
+  for (i = 0; i < new_ysize; i++) {
+    if (lores[j] == i) {
+      curr_colour = 1 - curr_colour;
+      j++;
+    }
+    *(new_line.pixels + i) = curr_colour;
+  }
+  target_image.put_column (0, 0, new_ysize, &new_line, 0);
+
+  for (i = 0; i < n; i++) {
+    oldhires[i] = hires[i];
+    oldlores[i] = lores[i];
+  }
+  for (i = n; i < oldn; i++) {
+    oldhires[i] = 0;
+    oldlores[i] = 0;
+  }
+  oldn = n;
+
+  for (col = 1; col < xsize; col++) {
+    image2.get_column (col, 0, ysize, &line, 0);
+    /* each line nominally begins with white */
+    curr_colour = 1;
+    n = 0;
+    for (i = 0; i < ysize; i++) {
+      new_colour = *(line.pixels + i);
+      if (new_colour != curr_colour) {
+        hires[n] = i;
+        n++;
+        curr_colour = new_colour;
+      }
+    }
+    for (i = n; i < oldn; i++) {
+      hires[i] = 0;
+      lores[i] = 0;
+    }
+
+    if (offset != 0)
+      for (i = 0; i < n; i++)
+        hires[i] += offset;
+
+    if (n > new_ysize) {
+      tprintf ("Too many transitions (%d) on column %d\n", n, col);
+      scale_image_cop_out(image,
+                          target_image,
+                          factor,
+                          hires,
+                          lores,
+                          oldhires,
+                          oldlores);
+      return;
+    }
+    else if (n > 0)
+      dyn_prog(n, hires, lores, new_ysize, oldhires, oldlores, oldn, factor); 
+    else
+      lores[0] = new_ysize;
+
+    curr_colour = 1;
+    j = 0;
+    for (i = 0; i < new_ysize; i++) {
+      if (lores[j] == i) {
+        curr_colour = 1 - curr_colour;
+        j++;
+      }
+      *(new_line.pixels + i) = curr_colour;
+    }
+    target_image.put_column (col, 0, new_ysize, &new_line, 0);
+
+    for (i = 0; i < n; i++) {
+      oldhires[i] = hires[i];
+      oldlores[i] = lores[i];
+    }
+    for (i = n; i < oldn; i++) {
+      oldhires[i] = 0;
+      oldlores[i] = 0;
+    }
+    oldn = n;
+  }
+  free(hires); 
+  free(lores); 
+  free(oldhires); 
+  free(oldlores); 
+}
+
+
+/**********************************************************************
+ * scale_image_cop_out
+ *
+ * Cop-out of scale_image by doing it the easy way and free the data.
+ **********************************************************************/
+
+void scale_image_cop_out(                      //scale an image
+                         IMAGE &image,         //source image
+                         IMAGE &target_image,  //target image
+                         float factor,         //scale factor
+                         int *hires,
+                         int *lores,
+                         int *oldhires,
+                         int *oldlores) {
+  INT32 xsize, ysize, new_xsize, new_ysize;
+
+  xsize = image.get_xsize ();
+  ysize = image.get_ysize ();
+  new_xsize = target_image.get_xsize ();
+  new_ysize = target_image.get_ysize ();
+
+  if (factor <= 0.5)
+    reduce_sub_image (&image, 0, 0, xsize, ysize,
+      &target_image, 0, 0, (INT32) (1.0 / factor), FALSE);
+  else if (factor >= 2)
+    enlarge_sub_image (&image, 0, 0, &target_image,
+        0, 0, new_xsize, new_ysize, (INT32) factor, FALSE);
+  else
+    copy_sub_image (&image, 0, 0, xsize, ysize, &target_image, 0, 0, FALSE);
+  free(hires); 
+  free(lores); 
+  free(oldhires); 
+  free(oldlores); 
+}

ccmain/scaleimg.h

@@ -0,0 +1,35 @@
+/**********************************************************************
+ * File:        scaleimg.h  (Formerly scaleim.h)
+ * Description: Smart scaling of images.
+ * Author:      Phil Cheatle
+ * Created:     Wed Nov 18 16:12:03 GMT 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           SCALEIMG_H
+#define           SCALEIMG_H
+
+void scale_image(                     //scale an image
+                 IMAGE &image,        //source image
+                 IMAGE &target_image  //target image
+                );
+void scale_image_cop_out(                      //scale an image
+                         IMAGE &image,         //source image
+                         IMAGE &target_image,  //target image
+                         float factor,         //scale factor
+                         int *hires,
+                         int *lores,
+                         int *oldhires,
+                         int *oldlores);
+#endif

ccmain/tessbox.cpp

@@ -0,0 +1,370 @@
+/**********************************************************************
+ * File:        tessbox.cpp  (Formerly tessbox.c)
+ * Description: Black boxed Tess for developing a resaljet.
+ * Author:					Ray Smith
+ * Created:					Thu Apr 23 11:03:36 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          "tfacep.h"
+#include          "tfacepp.h"
+#include          "tessbox.h"
+#include "mfoutline.h"
+
+#define EXTERN
+
+/**********************************************************************
+ * tess_segment_pass1
+ *
+ * Segment a word using the pass1 conditions of the tess segmenter.
+ **********************************************************************/
+
+WERD_CHOICE *tess_segment_pass1(                           //recog one word
+                                WERD *word,                //bln word to do
+                                DENORM *denorm,            //de-normaliser
+                                POLY_MATCHER matcher,      //matcher function
+                                WERD_CHOICE *&raw_choice,  //raw result //list of blob lists
+                                BLOB_CHOICE_LIST_CLIST *blob_choices,
+                                WERD *&outword             //bln word output
+                               ) {
+  WERD_CHOICE *result;           //return value
+  int saved_enable_assoc = 0;
+  int saved_chop_enable = 0;
+
+  if (word->flag (W_DONT_CHOP)) {
+    saved_enable_assoc = enable_assoc;
+    saved_chop_enable = chop_enable;
+    enable_assoc = 0;
+    chop_enable = 0;
+    if (word->flag (W_REP_CHAR))
+      permute_only_top = 1;
+  }
+  set_pass1();
+  //      tprintf("pass1 chop on=%d, seg=%d, onlytop=%d",chop_enable,enable_assoc,permute_only_top);
+  result = recog_word (word, denorm, matcher, NULL, NULL, FALSE,
+    raw_choice, blob_choices, outword);
+  if (word->flag (W_DONT_CHOP)) {
+    enable_assoc = saved_enable_assoc;
+    chop_enable = saved_chop_enable;
+    permute_only_top = 0;
+  }
+  return result;
+}
+
+
+/**********************************************************************
+ * tess_segment_pass2
+ *
+ * Segment a word using the pass2 conditions of the tess segmenter.
+ **********************************************************************/
+
+WERD_CHOICE *tess_segment_pass2(                           //recog one word
+                                WERD *word,                //bln word to do
+                                DENORM *denorm,            //de-normaliser
+                                POLY_MATCHER matcher,      //matcher function
+                                WERD_CHOICE *&raw_choice,  //raw result //list of blob lists
+                                BLOB_CHOICE_LIST_CLIST *blob_choices,
+                                WERD *&outword             //bln word output
+                               ) {
+  WERD_CHOICE *result;           //return value
+  int saved_enable_assoc = 0;
+  int saved_chop_enable = 0;
+
+  if (word->flag (W_DONT_CHOP)) {
+    saved_enable_assoc = enable_assoc;
+    saved_chop_enable = chop_enable;
+    enable_assoc = 0;
+    chop_enable = 0;
+    if (word->flag (W_REP_CHAR))
+      permute_only_top = 1;
+  }
+  set_pass2();
+  result = recog_word (word, denorm, matcher, NULL, NULL, FALSE,
+    raw_choice, blob_choices, outword);
+  if (word->flag (W_DONT_CHOP)) {
+    enable_assoc = saved_enable_assoc;
+    chop_enable = saved_chop_enable;
+    permute_only_top = 0;
+  }
+  return result;
+}
+
+
+/**********************************************************************
+ * correct_segment_pass2
+ *
+ * Segment a word correctly using the pass2 conditions of the tess segmenter.
+ * Then call the tester with all the correctly segmented blobs.
+ * If the correct segmentation cannot be found, the tester is called
+ * with the segmentation found by tess and all the correct flags set to
+ * false and all strings are NULL.
+ **********************************************************************/
+
+WERD_CHOICE *correct_segment_pass2(                           //recog one word
+                                   WERD *word,                //bln word to do
+                                   DENORM *denorm,            //de-normaliser
+                                   POLY_MATCHER matcher,      //matcher function
+                                   POLY_TESTER tester,        //tester function
+                                   WERD_CHOICE *&raw_choice,  //raw result //list of blob lists
+                                   BLOB_CHOICE_LIST_CLIST *blob_choices,
+                                   WERD *&outword             //bln word output
+                                  ) {
+  set_pass2();
+  return recog_word (word, denorm, matcher, NULL, tester, TRUE,
+    raw_choice, blob_choices, outword);
+}
+
+
+/**********************************************************************
+ * test_segment_pass2
+ *
+ * Segment a word correctly using the pass2 conditions of the tess segmenter.
+ * Then call the tester on all words used by tess in its search.
+ * Do this only on words where the correct segmentation could be found.
+ **********************************************************************/
+
+WERD_CHOICE *test_segment_pass2(                           //recog one word
+                                WERD *word,                //bln word to do
+                                DENORM *denorm,            //de-normaliser
+                                POLY_MATCHER matcher,      //matcher function
+                                POLY_TESTER tester,        //tester function
+                                WERD_CHOICE *&raw_choice,  //raw result //list of blob lists
+                                BLOB_CHOICE_LIST_CLIST *blob_choices,
+                                WERD *&outword             //bln word output
+                               ) {
+  set_pass2();
+  return recog_word (word, denorm, matcher, tester, NULL, TRUE,
+    raw_choice, blob_choices, outword);
+}
+
+
+/**********************************************************************
+ * tess_acceptable_word
+ *
+ * Return true if the word is regarded as "good enough".
+ **********************************************************************/
+
+BOOL8 tess_acceptable_word(                           //test acceptability
+                           WERD_CHOICE *word_choice,  //after context
+                           WERD_CHOICE *raw_choice    //before context
+                          ) {
+  A_CHOICE choice;               //after context
+  A_CHOICE tess_raw;             //before
+
+  choice.rating = word_choice->rating ();
+  choice.certainty = word_choice->certainty ();
+  choice.string = (char *) word_choice->string ().string ();
+  tess_raw.rating = raw_choice->rating ();
+  tess_raw.certainty = raw_choice->certainty ();
+  tess_raw.string = (char *) raw_choice->string ().string ();
+                                 //call tess
+  return AcceptableResult (&choice, &tess_raw);
+}
+
+
+/**********************************************************************
+ * tess_adaptable_word
+ *
+ * Return true if the word is regarded as "good enough".
+ **********************************************************************/
+
+BOOL8 tess_adaptable_word(                           //test adaptability
+                          WERD *word,                //word to test
+                          WERD_CHOICE *word_choice,  //after context
+                          WERD_CHOICE *raw_choice    //before context
+                         ) {
+  TWERD *tessword;               //converted word
+  INT32 result;                  //answer
+
+  tessword = make_tess_word (word, NULL);
+  result = AdaptableWord (tessword, word_choice->string ().string (),
+    raw_choice->string ().string ());
+  delete_word(tessword);
+  return result != 0;
+}
+
+
+/**********************************************************************
+ * tess_cn_matcher
+ *
+ * Match a blob using the Tess Char Normalized (non-adaptive) matcher
+ * only.
+ **********************************************************************/
+
+void tess_cn_matcher(                           //call tess
+                     PBLOB *pblob,              //previous blob
+                     PBLOB *blob,               //blob to match
+                     PBLOB *nblob,              //next blob
+                     WERD *word,                //word it came from
+                     DENORM *denorm,            //de-normaliser
+                     BLOB_CHOICE_LIST &ratings  //list of results
+                    ) {
+  LIST result;                   //tess output
+  TBLOB *tessblob;               //converted blob
+  TEXTROW tessrow;               //dummy row
+
+  tess_cn_matching = TRUE;       //turn it on
+  tess_bn_matching = FALSE;
+                                 //convert blob
+  tessblob = make_tess_blob (blob, TRUE);
+                                 //make dummy row
+  make_tess_row(denorm, &tessrow);
+                                 //classify
+  result = AdaptiveClassifier (tessblob, NULL, &tessrow);
+  free_blob(tessblob);
+                                 //make our format
+  convert_choice_list(result, ratings);
+}
+
+
+/**********************************************************************
+ * tess_bn_matcher
+ *
+ * Match a blob using the Tess Baseline Normalized (adaptive) matcher
+ * only.
+ **********************************************************************/
+
+void tess_bn_matcher(                           //call tess
+                     PBLOB *pblob,              //previous blob
+                     PBLOB *blob,               //blob to match
+                     PBLOB *nblob,              //next blob
+                     WERD *word,                //word it came from
+                     DENORM *denorm,            //de-normaliser
+                     BLOB_CHOICE_LIST &ratings  //list of results
+                    ) {
+  LIST result;                   //tess output
+  TBLOB *tessblob;               //converted blob
+  TEXTROW tessrow;               //dummy row
+
+  tess_bn_matching = TRUE;       //turn it on
+  tess_cn_matching = FALSE;
+                                 //convert blob
+  tessblob = make_tess_blob (blob, TRUE);
+                                 //make dummy row
+  make_tess_row(denorm, &tessrow);
+                                 //classify
+  result = AdaptiveClassifier (tessblob, NULL, &tessrow);
+  free_blob(tessblob);
+                                 //make our format
+  convert_choice_list(result, ratings);
+}
+
+
+/**********************************************************************
+ * tess_default_matcher
+ *
+ * Match a blob using the default functionality of the Tess matcher.
+ **********************************************************************/
+
+void tess_default_matcher(                           //call tess
+                          PBLOB *pblob,              //previous blob
+                          PBLOB *blob,               //blob to match
+                          PBLOB *nblob,              //next blob
+                          WERD *word,                //word it came from
+                          DENORM *denorm,            //de-normaliser
+                          BLOB_CHOICE_LIST &ratings  //list of results
+                         ) {
+  LIST result;                   //tess output
+  TBLOB *tessblob;               //converted blob
+  TEXTROW tessrow;               //dummy row
+
+  tess_bn_matching = FALSE;      //turn it off
+  tess_cn_matching = FALSE;
+                                 //convert blob
+  tessblob = make_tess_blob (blob, TRUE);
+                                 //make dummy row
+  make_tess_row(denorm, &tessrow);
+                                 //classify
+  result = AdaptiveClassifier (tessblob, NULL, &tessrow);
+  free_blob(tessblob);
+                                 //make our format
+  convert_choice_list(result, ratings);
+}
+
+
+/**********************************************************************
+ * tess_training_tester
+ *
+ * Matcher tester function which actually trains tess.
+ **********************************************************************/
+
+void tess_training_tester(                           //call tess
+                          PBLOB *blob,               //blob to match
+                          DENORM *denorm,            //de-normaliser
+                          BOOL8 correct,             //ly segmented
+                          char *text,                //correct text
+                          INT32 count,               //chars in text
+                          BLOB_CHOICE_LIST *ratings  //list of results
+                         ) {
+  TBLOB *tessblob;               //converted blob
+  TEXTROW tessrow;               //dummy row
+
+  if (correct) {
+    NormMethod = character;              //Force char norm spc 30/11/93
+    tess_bn_matching = FALSE;    //turn it off
+    tess_cn_matching = FALSE;
+                                 //convert blob
+    tessblob = make_tess_blob (blob, TRUE);
+                                 //make dummy row
+    make_tess_row(denorm, &tessrow);
+                                 //learn it
+    LearnBlob(tessblob, &tessrow, text, count);
+    free_blob(tessblob);
+  }
+}
+
+
+/**********************************************************************
+ * tess_adapter
+ *
+ * Adapt to the word using the Tesseract mechanism.
+ **********************************************************************/
+
+void tess_adapter(                         //adapt to word
+                  WERD *word,              //bln word
+                  DENORM *denorm,          //de-normalise
+                  const char *string,      //string for word
+                  const char *raw_string,  //before context
+                  const char *rejmap       //reject map
+                 ) {
+  TWERD *tessword;               //converted word
+  static TEXTROW tessrow;        //dummy row
+
+                                 //make dummy row
+  make_tess_row(denorm, &tessrow);
+                                 //make a word
+  tessword = make_tess_word (word, &tessrow);
+  AdaptToWord(tessword, &tessrow, string, raw_string, rejmap);
+  //adapt to it
+  delete_word(tessword);  //free it
+}
+
+
+/**********************************************************************
+ * tess_add_doc_word
+ *
+ * Add the given word to the document dictionary
+ **********************************************************************/
+
+void tess_add_doc_word(                          //test acceptability
+                       WERD_CHOICE *word_choice  //after context
+                      ) {
+  A_CHOICE choice;               //after context
+
+  choice.rating = word_choice->rating ();
+  choice.certainty = word_choice->certainty ();
+  choice.string = (char *) word_choice->string ().string ();
+  add_document_word(&choice);
+}

ccmain/tessbox.h

@@ -0,0 +1,110 @@
+/**********************************************************************
+ * File:        tessbox.h  (Formerly tessbox.h)
+ * Description: Black boxed Tess for developing a resaljet.
+ * Author:					Ray Smith
+ * Created:					Thu Apr 23 11:03:36 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           TESSBOX_H
+#define           TESSBOX_H
+
+#include          "ratngs.h"
+#include          "notdll.h"
+
+WERD_CHOICE *tess_segment_pass1(                           //recog one word
+                                WERD *word,                //bln word to do
+                                DENORM *denorm,            //de-normaliser
+                                POLY_MATCHER matcher,      //matcher function
+                                WERD_CHOICE *&raw_choice,  //raw result //list of blob lists
+                                BLOB_CHOICE_LIST_CLIST *blob_choices,
+                                WERD *&outword             //bln word output
+                               );
+WERD_CHOICE *tess_segment_pass2(                           //recog one word
+                                WERD *word,                //bln word to do
+                                DENORM *denorm,            //de-normaliser
+                                POLY_MATCHER matcher,      //matcher function
+                                WERD_CHOICE *&raw_choice,  //raw result //list of blob lists
+                                BLOB_CHOICE_LIST_CLIST *blob_choices,
+                                WERD *&outword             //bln word output
+                               );
+                                 //recog one word
+WERD_CHOICE *correct_segment_pass2(WERD *word,                //bln word to do
+                                   DENORM *denorm,            //de-normaliser
+                                   POLY_MATCHER matcher,      //matcher function
+                                   POLY_TESTER tester,        //tester function
+                                   WERD_CHOICE *&raw_choice,  //raw result //list of blob lists
+                                   BLOB_CHOICE_LIST_CLIST *blob_choices,
+                                   WERD *&outword             //bln word output
+                                  );
+WERD_CHOICE *test_segment_pass2(                           //recog one word
+                                WERD *word,                //bln word to do
+                                DENORM *denorm,            //de-normaliser
+                                POLY_MATCHER matcher,      //matcher function
+                                POLY_TESTER tester,        //tester function
+                                WERD_CHOICE *&raw_choice,  //raw result //list of blob lists
+                                BLOB_CHOICE_LIST_CLIST *blob_choices,
+                                WERD *&outword             //bln word output
+                               );
+BOOL8 tess_acceptable_word(                           //test acceptability
+                           WERD_CHOICE *word_choice,  //after context
+                           WERD_CHOICE *raw_choice    //before context
+                          );
+BOOL8 tess_adaptable_word(                           //test adaptability
+                          WERD *word,                //word to test
+                          WERD_CHOICE *word_choice,  //after context
+                          WERD_CHOICE *raw_choice    //before context
+                         );
+void tess_cn_matcher(                           //call tess
+                     PBLOB *pblob,              //previous blob
+                     PBLOB *blob,               //blob to match
+                     PBLOB *nblob,              //next blob
+                     WERD *word,                //word it came from
+                     DENORM *denorm,            //de-normaliser
+                     BLOB_CHOICE_LIST &ratings  //list of results
+                    );
+void tess_bn_matcher(                           //call tess
+                     PBLOB *pblob,              //previous blob
+                     PBLOB *blob,               //blob to match
+                     PBLOB *nblob,              //next blob
+                     WERD *word,                //word it came from
+                     DENORM *denorm,            //de-normaliser
+                     BLOB_CHOICE_LIST &ratings  //list of results
+                    );
+void tess_default_matcher(                           //call tess
+                          PBLOB *pblob,              //previous blob
+                          PBLOB *blob,               //blob to match
+                          PBLOB *nblob,              //next blob
+                          WERD *word,                //word it came from
+                          DENORM *denorm,            //de-normaliser
+                          BLOB_CHOICE_LIST &ratings  //list of results
+                         );
+void tess_training_tester(                           //call tess
+                          PBLOB *blob,               //blob to match
+                          DENORM *denorm,            //de-normaliser
+                          BOOL8 correct,             //ly segmented
+                          char *text,                //correct text
+                          INT32 count,               //chars in text
+                          BLOB_CHOICE_LIST *ratings  //list of results
+                         );
+void tess_adapter(                         //adapt to word
+                  WERD *word,              //bln word
+                  DENORM *denorm,          //de-normalise
+                  const char *string,      //string for word
+                  const char *raw_string,  //before context
+                  const char *rejmap);
+void tess_add_doc_word(                          //test acceptability
+                       WERD_CHOICE *word_choice  //after context
+                      );
+#endif

ccmain/tessedit.cpp

@@ -0,0 +1,321 @@
+/**********************************************************************
+ * File:        tessedit.cpp  (Formerly tessedit.c)
+ * Description: Main program for merge of tess and editor.
+ * Author:					Ray Smith
+ * Created:					Tue Jan 07 15:21:46 GMT 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+//#include                                                      <osfcn.h>
+//#include                                                      <signal.h>
+//#include                                                      <time.h>
+//#include                                                      <unistd.h>
+#include          "tfacep.h"     //must be before main.h
+//#include                                                      "fileerr.h"
+#include          "stderr.h"
+#include          "basedir.h"
+#include          "tessvars.h"
+//#include                                                      "debgwin.h"
+//#include                                      "epapdest.h"
+#include          "control.h"
+#include          "imgs.h"
+#include          "reject.h"
+#include          "pageres.h"
+//#include                                                      "gpapdest.h"
+#include          "mainblk.h"
+#include          "nwmain.h"
+#include          "pgedit.h"
+#include          "ocrshell.h"
+#include          "tprintf.h"
+//#include                                      "ipeerr.h"
+//#include                                                      "restart.h"
+#include          "tessedit.h"
+//#include                                                      "fontfind.h"
+#include "permute.h"
+#include "permdawg.h"
+#include "permnum.h"
+#include "stopper.h"
+#include "adaptmatch.h"
+#include "intmatcher.h"
+#include "chop.h"
+#include "globals.h"
+
+//extern "C" {
+#include          "callnet.h"    //phils nn stuff
+//}
+#include          "notdll.h"     //phils nn stuff
+
+#define VARDIR        "configs/" /*variables files */
+                                 //config under api
+#define API_CONFIG      "configs/api_config"
+#define EXTERN
+
+EXTERN BOOL_EVAR (tessedit_write_vars, FALSE, "Write all vars to file");
+EXTERN BOOL_VAR (tessedit_tweaking_tess_vars, FALSE,
+"Fiddle tess config values");
+
+EXTERN INT_VAR (tweak_ReliableConfigThreshold, 2, "Tess VAR");
+
+EXTERN double_VAR (tweak_garbage, 1.5, "Tess VAR");
+EXTERN double_VAR (tweak_ok_word, 1.25, "Tess VAR");
+EXTERN double_VAR (tweak_good_word, 1.1, "Tess VAR");
+EXTERN double_VAR (tweak_freq_word, 1.0, "Tess VAR");
+EXTERN double_VAR (tweak_ok_number, 1.4, "Tess VAR");
+EXTERN double_VAR (tweak_good_number, 1.1, "Tess VAR");
+EXTERN double_VAR (tweak_non_word, 1.25, "Tess VAR");
+EXTERN double_VAR (tweak_CertaintyPerChar, -0.5, "Tess VAR");
+EXTERN double_VAR (tweak_NonDictCertainty, -2.5, "Tess VAR");
+EXTERN double_VAR (tweak_RejectCertaintyOffset, 1.0, "Tess VAR");
+EXTERN double_VAR (tweak_GoodAdaptiveMatch, 0.125, "Tess VAR");
+EXTERN double_VAR (tweak_GreatAdaptiveMatch, 0.10, "Tess VAR");
+EXTERN INT_VAR (tweak_AdaptProtoThresh, 230, "Tess VAR");
+EXTERN INT_VAR (tweak_AdaptFeatureThresh, 230, "Tess VAR");
+EXTERN INT_VAR (tweak_min_outline_points, 6, "Tess VAR");
+EXTERN INT_VAR (tweak_min_outline_area, 2000, "Tess VAR");
+EXTERN double_VAR (tweak_good_split, 50.0, "Tess VAR");
+EXTERN double_VAR (tweak_ok_split, 100.0, "Tess VAR");
+
+extern INT16 XOFFSET;
+extern INT16 YOFFSET;
+extern int NO_BLOCK;
+
+                                 //progress monitor
+ETEXT_DESC *global_monitor = NULL;
+
+int init_tesseract(const char *arg0,
+                   const char *textbase,
+                   const char *configfile,
+                   int configc,
+                   const char *const *configv) {
+  FILE *var_file;
+  static char c_path[MAX_PATH];  //path for c code
+
+  // Set the basename, compute the data directory and read C++ configs.
+  main_setup(arg0, textbase, configc, configv);
+  debug_window_on.set_value (FALSE);
+
+  if (tessedit_write_vars) {
+    var_file = fopen ("edited.cfg", "w");
+    if (var_file != NULL) {
+      print_variables(var_file);
+      fclose(var_file);
+    }
+  }
+  strcpy (c_path, datadir.string ());
+  c_path[strlen (c_path) - strlen (m_data_sub_dir.string ())] = '\0';
+  demodir = c_path;
+  start_recog(configfile, textbase);
+
+  ReliableConfigThreshold = tweak_ReliableConfigThreshold;
+
+  set_tess_tweak_vars();
+
+  if (tessedit_use_nn)           //phils nn stuff
+    init_net();
+  return 0;                      //Normal exit
+}
+
+void end_tesseract() {
+  end_recog();
+}
+
+#ifdef _TIFFIO_
+void read_tiff_image(TIFF* tif, IMAGE* image) {
+  tdata_t buf;
+  uint32 image_width, image_height;
+  uint16 photometric;
+  short bpp;
+  TIFFGetField(tif, TIFFTAG_IMAGEWIDTH, &image_width);
+  TIFFGetField(tif, TIFFTAG_IMAGELENGTH, &image_height);
+  TIFFGetField(tif, TIFFTAG_BITSPERSAMPLE, &bpp);
+  TIFFGetField(tif, TIFFTAG_PHOTOMETRIC, &photometric);
+  // Tesseract's internal representation is 0-is-black,
+  // so if the photometric is 1 (min is black) then high-valued pixels
+  // are 1 (white), otherwise they are 0 (black).
+  UINT8 high_value = photometric == 1;
+  image->create(image_width, image_height, bpp);
+  IMAGELINE line;
+  line.init(image_width);
+
+  buf = _TIFFmalloc(TIFFScanlineSize(tif));
+  int bytes_per_line = (image_width*bpp + 7)/8;
+  UINT8* dest_buf = image->get_buffer();
+  // This will go badly wrong with one of the more exotic tiff formats,
+  // but the majority will work OK.
+  for (int y = 0; y < image_height; ++y) {
+    TIFFReadScanline(tif, buf, y);
+    memcpy(dest_buf, buf, bytes_per_line);
+    dest_buf += bytes_per_line;
+  }
+  if (high_value == 0)
+    invert_image(image);
+  _TIFFfree(buf);
+}
+#endif
+
+/* Define command type identifiers */
+
+enum CMD_EVENTS
+{
+  ACTION_1_CMD_EVENT,
+  RECOG_WERDS,
+  RECOG_PSEUDO,
+  ACTION_2_CMD_EVENT
+};
+
+/**********************************************************************
+ *  extend_menu()
+ *
+ *  Function called by pgeditor to let you extend the command menu.
+ *  Items can be added to the "MODES" and "OTHER" menus.  The modes_id_base
+ *  and other_id_base parameters are required to offset your command event ids
+ *  from those of pgeditor, and to let the pgeditor which commands are mode
+ *  changes and which are unmoded commands.  (Sorry if you think these offsets
+ *  are a bit kludgy, the alternative would be to duplicate all the menu
+ *  constructor modes within pgeditor so that the offsets could be hidden.)
+ *
+ *  Items for the "MODES" menu may only be simple menu items (just a name and
+ *  id).  Items for the "OTHER" menu can be editable parameters or boolean
+ *  toggles.  Refer to menu.h to see how to build different types.
+ **********************************************************************/
+
+void extend_menu(                             //handle for "MODES"
+                 RADIO_MENU *modes_menu,
+                 INT16 modes_id_base,         //mode cmd ids offset
+                 NON_RADIO_MENU *other_menu,  //handle for "OTHER"
+                 INT16 other_id_base          //mode cmd ids offset
+                ) {
+  /* Example new mode */
+
+  modes_menu->add_child (new RADIO_MENU_LEAF ("Recog Words",
+    modes_id_base + RECOG_WERDS));
+  modes_menu->add_child (new RADIO_MENU_LEAF ("Recog Blobs",
+    modes_id_base + RECOG_PSEUDO));
+
+  /* Example toggle
+
+  other_menu->add_child(
+    new TOGGLE_MENU_LEAF( "Action 2",					//Display string
+            other_id_base + ACTION_2_CMD_EVENT,	//offset command id
+            FALSE ) );							//Initial value
+
+   Example text parm  (commented out)
+
+    other_menu->add_child(
+    new VARIABLE_MENU_LEAF( "Parm change",				//Display string
+            other_id_base + ACTION_3_CMD_EVENT,	//offset command id
+            "default value" ) );				//default value string
+  */
+}
+
+
+/**********************************************************************
+ *  extend_moded_commands()
+ *
+ * Function called by pgeditor when the user is in one of the extended modes
+ * defined by extend_menu() and the user has selected an area in the image
+ * window.
+ **********************************************************************/
+
+void extend_moded_commands(                   //current mode
+                           INT32 mode,
+                           BOX selection_box  //area selected
+                          ) {
+  char msg[MAX_CHARS + 1];
+
+  switch (mode) {
+    case RECOG_WERDS:
+      command_window->msg ("Recogging selected words");
+
+      /* This is how to apply a "word processor" function to each selected word */
+
+      process_selected_words(current_block_list,
+                             selection_box,
+                             &recog_interactive);
+      break;
+    case RECOG_PSEUDO:
+      command_window->msg ("Recogging selected blobs");
+
+      /* This is how to apply a "word processor" function to each selected word */
+
+      recog_pseudo_word(current_block_list, selection_box);
+      break;
+    default:
+      sprintf (msg, "Unexpected extended mode " INT32FORMAT, mode);
+      command_window->msg (msg);
+  }
+}
+
+
+/**********************************************************************
+ *  extend_unmoded_commands()
+ *
+ * Function called by pgeditor when the user has selected one of the unmoded
+ * extended menu options.
+ **********************************************************************/
+
+void extend_unmoded_commands(                 //current mode
+                             INT32 cmd_event,
+                             char *new_value  //changed value if any
+                            ) {
+  char msg[MAX_CHARS + 1];
+
+  switch (cmd_event) {
+    case ACTION_2_CMD_EVENT:     //a toggle event
+      if (new_value[0] == 'T')
+                                 //Display message
+        command_window->msg ("Extended Action 2 ON!!");
+      else
+        command_window->msg ("Extended Action 2 OFF!!");
+      break;
+    default:
+      sprintf (msg, "Unrecognised extended command " INT32FORMAT " (%s)",
+        cmd_event, new_value);
+      command_window->msg (msg);
+      break;
+  }
+}
+
+
+/*************************************************************************
+ * set_tess_tweak_vars()
+ * Set TESS vars from the tweek value - This is only really of use during search
+ * of the space of tess configs - othertimes the default values are set
+ *
+ *************************************************************************/
+void set_tess_tweak_vars() {
+  if (tessedit_tweaking_tess_vars) {
+    garbage = tweak_garbage;
+    ok_word = tweak_ok_word;
+    good_word = tweak_good_word;
+    freq_word = tweak_freq_word;
+    ok_number = tweak_ok_number;
+    good_number = tweak_good_number;
+    non_word = tweak_non_word;
+    CertaintyPerChar = tweak_CertaintyPerChar;
+    NonDictCertainty = tweak_NonDictCertainty;
+    RejectCertaintyOffset = tweak_RejectCertaintyOffset;
+    GoodAdaptiveMatch = tweak_GoodAdaptiveMatch;
+    GreatAdaptiveMatch = tweak_GreatAdaptiveMatch;
+    AdaptProtoThresh = tweak_AdaptProtoThresh;
+    AdaptFeatureThresh = tweak_AdaptFeatureThresh;
+    min_outline_points = tweak_min_outline_points;
+    min_outline_area = tweak_min_outline_area;
+    good_split = tweak_good_split;
+    ok_split = tweak_ok_split;
+  }
+  //   if (expiry_day * 24 * 60 * 60 < time(NULL))
+  //         err_exit();
+}

ccmain/tessedit.h

@@ -0,0 +1,67 @@
+/**********************************************************************
+ * File:        tessedit.h  (Formerly tessedit.h)
+ * Description: Main program for merge of tess and editor.
+ * Author:		Ray Smith
+ * Created:		Tue Jan 07 15:21:46 GMT 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           TESSEDIT_H
+#define           TESSEDIT_H
+
+#include          "tessclas.h"
+#include          "ocrclass.h"
+#include                    "pgedit.h"
+#include          "notdll.h"
+
+// Includes libtiff if HAVE_LIBTIFF is defined
+#ifdef HAVE_LIBTIFF
+#ifdef GOOGLE3
+#include "third_party/tiff/tiffio.h"
+#else
+#include "tiffio.h"
+#endif
+#endif
+
+                                 //progress monitor
+extern ETEXT_DESC *global_monitor;
+
+int init_tesseract(const char *arg0,
+                   const char *textbase,
+                   const char *configfile,
+                   int configc,
+                   const char *const *configv);
+void recognize_page(STRING& image_name);
+void end_tesseract();
+
+#ifdef _TIFFIO_
+void read_tiff_image(TIFF* tif, IMAGE* image);
+#endif
+
+//handle for "MODES"
+void extend_menu(RADIO_MENU *modes_menu,
+                 INT16 modes_id_base,         //mode cmd ids offset
+                 NON_RADIO_MENU *other_menu,  //handle for "OTHER"
+                 INT16 other_id_base          //mode cmd ids offset
+                );
+                                 //current mode
+void extend_moded_commands(INT32 mode,
+                           BOX selection_box  //area selected
+                          );
+                                 //current mode
+void extend_unmoded_commands(INT32 cmd_event,
+                             char *new_value  //changed value if any
+                            );
+void set_tess_tweak_vars();
+#endif

ccmain/tessembedded.h

@@ -0,0 +1,38 @@
+/**********************************************************************
+ * File:        tessembedded.h 
+ * Description: Access to initialization functions in embedded environment
+ * Author:    Marius Renn
+ * Created:   Sun Oct 21
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef TESSEMBEDDED_H
+#define TESSEMBEDDED_H
+
+#include          "ocrblock.h"
+#include          "varable.h"
+#include          "notdll.h"
+
+int init_tessembedded(const char *arg0,
+                      const char *textbase,
+                      const char *configfile,
+                      int configc,
+                      const char *const *configv);
+                      
+void tessembedded_read_file(STRING &name,
+                            BLOCK_LIST *blocks);
+
+void end_tessembedded();
+
+#endif

ccmain/tesseractmain.cpp

@@ -0,0 +1,311 @@
+/**********************************************************************
+ * File:        tessedit.cpp  (Formerly tessedit.c)
+ * Description: Main program for merge of tess and editor.
+ * Author:					Ray Smith
+ * Created:					Tue Jan 07 15:21:46 GMT 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include "applybox.h"
+#include "control.h"
+#include "tessvars.h"
+#include "tessedit.h"
+#include "baseapi.h"
+#include "pageres.h"
+#include "imgs.h"
+#include "varabled.h"
+#include "tprintf.h"
+#include "tesseractmain.h"
+#include "stderr.h"
+#include "notdll.h"
+#include "mainblk.h"
+#include "globals.h"
+#include "tfacep.h"
+#include "callnet.h"
+
+#define VARDIR        "configs/" /*variables files */
+                                 //config under api
+#define API_CONFIG      "configs/api_config"
+#define EXTERN
+
+EXTERN BOOL_VAR (tessedit_read_image, TRUE, "Ensure the image is read");
+EXTERN BOOL_VAR (tessedit_write_images, FALSE,
+"Capture the image from the IPE");
+EXTERN BOOL_VAR (tessedit_debug_to_screen, FALSE, "Dont use debug file");
+
+extern INT16 XOFFSET;
+extern INT16 YOFFSET;
+extern int NO_BLOCK;
+
+const ERRCODE USAGE = "Usage";
+char szAppName[] = "Tessedit";   //app name
+
+/**********************************************************************
+ *  main()
+ *
+ **********************************************************************/
+
+#ifndef GRAPHICS_DISABLED
+int main(int argc, char **argv) {
+  STRING outfile;               //output file
+
+  if (argc < 3) {
+    USAGE.error (argv[0], EXIT,
+      "%s imagename outputbase [configfile [[+|-]varfile]...]\n", argv[0]);
+  }
+
+  if (argc == 3)
+    TessBaseAPI::Init(argv[0], argv[1], NULL, false, 0, argv + 2);
+  else
+    TessBaseAPI::Init(argv[0], argv[1], argv[3], false, argc - 4, argv + 4);
+
+  tprintf ("Tesseract Open Source OCR Engine\n");
+
+  IMAGE image;
+#ifdef _TIFFIO_
+  TIFF* tif = TIFFOpen(argv[1], "r");
+  if (tif) {
+    read_tiff_image(tif, &image);
+    TIFFClose(tif);
+  } else {
+    READFAILED.error (argv[0], EXIT, argv[1]);
+  }
+#else
+  if (image.read_header(argv[1]) < 0)
+    READFAILED.error (argv[0], EXIT, argv[1]);
+  if (image.read(image.get_ysize ()) < 0) {
+    MEMORY_OUT.error(argv[0], EXIT, "Read of image %s",
+      argv[1]);
+  }
+#endif
+  int bytes_per_line = check_legal_image_size(image.get_xsize(),
+                                              image.get_ysize(),
+                                              image.get_bpp());
+  char* text = TessBaseAPI::TesseractRect(image.get_buffer(), image.get_bpp()/8,
+                                          bytes_per_line, 0, 0,
+                                          image.get_xsize(), image.get_ysize());
+  outfile = argv[2];
+  outfile += ".txt";
+  FILE* fp = fopen(outfile.string(), "w");
+  if (fp != NULL) {
+    fwrite(text, 1, strlen(text), fp);
+    fclose(fp);
+  }
+  delete [] text;
+  TessBaseAPI::End();
+
+  return 0;                      //Normal exit
+}
+#else
+
+int main(int argc, char **argv) {
+  UINT16 lang;                   //language
+  STRING pagefile;               //input file
+
+  if (argc < 4) {
+    USAGE.error (argv[0], EXIT,
+      "%s imagename outputbase configfile [[+|-]varfile]...\n", argv[0]);
+  }
+
+  time_t t_start = time(NULL);
+
+  init_tessembedded (argv[0], argv[2], argv[3], argc - 4, argv + 4);
+
+  tprintf ("Tesseract Open Source OCR Engine (graphics disabled)\n");
+
+  if (tessedit_read_image) {
+#ifdef _TIFFIO_
+    TIFF* tif = TIFFOpen(argv[1], "r");
+    if (tif) {
+      read_tiff_image(tif);
+      TIFFClose(tif);
+    } else
+    READFAILED.error (argv[0], EXIT, argv[1]);
+
+#else
+    if (page_image.read_header (argv[1]) < 0)
+      READFAILED.error (argv[0], EXIT, argv[1]);
+    if (page_image.read (page_image.get_ysize ()) < 0) {
+      MEMORY_OUT.error (argv[0], EXIT, "Read of image %s",
+        argv[1]);
+    }
+#endif
+  }
+
+  pagefile = argv[1];
+
+  BLOCK_LIST current_block_list;
+  tessembedded_read_file(pagefile, &current_block_list);
+  tprintf ("Done reading files.\n");
+
+  PAGE_RES page_res(&current_block_list);
+
+  recog_all_words(&page_res, NULL);
+
+  current_block_list.clear();
+  ResetAdaptiveClassifier();
+
+  time_t t_end = time(NULL);
+  double secs = difftime(t_end, t_start);
+  tprintf ("Done. Number of seconds: %d\n", (int)secs);
+  return 0;                      //Normal exit
+}
+
+#endif
+
+int initialized = 0;
+
+#ifdef __MSW32__
+/**********************************************************************
+ * WinMain
+ *
+ * Main function for a windows program.
+ **********************************************************************/
+
+int WINAPI WinMain(  //main for windows //command line
+                   HINSTANCE hInstance,
+                   HINSTANCE hPrevInstance,
+                   LPSTR lpszCmdLine,
+                   int nCmdShow) {
+  WNDCLASS wc;
+  HWND hwnd;
+  MSG msg;
+
+  char **argv;
+  char *argsin[2];
+  int argc;
+  int exit_code;
+
+  wc.style = CS_NOCLOSE | CS_OWNDC;
+  wc.lpfnWndProc = (WNDPROC) WndProc;
+  wc.cbClsExtra = 0;
+  wc.cbWndExtra = 0;
+  wc.hInstance = hInstance;
+  wc.hIcon = NULL;               //LoadIcon (NULL, IDI_APPLICATION);
+  wc.hCursor = NULL;             //LoadCursor (NULL, IDC_ARROW);
+  wc.hbrBackground = (HBRUSH) (COLOR_WINDOW + 1);
+  wc.lpszMenuName = NULL;
+  wc.lpszClassName = szAppName;
+
+  RegisterClass(&wc);
+
+  hwnd = CreateWindow (szAppName, szAppName,
+    WS_OVERLAPPEDWINDOW | WS_DISABLED,
+    CW_USEDEFAULT, CW_USEDEFAULT, CW_USEDEFAULT,
+    CW_USEDEFAULT, HWND_DESKTOP, NULL, hInstance, NULL);
+
+  argsin[0] = strdup (szAppName);
+  argsin[1] = strdup (lpszCmdLine);
+  /*allocate memory for the args. There can never be more than half*/
+  /*the total number of characters in the arguments.*/
+  argv =
+    (char **) malloc (((strlen (argsin[0]) + strlen (argsin[1])) / 2 + 1) *
+    sizeof (char *));
+
+  /*now construct argv as it should be for C.*/
+  argc = parse_args (2, argsin, argv);
+
+  //  ShowWindow (hwnd, nCmdShow);
+  //  UpdateWindow (hwnd);
+
+  if (initialized) {
+    exit_code = main (argc, argv);
+    free (argsin[0]);
+    free (argsin[1]);
+    free(argv);
+    return exit_code;
+  }
+  while (GetMessage (&msg, NULL, 0, 0)) {
+    TranslateMessage(&msg);
+    DispatchMessage(&msg);
+    if (initialized) {
+      exit_code = main (argc, argv);
+      break;
+    }
+    else
+      exit_code = msg.wParam;
+  }
+  free (argsin[0]);
+  free (argsin[1]);
+  free(argv);
+  return exit_code;
+}
+
+
+/**********************************************************************
+ * WndProc
+ *
+ * Function to respond to messages.
+ **********************************************************************/
+
+LONG WINAPI WndProc(            //message handler
+                    HWND hwnd,  //window with message
+                    UINT msg,   //message typ
+                    WPARAM wParam,
+                    LPARAM lParam) {
+  HDC hdc;
+
+  if (msg == WM_CREATE) {
+    //
+    // Create a rendering context.
+    //
+    hdc = GetDC (hwnd);
+    ReleaseDC(hwnd, hdc);
+    initialized = 1;
+    return 0;
+  }
+  return DefWindowProc (hwnd, msg, wParam, lParam);
+}
+
+
+/**********************************************************************
+ * parse_args
+ *
+ * Turn a list of args into a new list of args with each separate
+ * whitespace spaced string being an arg.
+ **********************************************************************/
+
+int
+parse_args (                     /*refine arg list */
+int argc,                        /*no of input args */
+char *argv[],                    /*input args */
+char *arglist[]                  /*output args */
+) {
+  int argcount;                  /*converted argc */
+  char *testchar;                /*char in option string */
+  int arg;                       /*current argument */
+
+  argcount = 0;                  /*no of options */
+  for (arg = 0; arg < argc; arg++) {
+    testchar = argv[arg];        /*start of arg */
+    do {
+      while (*testchar
+        && (*testchar == ' ' || *testchar == '\n'
+        || *testchar == '\t'))
+        testchar++;              /*skip white space */
+      if (*testchar) {
+                                 /*new arg */
+        arglist[argcount++] = testchar;
+                                 /*skip to white space */
+        for (testchar++; *testchar && *testchar != ' ' && *testchar != '\n' && *testchar != '\t'; testchar++);
+        if (*testchar)
+          *testchar++ = '\0';    /*turn to separate args */
+      }
+    }
+    while (*testchar);
+  }
+  return argcount;               /*new number of args */
+}
+#endif

ccmain/tesseractmain.h

@@ -0,0 +1,58 @@
+/**********************************************************************
+ * File:        tessedit.h  (Formerly tessedit.h)
+ * Description: Main program for merge of tess and editor.
+ * Author:		Ray Smith
+ * Created:		Tue Jan 07 15:21:46 GMT 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           TESSERACTMAIN_H
+#define           TESSERACTMAIN_H
+
+#include          "varable.h"
+#include          "tessclas.h"
+#include          "notdll.h"
+#include          "tessembedded.h"
+
+extern BOOL_VAR_H (tessedit_read_image, TRUE, "Ensure the image is read");
+INT32 api_main(                   //run from api
+               const char *arg0,  //program name
+               UINT16 lang        //language
+              );
+INT16 setup_info(                     //setup dummy engine info
+                 UINT16 lang,         //user language
+                 const char *name,    //of engine
+                 const char *version  //of engine
+                );
+INT16 read_image(               //read dummy image info
+                 IMAGE *im_out  //output image
+                );
+#ifdef __MSW32__
+int WINAPI WinMain(  //main for windows //command line
+                   HINSTANCE hInstance,
+                   HINSTANCE hPrevInstance,
+                   LPSTR lpszCmdLine,
+                   int nCmdShow);
+LONG WINAPI WndProc(            //message handler
+                    HWND hwnd,  //window with message
+                    UINT msg,   //message typ
+                    WPARAM wParam,
+                    LPARAM lParam);
+int parse_args (                 /*refine arg list */
+int argc,                        /*no of input args */
+char *argv[],                    /*input args */
+char *arglist[]                  /*output args */
+);
+#endif
+#endif

ccmain/tessvars.cpp

@@ -0,0 +1,38 @@
+/**********************************************************************
+ * File:        tessvars.cpp  (Formerly tessvars.c)
+ * Description: Variables and other globals for tessedit.
+ * Author:		Ray Smith
+ * Created:		Mon Apr 13 13:13:23 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          "tessvars.h"
+
+#define EXTERN
+
+EXTERN INT_VAR (tessedit_adapt_kludge, 0,
+"Use acceptable result or dangambigs");
+EXTERN BOOL_VAR (interactive_mode, FALSE, "Run interactively?");
+EXTERN BOOL_VAR (edit_variables, FALSE, "Variables Editor Window?");
+// xiaofan EXTERN STRING_VAR(file_type,".bl","Filename extension");
+EXTERN STRING_VAR (file_type, ".tif", "Filename extension");
+INT_VAR (testedit_match_debug, 0, "Integer match debug ctrl");
+EXTERN INT_VAR (tessedit_dangambigs_chop, FALSE,
+"Use DangAmbigs to direct chop");
+EXTERN INT_VAR (tessedit_dangambigs_assoc, FALSE,
+"Use DangAmbigs to direct assoc");
+
+EXTERN IMAGE page_image;         //image of page
+EXTERN FILE *debug_fp;           //write debug stuff here

ccmain/tessvars.h

@@ -0,0 +1,48 @@
+/**********************************************************************
+ * File:        tessvars.h  (Formerly tessvars.h)
+ * Description: Variables and other globals for tessedit.
+ * Author:		Ray Smith
+ * Created:		Mon Apr 13 13:13:23 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           TESSVARS_H
+#define           TESSVARS_H
+
+#include          "varable.h"
+#include          "img.h"
+#include          "tordmain.h"
+#include          "notdll.h"
+
+extern INT_VAR_H (tessedit_adapt_kludge, 0,
+"Use acceptable result or dangambigs");
+extern BOOL_VAR_H (interactive_mode, FALSE, "Run interactively?");
+extern BOOL_VAR_H (edit_variables, FALSE, "Variables Editor Window?");
+//xiaofan extern STRING_VAR_H(file_type,".bl","Filename extension");
+extern STRING_VAR_H (file_type, ".tif", "Filename extension");
+extern INT_VAR_H (tessedit_truncate_wordchoice_log, 10,
+"Max words to keep in list");
+extern INT_VAR_H (testedit_match_debug, 0, "Integer match debug ctrl");
+extern INT_VAR_H (tessedit_truncate_chopper, 1,
+"Shorten chopper seam search");
+extern INT_VAR_H (tessedit_fix_sideways_chops, 1,
+"Fix sideways chop problem");
+extern INT_VAR_H (tessedit_dangambigs_chop, FALSE,
+"Use DangAmbigs to direct chop");
+extern INT_VAR_H (tessedit_dangambigs_assoc, FALSE,
+"Use DangAmbigs to direct assoc");
+
+extern IMAGE page_image;         //image of page
+extern FILE *debug_fp;           //write debug stuff here
+#endif

ccmain/tfacep.h

@@ -0,0 +1,121 @@
+/**********************************************************************
+ * File:        tfacep.h  (Formerly tfacep.h)
+ * Description: Declarations of C functions and C owned data.
+ * Author:					Ray Smith
+ * Created:					Mon Apr 27 12:51:28 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           TFACEP_H
+#define           TFACEP_H
+
+#include          "hosthplb.h"
+#include          "tessclas.h"
+#include          "tessarray.h"
+#include          "tstruct.h"
+#include          "notdll.h"
+#include "choices.h"
+#include "oldlist.h"
+#include "hyphen.h"
+#include "tface.h"
+#include "permute.h"
+#include "adaptmatch.h"
+#include "blobclass.h"
+#include "stopper.h"
+#include "associate.h"
+#include "chop.h"
+#include "expandblob.h"
+#include "tordvars.h"
+#include "metrics.h"
+#include "tface.h"
+#include "badwords.h"
+#include "structures.h"
+
+#define BLOB_MATCHING_ON
+typedef void (*TESS_TESTER) (TBLOB *, BOOL8, char *, INT32, LIST);
+typedef LIST (*TESS_MATCHER) (TBLOB *, TBLOB *, TBLOB *, void *, TEXTROW *);
+
+extern "C"
+{
+  /*
+  int							start_recog(				//Real main in C
+  int							argc,
+  char						*argv[]);
+  void						program_editup2(			//afterforking part
+  int							argc,
+  char**						argv);
+
+  int							end_recog(					//Real main in C
+  int							argc,
+  char						*argv[]);
+  void						set_interactive_pass();
+  void						set_pass1();
+  void						set_pass2();
+  //ARRAY						cc_recog(TWERD*,TESS_CHOICE*,TESS_CHOICE*,TESS_TESTER,
+  //										TESS_TESTER);*/
+  //void                                          wo_learn_blob(TBLOB*,TEXTROW*,char*,INT32);
+  //LIST                                  AdaptiveClassifier(TBLOB*,TBLOB*,TEXTROW*);
+  //void                                          LearnBlob(TBLOB*,TEXTROW*,char*,INT32);
+  //TWERD                                         *newword();
+  //TBLOB                                         *newblob();
+  //TESSLINE                                      *newoutline();
+  //EDGEPT                                                *newedgept();
+  //void                                          oldedgept(EDGEPT*);
+  //void                                          destroy_nodes(void*,void (*)(void*));
+  //TESS_LIST                                     *append_choice(TESS_LIST*,char*,double,double,char);
+  //void                                          fix_quotes (char*);
+  //void                                          record_certainty(double,int);
+  //int                                                   AcceptableResult(A_CHOICE*,A_CHOICE*);
+  //int                                                   AdaptableWord(TWERD*,const char*,const char*);
+  //void                                          delete_word(TWERD*);
+  //void                                          free_blob(TBLOB*);
+  //void                                          add_document_word(A_CHOICE*);
+  //void                                          AdaptToWord(TWERD*,TEXTROW*,const char*,const char*,const char*);
+  //void                                          SaveBadWord(const char*,double);
+  //void                                          free_choice(TESS_CHOICE*);
+  //TWERD                                         *newword();
+  //TBLOB                                         *newblob();
+  //void                                          free_blob(                                      //free a blob
+  //      TBLOB                                           *blob);                                         //blob to free
+
+  //int                                                   dict_word( const char* );
+
+  //extern int                                    tess_cn_matching;
+  //extern int                                    tess_bn_matching;
+  //extern int                                    last_word_on_line;
+  extern TEXTROW normalized_row;
+  //extern TESS_MATCHER                   blob_matchers[];
+  //extern FILE                                   *rawfile;
+  //extern FILE                                   *textfile;
+  //extern int                                    character_count;
+  //extern int                                    word_count;
+  //extern int                                    enable_assoc;
+  //extern int                                    chop_enable;
+  //extern int                                    permute_only_top;
+  extern int display_ratings;
+
+};
+
+#if 0
+#define strsave(s)    \
+	((s) ?  \
+	((char*) strcpy ((char*)alloc_string (strlen(s)+1), s))  :  \
+	(NULL))
+#endif
+
+#define BOLD_ON				"&dB(s3B"
+#define BOLD_OFF			"&d@(s0B"
+#define UNDERLINE_ON		"&dD"
+#define UNDERLINE_OFF		"&d@"
+#endif

ccmain/tfacepp.cpp

@@ -0,0 +1,411 @@
+/**********************************************************************
+ * File:        tfacepp.cpp  (Formerly tface++.c)
+ * Description: C++ side of the C/C++ Tess/Editor interface.
+ * Author:					Ray Smith
+ * Created:					Thu Apr 23 15:39:23 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#ifdef __UNIX__
+#include          <assert.h>
+#endif
+#include          "errcode.h"
+#include          "tessarray.h"
+//#include                                                      "fxtop.h"
+#include          "werd.h"
+#include          "tfacep.h"
+#include          "tstruct.h"
+#include          "tfacepp.h"
+#include          "tessvars.h"
+#include          "reject.h"
+
+#define EXTERN
+
+EXTERN BOOL_VAR (tessedit_override_permuter, TRUE, "According to dict_word");
+
+static POLY_MATCHER tess_matcher;//current matcher
+static POLY_TESTER tess_tester;  //current tester
+static POLY_TESTER tess_trainer; //current trainer
+static DENORM *tess_denorm;      //current denorm
+static WERD *tess_word;          //current word
+
+#define MAX_UNDIVIDED_LENGTH 24
+/**********************************************************************
+ * recog_word
+ *
+ * Convert the word to tess form and pass it to the tess segmenter.
+ * Convert the output back to editor form.
+ **********************************************************************/
+WERD_CHOICE *recog_word(                           //recog one owrd
+                        WERD *word,                //word to do
+                        DENORM *denorm,            //de-normaliser
+                        POLY_MATCHER matcher,      //matcher function
+                        POLY_TESTER tester,        //tester function
+                        POLY_TESTER trainer,       //trainer function
+                        BOOL8 testing,             //true if answer driven
+                        WERD_CHOICE *&raw_choice,  //raw result //list of blob lists
+                        BLOB_CHOICE_LIST_CLIST *blob_choices,
+                        WERD *&outword             //bln word output
+                       ) {
+  WERD_CHOICE *word_choice;
+  UINT8 perm_type;
+  UINT8 real_dict_perm_type;
+
+  if (word->blob_list ()->empty ()) {
+    word_choice = new WERD_CHOICE ("", 10.0f, -1.0f, TOP_CHOICE_PERM);
+    raw_choice = new WERD_CHOICE ("", 10.0f, -1.0f, TOP_CHOICE_PERM);
+    outword = word->poly_copy (denorm->row ()->x_height ());
+  }
+  else
+    word_choice = recog_word_recursive (word, denorm, matcher, tester,
+      trainer, testing, raw_choice,
+      blob_choices, outword);
+  if ((word_choice->string ().length () !=
+    outword->blob_list ()->length ()) ||
+  (word_choice->string ().length () != blob_choices->length ())) {
+    tprintf
+      ("recog_word ASSERT FAIL String:\"%s\"; Strlen=%d; #Blobs=%d; #Choices=%d\n",
+      word_choice->string ().string (), word_choice->string ().length (),
+      outword->blob_list ()->length (), blob_choices->length ());
+  }
+  ASSERT_HOST (word_choice->string ().length () ==
+    outword->blob_list ()->length ());
+  ASSERT_HOST (word_choice->string ().length () == blob_choices->length ());
+
+  /* Copy any reject blobs into the outword */
+  outword->rej_blob_list ()->deep_copy (word->rej_blob_list ());
+
+  if (tessedit_override_permuter) {
+    /* Override the permuter type if a straight dictionary check disagrees. */
+    perm_type = word_choice->permuter ();
+    if ((perm_type != SYSTEM_DAWG_PERM) &&
+    (perm_type != FREQ_DAWG_PERM) && (perm_type != USER_DAWG_PERM)) {
+      real_dict_perm_type = dict_word (word_choice->string ().string ());
+      if (((real_dict_perm_type == SYSTEM_DAWG_PERM) ||
+        (real_dict_perm_type == FREQ_DAWG_PERM) ||
+        (real_dict_perm_type == USER_DAWG_PERM)) &&
+        (alpha_count (word_choice->string ().string ()) > 0))
+        word_choice->set_permuter (real_dict_perm_type);
+      //Use dict perm
+    }
+    if (tessedit_rejection_debug && perm_type != word_choice->permuter ()) {
+      tprintf ("Permuter Type Flipped from %d to %d\n",
+        perm_type, word_choice->permuter ());
+    }
+  }
+  assert ((word_choice == NULL) == (raw_choice == NULL));
+  return word_choice;
+}
+
+
+/**********************************************************************
+ * recog_word_recursive
+ *
+ * Convert the word to tess form and pass it to the tess segmenter.
+ * Convert the output back to editor form.
+ **********************************************************************/
+
+WERD_CHOICE *recog_word_recursive(                           //recog one owrd
+                                  WERD *word,                //word to do
+                                  DENORM *denorm,            //de-normaliser
+                                  POLY_MATCHER matcher,      //matcher function
+                                  POLY_TESTER tester,        //tester function
+                                  POLY_TESTER trainer,       //trainer function
+                                  BOOL8 testing,             //true if answer driven
+                                  WERD_CHOICE *&raw_choice,  //raw result //list of blob lists
+                                  BLOB_CHOICE_LIST_CLIST *blob_choices,
+                                  WERD *&outword             //bln word output
+                                 ) {
+  INT32 initial_blob_choice_len;
+  INT32 word_length;             //no of blobs
+  STRING word_string;            //converted from tess
+  ARRAY tess_ratings;            //tess results
+  A_CHOICE tess_choice;          //best word
+  A_CHOICE tess_raw;             //raw result
+  TWERD *tessword;               //tess format
+  BLOB_CHOICE_LIST *choice_list; //fake list
+                                 //iterator
+  BLOB_CHOICE_LIST_C_IT choice_it;
+
+  tess_matcher = matcher;        //install matcher
+  tess_tester = testing ? tester : NULL;
+  tess_trainer = testing ? trainer : NULL;
+  tess_denorm = denorm;
+  tess_word = word;
+  //      blob_matchers[1]=call_matcher;
+  if (word->blob_list ()->length () > MAX_UNDIVIDED_LENGTH) {
+    return split_and_recog_word (word, denorm, matcher, tester, trainer,
+      testing, raw_choice, blob_choices,
+      outword);
+  }
+  else {
+    if (word->flag (W_EOL))
+      last_word_on_line = TRUE;
+    else
+      last_word_on_line = FALSE;
+    initial_blob_choice_len = blob_choices->length ();
+    tessword = make_tess_word (word, NULL);
+    tess_ratings = cc_recog (tessword, &tess_choice, &tess_raw,
+      testing
+      && tester != NULL /* ? call_tester : NULL */ ,
+      testing
+      && trainer !=
+      NULL /* ? call_train_tester : NULL */ );
+                                 //convert word
+    outword = make_ed_word (tessword, word);
+    if (outword == NULL) {
+      outword = word->poly_copy (denorm->row ()->x_height ());
+    }
+    delete_word(tessword);  //get rid of it
+                                 //no of blobs
+    word_length = outword->blob_list ()->length ();
+                                 //convert all ratings
+    convert_choice_lists(tess_ratings, blob_choices);
+                                 //copy string
+    word_string = tess_raw.string;
+    while (word_string.length () < word_length)
+      word_string += " ";        //pad with blanks
+    raw_choice = new WERD_CHOICE (word_string.string (),
+      tess_raw.rating, tess_raw.certainty,
+      tess_raw.permuter);
+    word_string = tess_choice.string;
+    if (word_string.length () > word_length) {
+      tprintf ("recog_word: Discarded long string \"%s\"\n",
+        word_string.string ());
+      word_string = NULL;        //should never happen
+    }
+    if (blob_choices->length () - initial_blob_choice_len != word_length) {
+      word_string = NULL;        //force rejection
+      tprintf ("recog_word: Choices list len:%d; blob lists len:%d\n",
+        blob_choices->length (), word_length);
+                                 //list of lists
+      choice_it.set_to_list (blob_choices);
+      while (blob_choices->length () - initial_blob_choice_len <
+      word_length) {
+                                 //get fake one
+        choice_list = new BLOB_CHOICE_LIST;
+                                 //add to list
+        choice_it.add_to_end (choice_list);
+        tprintf ("recog_word: Added dummy choice list\n");
+      }
+      while (blob_choices->length () - initial_blob_choice_len >
+      word_length) {
+        choice_it.move_to_last ();
+                                 //should never happen
+        delete choice_it.extract ();
+        tprintf ("recog_word: Deleted choice list\n");
+      }
+    }
+    while (word_string.length () < word_length)
+      word_string += " ";        //pad with blanks
+
+    assert (raw_choice != NULL);
+    if (tess_choice.string)
+      strfree(tess_choice.string);
+    if (tess_raw.string)
+      strfree(tess_raw.string);
+    return new WERD_CHOICE (word_string.string (),
+      tess_choice.rating, tess_choice.certainty,
+      tess_choice.permuter);
+  }
+}
+
+
+/**********************************************************************
+ * split_and_recog_word
+ *
+ * Convert the word to tess form and pass it to the tess segmenter.
+ * Convert the output back to editor form.
+ **********************************************************************/
+
+WERD_CHOICE *split_and_recog_word(                           //recog one owrd
+                                  WERD *word,                //word to do
+                                  DENORM *denorm,            //de-normaliser
+                                  POLY_MATCHER matcher,      //matcher function
+                                  POLY_TESTER tester,        //tester function
+                                  POLY_TESTER trainer,       //trainer function
+                                  BOOL8 testing,             //true if answer driven
+                                  WERD_CHOICE *&raw_choice,  //raw result //list of blob lists
+                                  BLOB_CHOICE_LIST_CLIST *blob_choices,
+                                  WERD *&outword             //bln word output
+                                 ) {
+  //   INT32                                                      outword1_len;
+  //   INT32                                                      outword2_len;
+  WERD *first_word;              //poly copy of word
+  WERD *second_word;             //fabricated word
+  WERD *outword2;                //2nd output word
+  PBLOB *blob;
+  WERD_CHOICE *result;           //resturn value
+  WERD_CHOICE *result2;          //output of 2nd word
+  WERD_CHOICE *raw_choice2;      //raw version of 2nd
+  float gap;                     //blob gap
+  float bestgap;                 //biggest gap
+  PBLOB_LIST new_blobs;          //list of gathered blobs
+  PBLOB_IT blob_it;
+                                 //iterator
+  PBLOB_IT new_blob_it = &new_blobs;
+
+  first_word = word->poly_copy (denorm->row ()->x_height ());
+  blob_it.set_to_list (first_word->blob_list ());
+  bestgap = -MAX_INT32;
+  while (!blob_it.at_last ()) {
+    blob = blob_it.data ();
+                                 //gap to next
+    gap = blob_it.data_relative (1)->bounding_box ().left () - blob->bounding_box ().right ();
+    blob_it.forward ();
+    if (gap > bestgap) {
+      bestgap = gap;             //find biggest
+      new_blob_it = blob_it;     //save position
+    }
+  }
+                                 //take 2nd half
+  new_blobs.assign_to_sublist (&new_blob_it, &blob_it);
+                                 //make it a word
+  second_word = new WERD (&new_blobs, 1, NULL);
+  ASSERT_HOST (word->blob_list ()->length () ==
+    first_word->blob_list ()->length () +
+    second_word->blob_list ()->length ());
+
+  result = recog_word_recursive (first_word, denorm, matcher,
+    tester, trainer, testing, raw_choice,
+    blob_choices, outword);
+  delete first_word;             //done that one
+  result2 = recog_word_recursive (second_word, denorm, matcher,
+    tester, trainer, testing, raw_choice2,
+    blob_choices, outword2);
+  delete second_word;            //done that too
+  *result += *result2;           //combine ratings
+  delete result2;
+  *raw_choice += *raw_choice2;
+  delete raw_choice2;            //finished with it
+  //   outword1_len= outword->blob_list()->length();
+  //   outword2_len= outword2->blob_list()->length();
+  outword->join_on (outword2);   //join words
+  delete outword2;
+  //   if ( outword->blob_list()->length() != outword1_len + outword2_len )
+  //      tprintf( "Split&Recog: part1len=%d; part2len=%d; combinedlen=%d\n",
+  //                                outword1_len, outword2_len, outword->blob_list()->length() );
+  //   ASSERT_HOST( outword->blob_list()->length() == outword1_len + outword2_len );
+  return result;
+}
+
+
+/**********************************************************************
+ * call_matcher
+ *
+ * Called from Tess with a blob in tess form.
+ * Convert the blob to editor form.
+ * Call the matcher setup by the segmenter in tess_matcher.
+ * Convert the output choices back to tess form.
+ **********************************************************************/
+
+LIST call_matcher(                  //call a matcher
+                  TBLOB *ptblob,    //previous
+                  TBLOB *tessblob,  //blob to match
+                  TBLOB *ntblob,    //next
+                  void *,           //unused parameter
+                  TEXTROW *         //always null anyway
+                 ) {
+  PBLOB *pblob;                  //converted blob
+  PBLOB *blob;                   //converted blob
+  PBLOB *nblob;                  //converted blob
+  LIST result;                   //tess output
+  BLOB_CHOICE *choice;           //current choice
+  char string[2];                //char converted
+  BLOB_CHOICE_LIST ratings;      //matcher result
+  BLOB_CHOICE_IT it;             //iterator
+
+  blob = make_ed_blob (tessblob);//convert blob
+  if (blob == NULL)
+    return NULL;                 //can't do it
+  pblob = ptblob != NULL ? make_ed_blob (ptblob) : NULL;
+  nblob = ntblob != NULL ? make_ed_blob (ntblob) : NULL;
+  (*tess_matcher) (pblob, blob, nblob, tess_word, tess_denorm, ratings);
+  //match it
+  delete blob;                   //don't need that now
+  if (pblob != NULL)
+    delete pblob;
+  if (nblob != NULL)
+    delete nblob;
+  it.set_to_list (&ratings);     //get list
+  result = NULL;
+  string[1] = '\0';
+  for (it.mark_cycle_pt (); !it.cycled_list (); it.forward ()) {
+    choice = it.data ();
+    string[0] = choice->char_class ();
+    result = append_choice (result, string,
+      choice->rating (), choice->certainty (),
+      choice->config ());
+  }
+  return result;                 //converted list
+}
+
+
+/**********************************************************************
+ * call_tester
+ *
+ * Called from Tess with a blob in tess form.
+ * Convert the blob to editor form.
+ * Call the tester setup by the segmenter in tess_tester.
+ **********************************************************************/
+
+void call_tester(                     //call a tester
+                 TBLOB *tessblob,     //blob to test
+                 BOOL8 correct_blob,  //true if good
+                 char *text,          //source text
+                 INT32 count,         //chars in text
+                 LIST result          //output of matcher
+                ) {
+  PBLOB *blob;                   //converted blob
+  BLOB_CHOICE_LIST ratings;      //matcher result
+
+  blob = make_ed_blob (tessblob);//convert blob
+  if (blob == NULL)
+    return;
+                                 //make it right type
+  convert_choice_list(result, ratings);
+  if (tess_tester != NULL)
+    (*tess_tester) (blob, tess_denorm, correct_blob, text, count, &ratings);
+  delete blob;                   //don't need that now
+}
+
+
+/**********************************************************************
+ * call_train_tester
+ *
+ * Called from Tess with a blob in tess form.
+ * Convert the blob to editor form.
+ * Call the trainer setup by the segmenter in tess_trainer.
+ **********************************************************************/
+
+void call_train_tester(                     //call a tester
+                       TBLOB *tessblob,     //blob to test
+                       BOOL8 correct_blob,  //true if good
+                       char *text,          //source text
+                       INT32 count,         //chars in text
+                       LIST result          //output of matcher
+                      ) {
+  PBLOB *blob;                   //converted blob
+  BLOB_CHOICE_LIST ratings;      //matcher result
+
+  blob = make_ed_blob (tessblob);//convert blob
+  if (blob == NULL)
+    return;
+                                 //make it right type
+  convert_choice_list(result, ratings);
+  if (tess_trainer != NULL)
+    (*tess_trainer) (blob, tess_denorm, correct_blob, text, count, &ratings);
+  delete blob;                   //don't need that now
+}

ccmain/tfacepp.h

@@ -0,0 +1,85 @@
+/**********************************************************************
+ * File:        tfacepp.h  (Formerly tface++.h)
+ * Description: C++ side of the C/C++ Tess/Editor interface.
+ * Author:					Ray Smith
+ * Created:					Thu Apr 23 15:39:23 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           TFACEPP_H
+#define           TFACEPP_H
+
+#include          "varable.h"
+#include          "tstruct.h"
+#include          "ratngs.h"
+#include          "tessclas.h"
+#include          "notdll.h"
+
+extern BOOL_VAR_H (tessedit_override_permuter, TRUE,
+"According to dict_word");
+WERD_CHOICE *recog_word(                           //recog one owrd
+                        WERD *word,                //word to do
+                        DENORM *denorm,            //de-normaliser
+                        POLY_MATCHER matcher,      //matcher function
+                        POLY_TESTER tester,        //tester function
+                        POLY_TESTER trainer,       //trainer function
+                        BOOL8 testing,             //true if answer driven
+                        WERD_CHOICE *&raw_choice,  //raw result //list of blob lists
+                        BLOB_CHOICE_LIST_CLIST *blob_choices,
+                        WERD *&outword             //bln word output
+                       );
+                                 //recog one owrd
+WERD_CHOICE *recog_word_recursive(WERD *word,                //word to do
+                                  DENORM *denorm,            //de-normaliser
+                                  POLY_MATCHER matcher,      //matcher function
+                                  POLY_TESTER tester,        //tester function
+                                  POLY_TESTER trainer,       //trainer function
+                                  BOOL8 testing,             //true if answer driven
+                                  WERD_CHOICE *&raw_choice,  //raw result //list of blob lists
+                                  BLOB_CHOICE_LIST_CLIST *blob_choices,
+                                  WERD *&outword             //bln word output
+                                 );
+                                 //recog one owrd
+WERD_CHOICE *split_and_recog_word(WERD *word,                //word to do
+                                  DENORM *denorm,            //de-normaliser
+                                  POLY_MATCHER matcher,      //matcher function
+                                  POLY_TESTER tester,        //tester function
+                                  POLY_TESTER trainer,       //trainer function
+                                  BOOL8 testing,             //true if answer driven
+                                  WERD_CHOICE *&raw_choice,  //raw result //list of blob lists
+                                  BLOB_CHOICE_LIST_CLIST *blob_choices,
+                                  WERD *&outword             //bln word output
+                                 );
+LIST call_matcher(                  //call a matcher
+                  TBLOB *ptblob,    //previous
+                  TBLOB *tessblob,  //blob to match
+                  TBLOB *ntblob,    //next
+                  void *,           //unused parameter
+                  TEXTROW *         //always null anyway
+                 );
+void call_tester(                     //call a tester
+                 TBLOB *tessblob,     //blob to test
+                 BOOL8 correct_blob,  //true if good
+                 char *text,          //source text
+                 INT32 count,         //chars in text
+                 LIST result          //output of matcher
+                );
+void call_train_tester(                     //call a tester
+                       TBLOB *tessblob,     //blob to test
+                       BOOL8 correct_blob,  //true if good
+                       char *text,          //source text
+                       INT32 count,         //chars in text
+                       LIST result          //output of matcher
+                      );
+#endif

ccmain/tstruct.cpp

@@ -0,0 +1,511 @@
+/**********************************************************************
+ * File:        tstruct.cpp  (Formerly tstruct.c)
+ * Description: Code to manipulate the structures of the C++/C interface.
+ * Author:		Ray Smith
+ * Created:		Thu Apr 23 15:49:29 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          "tfacep.h"
+#include          "tstruct.h"
+//#include "structures.h"
+
+static ERRCODE BADFRAGMENTS = "Couldn't find matching fragment ends";
+
+ELISTIZE (FRAGMENT)
+//extern /*"C"*/ oldoutline(TESSLINE*);
+/**********************************************************************
+ * FRAGMENT::FRAGMENT
+ *
+ * Constructor for fragments.
+ **********************************************************************/
+FRAGMENT::FRAGMENT (             //constructor
+EDGEPT * head_pt,                //start point
+EDGEPT * tail_pt                 //end point
+):head (head_pt->pos.x, head_pt->pos.y), tail (tail_pt->pos.x,
+tail_pt->pos.y) {
+  headpt = head_pt;              //save ptrs
+  tailpt = tail_pt;
+}
+
+
+/**********************************************************************
+ * make_ed_word
+ *
+ * Make an editor format word from the tess style word.
+ **********************************************************************/
+
+WERD *make_ed_word(                  //construct word
+                   TWERD *tessword,  //word to convert
+                   WERD *clone       //clone this one
+                  ) {
+  WERD *word;                    //converted word
+  TBLOB *tblob;                  //current blob
+  PBLOB *blob;                   //new blob
+  PBLOB_LIST blobs;              //list of blobs
+  PBLOB_IT blob_it = &blobs;     //iterator
+
+  for (tblob = tessword->blobs; tblob != NULL; tblob = tblob->next) {
+    blob = make_ed_blob (tblob);
+    if (blob != NULL)
+      blob_it.add_after_then_move (blob);
+  }
+  if (!blobs.empty ())
+    word = new WERD (&blobs, clone);
+  else
+    word = NULL;
+  return word;
+}
+
+
+/**********************************************************************
+ * make_ed_blob
+ *
+ * Make an editor format blob from the tess style blob.
+ **********************************************************************/
+
+PBLOB *make_ed_blob(                 //construct blob
+                    TBLOB *tessblob  //blob to convert
+                   ) {
+  TESSLINE *tessol;              //tess outline
+  FRAGMENT_LIST fragments;       //list of fragments
+  OUTLINE *outline;              //current outline
+  OUTLINE_LIST out_list;         //list of outlines
+  OUTLINE_IT out_it = &out_list; //iterator
+
+  for (tessol = tessblob->outlines; tessol != NULL; tessol = tessol->next) {
+                                 //stick in list
+    register_outline(tessol, &fragments);
+  }
+  while (!fragments.empty ()) {
+    outline = make_ed_outline (&fragments);
+    if (outline != NULL)
+      out_it.add_after_then_move (outline);
+  }
+  if (out_it.empty())
+    return NULL;                 //couldn't do it
+  return new PBLOB (&out_list);  //turn to blob
+}
+
+
+/**********************************************************************
+ * make_ed_outline
+ *
+ * Make an editor format outline from the list of fragments.
+ **********************************************************************/
+
+OUTLINE *make_ed_outline(                     //constructoutline
+                         FRAGMENT_LIST *list  //list of fragments
+                        ) {
+  FRAGMENT *fragment;            //current fragment
+  EDGEPT *edgept;                //current point
+  ICOORD headpos;                //coords of head
+  ICOORD tailpos;                //coords of tail
+  FCOORD pos;                    //coords of edgept
+  FCOORD vec;                    //empty
+  POLYPT *polypt;                //current point
+  POLYPT_LIST poly_list;         //list of point
+  POLYPT_IT poly_it = &poly_list;//iterator
+  FRAGMENT_IT fragment_it = list;//fragment
+
+  headpos = fragment_it.data ()->head;
+  do {
+    fragment = fragment_it.data ();
+    edgept = fragment->headpt;   //start of segment
+    do {
+      pos = FCOORD (edgept->pos.x, edgept->pos.y);
+      vec = FCOORD (edgept->vec.x, edgept->vec.y);
+      polypt = new POLYPT (pos, vec);
+                                 //add to list
+      poly_it.add_after_then_move (polypt);
+      edgept = edgept->next;
+    }
+    while (edgept != fragment->tailpt);
+    tailpos = ICOORD (edgept->pos.x, edgept->pos.y);
+                                 //get rid of it
+    delete fragment_it.extract ();
+    if (tailpos != headpos) {
+      if (fragment_it.empty ()) {
+        //                              tprintf("Bad tailpos (%d,%d), Head=(%d,%d), no fragments.\n",
+        //                                      fragment->head.x(),fragment->head.y(),
+        //                                      headpos.x(),headpos.y());
+        return NULL;
+      }
+      fragment_it.forward ();
+                                 //find next segment
+      for (fragment_it.mark_cycle_pt (); !fragment_it.cycled_list () && fragment_it.data ()->head != tailpos;
+        fragment_it.forward ());
+      if (fragment_it.data ()->head != tailpos) {
+        //                              tprintf("Bad tailpos (%d,%d), Fragments are:\n",
+        //                                      tailpos.x(),tailpos.y());
+        for (fragment_it.mark_cycle_pt ();
+        !fragment_it.cycled_list (); fragment_it.forward ()) {
+          fragment = fragment_it.extract ();
+          //                                      tprintf("Head=(%d,%d), tail=(%d,%d)\n",
+          //                                              fragment->head.x(),fragment->head.y(),
+          //                                              fragment->tail.x(),fragment->tail.y());
+          delete fragment;
+        }
+        return NULL;             //can't do it
+        //                              BADFRAGMENTS.error("make_ed_blob",ABORT,NULL);
+      }
+    }
+  }
+  while (tailpos != headpos);
+  return new OUTLINE (&poly_it); //turn to outline
+}
+
+
+/**********************************************************************
+ * register_outline
+ *
+ * Add the fragments in the given outline to the list
+ **********************************************************************/
+
+void register_outline(                     //add fragments
+                      TESSLINE *outline,   //tess format
+                      FRAGMENT_LIST *list  //list to add to
+                     ) {
+  EDGEPT *startpt;               //start of outline
+  EDGEPT *headpt;                //start of fragment
+  EDGEPT *tailpt;                //end of fragment
+  FRAGMENT *fragment;            //new fragment
+  FRAGMENT_IT it = list;         //iterator
+
+  startpt = outline->loop;
+  do {
+    startpt = startpt->next;
+    if (startpt == NULL)
+      return;                    //illegal!
+  }
+  while (startpt->flags[0] == 0 && startpt != outline->loop);
+  headpt = startpt;
+  do
+  startpt = startpt->next;
+  while (startpt->flags[0] != 0 && startpt != headpt);
+  if (startpt->flags[0] != 0)
+    return;                      //all hidden!
+
+  headpt = startpt;
+  do {
+    tailpt = headpt;
+    do
+    tailpt = tailpt->next;
+    while (tailpt->flags[0] == 0 && tailpt != startpt);
+    fragment = new FRAGMENT (headpt, tailpt);
+    it.add_after_then_move (fragment);
+    while (tailpt->flags[0] != 0)
+      tailpt = tailpt->next;
+    headpt = tailpt;
+  }
+  while (tailpt != startpt);
+}
+
+
+/**********************************************************************
+ * convert_choice_lists
+ *
+ * Convert the ARRAY of TESS_LIST of TESS_CHOICEs into a BLOB_CHOICE_LIST.
+ **********************************************************************/
+
+void convert_choice_lists(                                 //convert lists
+                          ARRAY tessarray,                 //list from tess
+                          BLOB_CHOICE_LIST_CLIST *ratings  //list of results
+                         ) {
+  INT32 length;                  //elements in array
+  INT32 index;                   //index to array
+  LIST result;                   //tess output
+                                 //iterator
+  BLOB_CHOICE_LIST_C_IT it = ratings;
+  BLOB_CHOICE_LIST *choice;      //created choice
+
+  if (tessarray != NULL) {
+    length = array_count (tessarray);
+    for (index = 0; index < length; index++) {
+      result = (LIST) array_value (tessarray, index);
+                                 //make one
+      choice = new BLOB_CHOICE_LIST;
+                                 //convert blob choices
+      convert_choice_list(result, *choice);
+                                 //add to super list
+      it.add_after_then_move (choice);
+    }
+    free_mem(tessarray);  //lists already freed
+  }
+}
+
+
+/**********************************************************************
+ * convert_choice_list
+ *
+ * Convert the LIST of TESS_CHOICEs into a BLOB_CHOICE_LIST.
+ **********************************************************************/
+
+void convert_choice_list(                           //convert lists
+                         LIST list,                 //list from tess
+                         BLOB_CHOICE_LIST &ratings  //list of results
+                        ) {
+  LIST result;                   //tess output
+  BLOB_CHOICE_IT it = &ratings;  //iterator
+  BLOB_CHOICE *choice;           //created choice
+  A_CHOICE *tesschoice;          //choice to convert
+
+  for (result = list; result != NULL; result = result->next) {
+  //traverse list
+    tesschoice = (A_CHOICE *) result->node;
+                                 //make one
+    choice = new BLOB_CHOICE (tesschoice->string[0], tesschoice->rating, tesschoice->certainty, tesschoice->config);
+    it.add_after_then_move (choice);
+  }
+  destroy_nodes (list, (void (*)(void *)) free_choice);
+  //get rid of it
+}
+
+
+/**********************************************************************
+ * make_tess_row
+ *
+ * Make a fake row structure to pass to the tesseract matchers.
+ **********************************************************************/
+
+void make_tess_row(                  //make fake row
+                   DENORM *denorm,   //row info
+                   TEXTROW *tessrow  //output row
+                  ) {
+  tessrow->baseline.segments = 1;
+  tessrow->baseline.xstarts[0] = -32767;
+  tessrow->baseline.xstarts[1] = 32767;
+  tessrow->baseline.quads[0].a = 0;
+  tessrow->baseline.quads[0].b = 0;
+  tessrow->baseline.quads[0].c = bln_baseline_offset;
+  tessrow->xheight.segments = 1;
+  tessrow->xheight.xstarts[0] = -32767;
+  tessrow->xheight.xstarts[1] = 32767;
+  tessrow->xheight.quads[0].a = 0;
+  tessrow->xheight.quads[0].b = 0;
+  tessrow->xheight.quads[0].c = bln_x_height + bln_baseline_offset;
+  tessrow->lineheight = bln_x_height;
+  tessrow->ascrise = denorm->row ()->ascenders () * denorm->scale ();
+  tessrow->descdrop = denorm->row ()->descenders () * denorm->scale ();
+}
+
+
+/**********************************************************************
+ * make_tess_word
+ *
+ * Convert the word to Tess format.
+ **********************************************************************/
+
+TWERD *make_tess_word(              //convert owrd
+                      WERD *word,   //word to do
+                      TEXTROW *row  //fake row
+                     ) {
+  TWERD *tessword;               //tess format
+
+  tessword = newword ();         //use old allocator
+  tessword->row = row;           //give them something
+                                 //copy string
+  tessword->correct = strsave (word->text ());
+  tessword->guess = NULL;
+  tessword->blobs = make_tess_blobs (word->blob_list ());
+  tessword->blanks = 1;
+  tessword->blobcount = word->blob_list ()->length ();
+  tessword->next = NULL;
+  return tessword;
+}
+
+
+/**********************************************************************
+ * make_tess_blobs
+ *
+ * Make Tess style blobs from a list of BLOBs.
+ **********************************************************************/
+
+TBLOB *make_tess_blobs(                      //make tess blobs
+                       PBLOB_LIST *bloblist  //list to convert
+                      ) {
+  PBLOB_IT it = bloblist;        //iterator
+  PBLOB *blob;                   //current blob
+  TBLOB *head;                   //output list
+  TBLOB *tail;                   //end of list
+  TBLOB *tessblob;
+
+  head = NULL;
+  tail = NULL;
+  for (it.mark_cycle_pt (); !it.cycled_list (); it.forward ()) {
+    blob = it.data ();
+    tessblob = make_tess_blob (blob, TRUE);
+    if (head)
+      tail->next = tessblob;
+    else
+      head = tessblob;
+    tail = tessblob;
+  }
+  return head;
+}
+
+
+/**********************************************************************
+ * make_tess_blob
+ *
+ * Make a single Tess style blob
+ **********************************************************************/
+
+TBLOB *make_tess_blob(               //make tess blob
+                      PBLOB *blob,   //blob to convert
+                      BOOL8 flatten  //flatten outline structure
+                     ) {
+  INT32 index;
+  TBLOB *tessblob;
+
+  tessblob = newblob ();
+  tessblob->outlines = (struct olinestruct *)
+    make_tess_outlines (blob->out_list (), flatten);
+  for (index = 0; index < TBLOBFLAGS; index++)
+    tessblob->flags[index] = 0;  //!!
+  tessblob->correct = 0;
+  tessblob->guess = 0;
+  for (index = 0; index < MAX_WO_CLASSES; index++) {
+    tessblob->classes[index] = 0;
+    tessblob->values[index] = 0;
+  }
+  tessblob->next = NULL;
+  return tessblob;
+}
+
+
+/**********************************************************************
+ * make_tess_outlines
+ *
+ * Make Tess style outlines from a list of OUTLINEs.
+ **********************************************************************/
+
+TESSLINE *make_tess_outlines(                            //make tess outlines
+                             OUTLINE_LIST *outlinelist,  //list to convert
+                             BOOL8 flatten               //flatten outline structure
+                            ) {
+  OUTLINE_IT it = outlinelist;   //iterator
+  OUTLINE *outline;              //current outline
+  TESSLINE *head;                //output list
+  TESSLINE *tail;                //end of list
+  TESSLINE *tessoutline;
+
+  head = NULL;
+  tail = NULL;
+  for (it.mark_cycle_pt (); !it.cycled_list (); it.forward ()) {
+    outline = it.data ();
+    tessoutline = newoutline ();
+    tessoutline->compactloop = NULL;
+    tessoutline->loop = make_tess_edgepts (outline->polypts (),
+      tessoutline->topleft,
+      tessoutline->botright);
+    if (tessoutline->loop == NULL) {
+      oldoutline(tessoutline);
+      continue;
+    }
+    tessoutline->start = tessoutline->loop->pos;
+    tessoutline->node = NULL;
+    tessoutline->next = NULL;
+    tessoutline->child = NULL;
+    if (!outline->child ()->empty ()) {
+      if (flatten)
+        tessoutline->next = (struct olinestruct *)
+          make_tess_outlines (outline->child (), flatten);
+      else {
+        tessoutline->next = NULL;
+        tessoutline->child = (struct olinestruct *)
+          make_tess_outlines (outline->child (), flatten);
+      }
+    }
+    else
+      tessoutline->next = NULL;
+    if (head)
+      tail->next = tessoutline;
+    else
+      head = tessoutline;
+    while (tessoutline->next != NULL)
+      tessoutline = tessoutline->next;
+    tail = tessoutline;
+  }
+  return head;
+}
+
+
+/**********************************************************************
+ * make_tess_edgepts
+ *
+ * Make Tess style edgepts from a list of POLYPTs.
+ **********************************************************************/
+
+EDGEPT *make_tess_edgepts(                          //make tess edgepts
+                          POLYPT_LIST *edgeptlist,  //list to convert
+                          TPOINT &tl,               //bounding box
+                          TPOINT &br) {
+  INT32 index;
+  POLYPT_IT it = edgeptlist;     //iterator
+  POLYPT *edgept;                //current edgept
+  EDGEPT *head;                  //output list
+  EDGEPT *tail;                  //end of list
+  EDGEPT *tessedgept;
+
+  head = NULL;
+  tail = NULL;
+  tl.x = MAX_INT16;
+  tl.y = -MAX_INT16;
+  br.x = -MAX_INT16;
+  br.y = MAX_INT16;
+  for (it.mark_cycle_pt (); !it.cycled_list ();) {
+    edgept = it.data ();
+    tessedgept = newedgept ();
+    tessedgept->pos.x = (INT16) edgept->pos.x ();
+    tessedgept->pos.y = (INT16) edgept->pos.y ();
+    if (tessedgept->pos.x < tl.x)
+      tl.x = tessedgept->pos.x;
+    if (tessedgept->pos.x > br.x)
+      br.x = tessedgept->pos.x;
+    if (tessedgept->pos.y > tl.y)
+      tl.y = tessedgept->pos.y;
+    if (tessedgept->pos.y < br.y)
+      br.y = tessedgept->pos.y;
+    if (head != NULL && tessedgept->pos.x == tail->pos.x
+    && tessedgept->pos.y == tail->pos.y) {
+      oldedgept(tessedgept);
+    }
+    else {
+      for (index = 0; index < EDGEPTFLAGS; index++)
+        tessedgept->flags[index] = 0;
+      if (head != NULL) {
+        tail->vec.x = tessedgept->pos.x - tail->pos.x;
+        tail->vec.y = tessedgept->pos.y - tail->pos.y;
+        tessedgept->prev = tail;
+      }
+      tessedgept->next = head;
+      if (head)
+        tail->next = tessedgept;
+      else
+        head = tessedgept;
+      tail = tessedgept;
+    }
+    it.forward ();
+  }
+  head->prev = tail;
+  tail->vec.x = head->pos.x - tail->pos.x;
+  tail->vec.y = head->pos.y - tail->pos.y;
+  if (head == tail) {
+    oldedgept(head);
+    return NULL;                 //empty
+  }
+  return head;
+}

ccmain/tstruct.h

@@ -0,0 +1,108 @@
+/**********************************************************************
+ * File:        tstruct.h  (Formerly tstruct.h)
+ * Description: Code to manipulate the structures of the C++/C interface.
+ * Author:		Ray Smith
+ * Created:		Thu Apr 23 15:49:29 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           TSTRUCT_H
+#define           TSTRUCT_H
+
+#include          "tessarray.h"
+#include          "werd.h"
+#include          "tessclas.h"
+#include          "ratngs.h"
+#include          "notdll.h"
+#include "oldlist.h"
+
+/*
+struct TESS_LIST
+{
+  TESS_LIST				*node;						//data
+  TESS_LIST				*next;						//next in list
+};
+
+struct TESS_CHOICE
+{
+  float					rating;						//scaled
+  float					certainty;					//absolute
+  char					permuter;					//which permuter code
+  INT8					config;						//which config
+  char*					string;						//really can!
+};
+*/
+class FRAGMENT:public ELIST_LINK
+{
+  public:
+    FRAGMENT() {  //constructor
+    }
+    FRAGMENT(EDGEPT *head_pt,   //start
+             EDGEPT *tail_pt);  //end
+
+    ICOORD head;                 //coords of start
+    ICOORD tail;                 //coords of end
+    EDGEPT *headpt;              //start point
+    EDGEPT *tailpt;              //end point
+
+    NEWDELETE2 (FRAGMENT)
+};
+
+ELISTIZEH (FRAGMENT)
+WERD *make_ed_word(                  //construct word
+                   TWERD *tessword,  //word to convert
+                   WERD *clone       //clone this one
+                  );
+PBLOB *make_ed_blob(                 //construct blob
+                    TBLOB *tessblob  //blob to convert
+                   );
+OUTLINE *make_ed_outline(                     //constructoutline
+                         FRAGMENT_LIST *list  //list of fragments
+                        );
+void register_outline(                     //add fragments
+                      TESSLINE *outline,   //tess format
+                      FRAGMENT_LIST *list  //list to add to
+                     );
+void convert_choice_lists(                                 //convert lists
+                          ARRAY tessarray,                 //list from tess
+                          BLOB_CHOICE_LIST_CLIST *ratings  //list of results
+                         );
+void convert_choice_list(                           //convert lists
+                         LIST list,                 //list from tess
+                         BLOB_CHOICE_LIST &ratings  //list of results
+                        );
+void make_tess_row(                  //make fake row
+                   DENORM *denorm,   //row info
+                   TEXTROW *tessrow  //output row
+                  );
+TWERD *make_tess_word(              //convert owrd
+                      WERD *word,   //word to do
+                      TEXTROW *row  //fake row
+                     );
+TBLOB *make_tess_blobs(                      //make tess blobs
+                       PBLOB_LIST *bloblist  //list to convert
+                      );
+TBLOB *make_tess_blob(               //make tess blob
+                      PBLOB *blob,   //blob to convert
+                      BOOL8 flatten  //flatten outline structure
+                     );
+TESSLINE *make_tess_outlines(                            //make tess outlines
+                             OUTLINE_LIST *outlinelist,  //list to convert
+                             BOOL8 flatten               //flatten outline structure
+                            );
+EDGEPT *make_tess_edgepts(                          //make tess edgepts
+                          POLYPT_LIST *edgeptlist,  //list to convert
+                          TPOINT &tl,               //bounding box
+                          TPOINT &br);
+#endif

ccmain/werdit.cpp

@@ -0,0 +1,193 @@
+/**********************************************************************
+ * File:        werdit.cpp  (Formerly wordit.c)
+ * Description: An iterator for passing over all the words in a document.
+ * Author:		Ray Smith
+ * Created:		Mon Apr 27 08:51:22 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          "werdit.h"
+
+#define EXTERN
+
+//EXTERN BOOL_VAR(wordit_linearc,FALSE,"Pass poly of linearc to Tess");
+
+/**********************************************************************
+ * WERDIT::start_page
+ *
+ * Get ready to iterate over the page by setting the iterators.
+ **********************************************************************/
+
+void WERDIT::start_page(                        //set iterators
+                        BLOCK_LIST *block_list  //blocks to check
+                       ) {
+  block_it.set_to_list (block_list);
+  block_it.mark_cycle_pt ();
+  do {
+    while (block_it.data ()->row_list ()->empty ()
+    && !block_it.cycled_list ()) {
+      block_it.forward ();
+    }
+    if (!block_it.data ()->row_list ()->empty ()) {
+      row_it.set_to_list (block_it.data ()->row_list ());
+      row_it.mark_cycle_pt ();
+      while (row_it.data ()->word_list ()->empty ()
+      && !row_it.cycled_list ()) {
+        row_it.forward ();
+      }
+      if (!row_it.data ()->word_list ()->empty ()) {
+        word_it.set_to_list (row_it.data ()->word_list ());
+        word_it.mark_cycle_pt ();
+      }
+    }
+  }
+  while (!block_it.cycled_list () && row_it.data ()->word_list ()->empty ());
+}
+
+
+/**********************************************************************
+ * WERDIT::forward
+ *
+ * Give the next word on the page, or NULL if none left.
+ * This code assumes all rows to be non-empty, but blocks are allowed
+ * to be empty as eventually we will have non-text blocks.
+ * The output is always a copy and needs to be deleted by somebody.
+ **********************************************************************/
+
+WERD *WERDIT::forward() {  //use iterators
+  WERD *word;                    //actual word
+  //      WERD                                                    *larc_word;                                                     //linearc copy
+  WERD *result;                  //output word
+  ROW *row;                      //row of word
+
+  if (word_it.cycled_list ()) {
+    return NULL;                 //finished page
+  }
+  else {
+    word = word_it.data ();
+    row = row_it.data ();
+    word_it.forward ();
+    if (word_it.cycled_list ()) {
+      row_it.forward ();         //finished row
+      if (row_it.cycled_list ()) {
+        do {
+          block_it.forward ();   //finished block
+          if (!block_it.cycled_list ()) {
+            row_it.set_to_list (block_it.data ()->row_list ());
+            row_it.mark_cycle_pt ();
+          }
+        }
+                                 //find non-empty block
+        while (!block_it.cycled_list ()
+          && row_it.cycled_list ());
+      }
+      if (!row_it.cycled_list ()) {
+        word_it.set_to_list (row_it.data ()->word_list ());
+        word_it.mark_cycle_pt ();
+      }
+    }
+
+    //              if (wordit_linearc && !word->flag(W_POLYGON))
+    //              {
+    //                      larc_word=word->larc_copy(row->x_height());
+    //                      result=larc_word->poly_copy(row->x_height());
+    //                      delete larc_word;
+    //              }
+    //              else
+    result = word->poly_copy (row->x_height ());
+    return result;
+  }
+}
+
+
+/**********************************************************************
+ * make_pseudo_word
+ *
+ * Make all the blobs inside a selection into a single word.
+ * The word is always a copy and needs to be deleted.
+ **********************************************************************/
+
+WERD *make_pseudo_word(                         //make fake word
+                       BLOCK_LIST *block_list,  //blocks to check //block of selection
+                       BOX &selection_box,
+                       BLOCK *&pseudo_block,
+                       ROW *&pseudo_row         //row of selection
+                      ) {
+  BLOCK_IT block_it(block_list); 
+  BLOCK *block;
+  ROW_IT row_it;
+  ROW *row;
+  WERD_IT word_it;
+  WERD *word;
+  PBLOB_IT blob_it;
+  PBLOB *blob;
+  PBLOB_LIST new_blobs;          //list of gathered blobs
+                                 //iterator
+  PBLOB_IT new_blob_it = &new_blobs;
+  WERD *pseudo_word;             //fabricated word
+  WERD *poly_word;               //poly copy of word
+  //      WERD                                                    *larc_word;                                                     //linearc copy
+
+  for (block_it.mark_cycle_pt ();
+  !block_it.cycled_list (); block_it.forward ()) {
+    block = block_it.data ();
+    if (block->bounding_box ().overlap (selection_box)) {
+      pseudo_block = block;
+      row_it.set_to_list (block->row_list ());
+      for (row_it.mark_cycle_pt ();
+      !row_it.cycled_list (); row_it.forward ()) {
+        row = row_it.data ();
+        if (row->bounding_box ().overlap (selection_box)) {
+          word_it.set_to_list (row->word_list ());
+          for (word_it.mark_cycle_pt ();
+          !word_it.cycled_list (); word_it.forward ()) {
+            word = word_it.data ();
+            if (word->bounding_box ().overlap (selection_box)) {
+              //                                                      if (wordit_linearc && !word->flag(W_POLYGON))
+              //                                                      {
+              //                                                              larc_word=word->larc_copy(row->x_height());
+              //                                                              poly_word=larc_word->poly_copy(row->x_height());
+              //                                                              delete larc_word;
+              //                                                      }
+              //                                                      else
+              poly_word = word->poly_copy (row->x_height ());
+              blob_it.set_to_list (poly_word->blob_list ());
+              for (blob_it.mark_cycle_pt ();
+              !blob_it.cycled_list (); blob_it.forward ()) {
+                blob = blob_it.data ();
+                if (blob->bounding_box ().
+                overlap (selection_box)) {
+                  new_blob_it.add_after_then_move (blob_it.
+                    extract
+                    ());
+                                 //steal off list
+                  pseudo_row = row;
+                }
+              }
+              delete poly_word;  //get rid of it
+            }
+          }
+        }
+      }
+    }
+  }
+  if (!new_blobs.empty ()) {
+                                 //make new word
+    pseudo_word = new WERD (&new_blobs, 1, NULL);
+  }
+  else
+    pseudo_word = NULL;
+  return pseudo_word;
+}

ccmain/werdit.h

@@ -0,0 +1,67 @@
+/**********************************************************************
+ * File:			wordit.c
+ * Description: An iterator for passing over all the words in a document.
+ * Author:		Ray Smith
+ * Created:		Mon Apr 27 08:51:22 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           WERDIT_H
+#define           WERDIT_H
+
+#include          "varable.h"
+#include          "ocrblock.h"
+#include          "notdll.h"
+
+class WERDIT
+{
+  public:
+    WERDIT() { 
+    }                            //empty contructor
+    WERDIT(                          //empty contructor
+           BLOCK_LIST *blocklist) {  //blocks on page
+      start_page(blocklist);  //ready to scan
+    }
+
+    void start_page(                         //get ready
+                    BLOCK_LIST *blocklist);  //blocks on page
+
+    WERD *forward();  //get next word
+    WERD *next_word() {  //get next word
+      return word_it.data ();    //already at next
+    }
+    ROW *row() {  //get current row
+      return word_it.cycled_list ()? NULL : row_it.data ();
+    }
+    ROW *next_row() {  //get next row
+      return row_it.data_relative (1);
+    }
+    BLOCK *block() {  //get current block
+      return block_it.data ();
+    }
+
+  private:
+    BLOCK_IT block_it;           //iterators
+    ROW_IT row_it;
+    WERD_IT word_it;
+};
+
+//extern BOOL_VAR_H(wordit_linearc,FALSE,"Pass poly of linearc to Tess");
+WERD *make_pseudo_word(                         //make fake word
+                       BLOCK_LIST *block_list,  //blocks to check //block of selection
+                       BOX &selection_box,
+                       BLOCK *&pseudo_block,
+                       ROW *&pseudo_row         //row of selection
+                      );
+#endif

ccstruct/Makefile.am

@@ -0,0 +1,25 @@
+SUBDIRS =
+AM_CPPFLAGS = \
+    -I$(top_srcdir)/ccutil -I$(top_srcdir)/cutil \
+    -I$(top_srcdir)/image -I$(top_srcdir)/viewer
+
+EXTRA_DIST = \
+    blckerr.h blobbox.h blobs.h blread.h coutln.h crakedge.h \
+    genblob.h hpddef.h hpdsizes.h ipoints.h labls.h linlsq.h \
+    lmedsq.h mod128.h normalis.h ocrblock.h ocrrow.h pageblk.h \
+    pageres.h pdblock.h pdclass.h points.h polyaprx.h polyblk.h \
+    polyblob.h polyvert.h poutline.h quadlsq.h quadratc.h \
+    quspline.h ratngs.h rect.h rejctmap.h rwpoly.h statistc.h \
+    stepblob.h txtregn.h vecfuncs.h werd.h
+
+noinst_LIBRARIES = libtesseract_ccstruct.a
+libtesseract_ccstruct_a_SOURCES = \
+    blobbox.cpp blobs.cpp blread.cpp callcpp.cpp \
+    coutln.cpp genblob.cpp labls.cpp linlsq.cpp \
+    lmedsq.cpp mod128.cpp normalis.cpp ocrblock.cpp \
+    ocrrow.cpp pageblk.cpp pageres.cpp pdblock.cpp \
+    points.cpp polyaprx.cpp polyblk.cpp polyblob.cpp \
+    polyvert.cpp poutline.cpp quadlsq.cpp quadratc.cpp \
+    quspline.cpp ratngs.cpp rect.cpp rejctmap.cpp \
+    rwpoly.cpp statistc.cpp stepblob.cpp txtregn.cpp \
+    vecfuncs.cpp werd.cpp

ccstruct/Makefile.in

@@ -0,0 +1,587 @@
+# Makefile.in generated by automake 1.9.6 from Makefile.am.
+# @configure_input@
+
+# Copyright (C) 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002,
+# 2003, 2004, 2005  Free Software Foundation, Inc.
+# This Makefile.in is free software; the Free Software Foundation
+# gives unlimited permission to copy and/or distribute it,
+# with or without modifications, as long as this notice is preserved.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY, to the extent permitted by law; without
+# even the implied warranty of MERCHANTABILITY or FITNESS FOR A
+# PARTICULAR PURPOSE.
+
+@SET_MAKE@
+
+srcdir = @srcdir@
+top_srcdir = @top_srcdir@
+VPATH = @srcdir@
+pkgdatadir = $(datadir)/@PACKAGE@
+pkglibdir = $(libdir)/@PACKAGE@
+pkgincludedir = $(includedir)/@PACKAGE@
+top_builddir = ..
+am__cd = CDPATH="$${ZSH_VERSION+.}$(PATH_SEPARATOR)" && cd
+INSTALL = @INSTALL@
+install_sh_DATA = $(install_sh) -c -m 644
+install_sh_PROGRAM = $(install_sh) -c
+install_sh_SCRIPT = $(install_sh) -c
+INSTALL_HEADER = $(INSTALL_DATA)
+transform = $(program_transform_name)
+NORMAL_INSTALL = :
+PRE_INSTALL = :
+POST_INSTALL = :
+NORMAL_UNINSTALL = :
+PRE_UNINSTALL = :
+POST_UNINSTALL = :
+build_triplet = @build@
+host_triplet = @host@
+subdir = ccstruct
+DIST_COMMON = $(srcdir)/Makefile.am $(srcdir)/Makefile.in
+ACLOCAL_M4 = $(top_srcdir)/aclocal.m4
+am__aclocal_m4_deps = $(top_srcdir)/acinclude.m4 \
+	$(top_srcdir)/config/ac_define_versionlevel.m4 \
+	$(top_srcdir)/config/acinclude_custom.m4 \
+	$(top_srcdir)/configure.ac
+am__configure_deps = $(am__aclocal_m4_deps) $(CONFIGURE_DEPENDENCIES) \
+	$(ACLOCAL_M4)
+mkinstalldirs = $(SHELL) $(top_srcdir)/config/mkinstalldirs
+CONFIG_HEADER = $(top_builddir)/config_auto.h
+CONFIG_CLEAN_FILES =
+LIBRARIES = $(noinst_LIBRARIES)
+AR = ar
+ARFLAGS = cru
+libtesseract_ccstruct_a_AR = $(AR) $(ARFLAGS)
+libtesseract_ccstruct_a_LIBADD =
+am_libtesseract_ccstruct_a_OBJECTS = blobbox.$(OBJEXT) blobs.$(OBJEXT) \
+	blread.$(OBJEXT) callcpp.$(OBJEXT) coutln.$(OBJEXT) \
+	genblob.$(OBJEXT) labls.$(OBJEXT) linlsq.$(OBJEXT) \
+	lmedsq.$(OBJEXT) mod128.$(OBJEXT) normalis.$(OBJEXT) \
+	ocrblock.$(OBJEXT) ocrrow.$(OBJEXT) pageblk.$(OBJEXT) \
+	pageres.$(OBJEXT) pdblock.$(OBJEXT) points.$(OBJEXT) \
+	polyaprx.$(OBJEXT) polyblk.$(OBJEXT) polyblob.$(OBJEXT) \
+	polyvert.$(OBJEXT) poutline.$(OBJEXT) quadlsq.$(OBJEXT) \
+	quadratc.$(OBJEXT) quspline.$(OBJEXT) ratngs.$(OBJEXT) \
+	rect.$(OBJEXT) rejctmap.$(OBJEXT) rwpoly.$(OBJEXT) \
+	statistc.$(OBJEXT) stepblob.$(OBJEXT) txtregn.$(OBJEXT) \
+	vecfuncs.$(OBJEXT) werd.$(OBJEXT)
+libtesseract_ccstruct_a_OBJECTS =  \
+	$(am_libtesseract_ccstruct_a_OBJECTS)
+DEFAULT_INCLUDES = -I. -I$(srcdir) -I$(top_builddir)
+depcomp = $(SHELL) $(top_srcdir)/config/depcomp
+am__depfiles_maybe = depfiles
+CXXCOMPILE = $(CXX) $(DEFS) $(DEFAULT_INCLUDES) $(INCLUDES) \
+	$(AM_CPPFLAGS) $(CPPFLAGS) $(AM_CXXFLAGS) $(CXXFLAGS)
+CXXLD = $(CXX)
+CXXLINK = $(CXXLD) $(AM_CXXFLAGS) $(CXXFLAGS) $(AM_LDFLAGS) $(LDFLAGS) \
+	-o $@
+SOURCES = $(libtesseract_ccstruct_a_SOURCES)
+DIST_SOURCES = $(libtesseract_ccstruct_a_SOURCES)
+RECURSIVE_TARGETS = all-recursive check-recursive dvi-recursive \
+	html-recursive info-recursive install-data-recursive \
+	install-exec-recursive install-info-recursive \
+	install-recursive installcheck-recursive installdirs-recursive \
+	pdf-recursive ps-recursive uninstall-info-recursive \
+	uninstall-recursive
+ETAGS = etags
+CTAGS = ctags
+DIST_SUBDIRS = $(SUBDIRS)
+DISTFILES = $(DIST_COMMON) $(DIST_SOURCES) $(TEXINFOS) $(EXTRA_DIST)
+ACLOCAL = @ACLOCAL@
+AMDEP_FALSE = @AMDEP_FALSE@
+AMDEP_TRUE = @AMDEP_TRUE@
+AMTAR = @AMTAR@
+AUTOCONF = @AUTOCONF@
+AUTOHEADER = @AUTOHEADER@
+AUTOMAKE = @AUTOMAKE@
+AWK = @AWK@
+CC = @CC@
+CCDEPMODE = @CCDEPMODE@
+CFLAGS = @CFLAGS@
+CPPFLAGS = @CPPFLAGS@
+CXX = @CXX@
+CXXCPP = @CXXCPP@
+CXXDEPMODE = @CXXDEPMODE@
+CXXFLAGS = @CXXFLAGS@
+CXXRPOFLAGS = @CXXRPOFLAGS@
+CYGPATH_W = @CYGPATH_W@
+DEFS = @DEFS@
+DEPDIR = @DEPDIR@
+ECHO_C = @ECHO_C@
+ECHO_N = @ECHO_N@
+ECHO_T = @ECHO_T@
+EGREP = @EGREP@
+EXEEXT = @EXEEXT@
+GNUWIN32_DIR = @GNUWIN32_DIR@
+HAVE_GNUWIN32_FALSE = @HAVE_GNUWIN32_FALSE@
+HAVE_GNUWIN32_TRUE = @HAVE_GNUWIN32_TRUE@
+HAVE_LIBTIFF_FALSE = @HAVE_LIBTIFF_FALSE@
+HAVE_LIBTIFF_TRUE = @HAVE_LIBTIFF_TRUE@
+INSTALL_DATA = @INSTALL_DATA@
+INSTALL_PROGRAM = @INSTALL_PROGRAM@
+INSTALL_SCRIPT = @INSTALL_SCRIPT@
+INSTALL_STRIP_PROGRAM = @INSTALL_STRIP_PROGRAM@
+LDFLAGS = @LDFLAGS@
+LIBOBJS = @LIBOBJS@
+LIBS = @LIBS@
+LIBTIFF_CFLAGS = @LIBTIFF_CFLAGS@
+LIBTIFF_LIBS = @LIBTIFF_LIBS@
+LTLIBOBJS = @LTLIBOBJS@
+MAINT = @MAINT@
+MAINTAINER_MODE_FALSE = @MAINTAINER_MODE_FALSE@
+MAINTAINER_MODE_TRUE = @MAINTAINER_MODE_TRUE@
+MAKEINFO = @MAKEINFO@
+OBJEXT = @OBJEXT@
+OPTS = @OPTS@
+PACKAGE = @PACKAGE@
+PACKAGE_BUGREPORT = @PACKAGE_BUGREPORT@
+PACKAGE_DATE = @PACKAGE_DATE@
+PACKAGE_NAME = @PACKAGE_NAME@
+PACKAGE_STRING = @PACKAGE_STRING@
+PACKAGE_TARNAME = @PACKAGE_TARNAME@
+PACKAGE_VERSION = @PACKAGE_VERSION@
+PACKAGE_YEAR = @PACKAGE_YEAR@
+PATH_SEPARATOR = @PATH_SEPARATOR@
+RANLIB = @RANLIB@
+RPO_NO = @RPO_NO@
+RPO_YES = @RPO_YES@
+SET_MAKE = @SET_MAKE@
+SHELL = @SHELL@
+STRIP = @STRIP@
+USING_CL_FALSE = @USING_CL_FALSE@
+USING_CL_TRUE = @USING_CL_TRUE@
+VERSION = @VERSION@
+ac_ct_CC = @ac_ct_CC@
+ac_ct_CXX = @ac_ct_CXX@
+ac_ct_RANLIB = @ac_ct_RANLIB@
+ac_ct_STRIP = @ac_ct_STRIP@
+am__fastdepCC_FALSE = @am__fastdepCC_FALSE@
+am__fastdepCC_TRUE = @am__fastdepCC_TRUE@
+am__fastdepCXX_FALSE = @am__fastdepCXX_FALSE@
+am__fastdepCXX_TRUE = @am__fastdepCXX_TRUE@
+am__include = @am__include@
+am__leading_dot = @am__leading_dot@
+am__quote = @am__quote@
+am__tar = @am__tar@
+am__untar = @am__untar@
+bindir = @bindir@
+build = @build@
+build_alias = @build_alias@
+build_cpu = @build_cpu@
+build_os = @build_os@
+build_vendor = @build_vendor@
+datadir = @datadir@
+exec_prefix = @exec_prefix@
+host = @host@
+host_alias = @host_alias@
+host_cpu = @host_cpu@
+host_os = @host_os@
+host_vendor = @host_vendor@
+includedir = @includedir@
+infodir = @infodir@
+install_sh = @install_sh@
+libdir = @libdir@
+libexecdir = @libexecdir@
+localstatedir = @localstatedir@
+mandir = @mandir@
+mkdir_p = @mkdir_p@
+oldincludedir = @oldincludedir@
+prefix = @prefix@
+program_transform_name = @program_transform_name@
+sbindir = @sbindir@
+sharedstatedir = @sharedstatedir@
+sysconfdir = @sysconfdir@
+target_alias = @target_alias@
+SUBDIRS = 
+AM_CPPFLAGS = \
+    -I$(top_srcdir)/ccutil -I$(top_srcdir)/cutil \
+    -I$(top_srcdir)/image -I$(top_srcdir)/viewer
+
+EXTRA_DIST = \
+    blckerr.h blobbox.h blobs.h blread.h coutln.h crakedge.h \
+    genblob.h hpddef.h hpdsizes.h ipoints.h labls.h linlsq.h \
+    lmedsq.h mod128.h normalis.h ocrblock.h ocrrow.h pageblk.h \
+    pageres.h pdblock.h pdclass.h points.h polyaprx.h polyblk.h \
+    polyblob.h polyvert.h poutline.h quadlsq.h quadratc.h \
+    quspline.h ratngs.h rect.h rejctmap.h rwpoly.h statistc.h \
+    stepblob.h txtregn.h vecfuncs.h werd.h
+
+noinst_LIBRARIES = libtesseract_ccstruct.a
+libtesseract_ccstruct_a_SOURCES = \
+    blobbox.cpp blobs.cpp blread.cpp callcpp.cpp \
+    coutln.cpp genblob.cpp labls.cpp linlsq.cpp \
+    lmedsq.cpp mod128.cpp normalis.cpp ocrblock.cpp \
+    ocrrow.cpp pageblk.cpp pageres.cpp pdblock.cpp \
+    points.cpp polyaprx.cpp polyblk.cpp polyblob.cpp \
+    polyvert.cpp poutline.cpp quadlsq.cpp quadratc.cpp \
+    quspline.cpp ratngs.cpp rect.cpp rejctmap.cpp \
+    rwpoly.cpp statistc.cpp stepblob.cpp txtregn.cpp \
+    vecfuncs.cpp werd.cpp
+
+all: all-recursive
+
+.SUFFIXES:
+.SUFFIXES: .cpp .o .obj
+$(srcdir)/Makefile.in: @MAINTAINER_MODE_TRUE@ $(srcdir)/Makefile.am  $(am__configure_deps)
+	@for dep in $?; do \
+	  case '$(am__configure_deps)' in \
+	    *$$dep*) \
+	      cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh \
+		&& exit 0; \
+	      exit 1;; \
+	  esac; \
+	done; \
+	echo ' cd $(top_srcdir) && $(AUTOMAKE) --gnu  ccstruct/Makefile'; \
+	cd $(top_srcdir) && \
+	  $(AUTOMAKE) --gnu  ccstruct/Makefile
+.PRECIOUS: Makefile
+Makefile: $(srcdir)/Makefile.in $(top_builddir)/config.status
+	@case '$?' in \
+	  *config.status*) \
+	    cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh;; \
+	  *) \
+	    echo ' cd $(top_builddir) && $(SHELL) ./config.status $(subdir)/$@ $(am__depfiles_maybe)'; \
+	    cd $(top_builddir) && $(SHELL) ./config.status $(subdir)/$@ $(am__depfiles_maybe);; \
+	esac;
+
+$(top_builddir)/config.status: $(top_srcdir)/configure $(CONFIG_STATUS_DEPENDENCIES)
+	cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh
+
+$(top_srcdir)/configure: @MAINTAINER_MODE_TRUE@ $(am__configure_deps)
+	cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh
+$(ACLOCAL_M4): @MAINTAINER_MODE_TRUE@ $(am__aclocal_m4_deps)
+	cd $(top_builddir) && $(MAKE) $(AM_MAKEFLAGS) am--refresh
+
+clean-noinstLIBRARIES:
+	-test -z "$(noinst_LIBRARIES)" || rm -f $(noinst_LIBRARIES)
+libtesseract_ccstruct.a: $(libtesseract_ccstruct_a_OBJECTS) $(libtesseract_ccstruct_a_DEPENDENCIES) 
+	-rm -f libtesseract_ccstruct.a
+	$(libtesseract_ccstruct_a_AR) libtesseract_ccstruct.a $(libtesseract_ccstruct_a_OBJECTS) $(libtesseract_ccstruct_a_LIBADD)
+	$(RANLIB) libtesseract_ccstruct.a
+
+mostlyclean-compile:
+	-rm -f *.$(OBJEXT)
+
+distclean-compile:
+	-rm -f *.tab.c
+
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/blobbox.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/blobs.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/blread.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/callcpp.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/coutln.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/genblob.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/labls.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/linlsq.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/lmedsq.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/mod128.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/normalis.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/ocrblock.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/ocrrow.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/pageblk.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/pageres.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/pdblock.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/points.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/polyaprx.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/polyblk.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/polyblob.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/polyvert.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/poutline.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/quadlsq.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/quadratc.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/quspline.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/ratngs.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/rect.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/rejctmap.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/rwpoly.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/statistc.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/stepblob.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/txtregn.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/vecfuncs.Po@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/werd.Po@am__quote@
+
+.cpp.o:
+@am__fastdepCXX_TRUE@	if $(CXXCOMPILE) -MT $@ -MD -MP -MF "$(DEPDIR)/$*.Tpo" -c -o $@ $<; \
+@am__fastdepCXX_TRUE@	then mv -f "$(DEPDIR)/$*.Tpo" "$(DEPDIR)/$*.Po"; else rm -f "$(DEPDIR)/$*.Tpo"; exit 1; fi
+@AMDEP_TRUE@@am__fastdepCXX_FALSE@	source='$<' object='$@' libtool=no @AMDEPBACKSLASH@
+@AMDEP_TRUE@@am__fastdepCXX_FALSE@	DEPDIR=$(DEPDIR) $(CXXDEPMODE) $(depcomp) @AMDEPBACKSLASH@
+@am__fastdepCXX_FALSE@	$(CXXCOMPILE) -c -o $@ $<
+
+.cpp.obj:
+@am__fastdepCXX_TRUE@	if $(CXXCOMPILE) -MT $@ -MD -MP -MF "$(DEPDIR)/$*.Tpo" -c -o $@ `$(CYGPATH_W) '$<'`; \
+@am__fastdepCXX_TRUE@	then mv -f "$(DEPDIR)/$*.Tpo" "$(DEPDIR)/$*.Po"; else rm -f "$(DEPDIR)/$*.Tpo"; exit 1; fi
+@AMDEP_TRUE@@am__fastdepCXX_FALSE@	source='$<' object='$@' libtool=no @AMDEPBACKSLASH@
+@AMDEP_TRUE@@am__fastdepCXX_FALSE@	DEPDIR=$(DEPDIR) $(CXXDEPMODE) $(depcomp) @AMDEPBACKSLASH@
+@am__fastdepCXX_FALSE@	$(CXXCOMPILE) -c -o $@ `$(CYGPATH_W) '$<'`
+uninstall-info-am:
+
+# This directory's subdirectories are mostly independent; you can cd
+# into them and run `make' without going through this Makefile.
+# To change the values of `make' variables: instead of editing Makefiles,
+# (1) if the variable is set in `config.status', edit `config.status'
+#     (which will cause the Makefiles to be regenerated when you run `make');
+# (2) otherwise, pass the desired values on the `make' command line.
+$(RECURSIVE_TARGETS):
+	@failcom='exit 1'; \
+	for f in x $$MAKEFLAGS; do \
+	  case $$f in \
+	    *=* | --[!k]*);; \
+	    *k*) failcom='fail=yes';; \
+	  esac; \
+	done; \
+	dot_seen=no; \
+	target=`echo $@ | sed s/-recursive//`; \
+	list='$(SUBDIRS)'; for subdir in $$list; do \
+	  echo "Making $$target in $$subdir"; \
+	  if test "$$subdir" = "."; then \
+	    dot_seen=yes; \
+	    local_target="$$target-am"; \
+	  else \
+	    local_target="$$target"; \
+	  fi; \
+	  (cd $$subdir && $(MAKE) $(AM_MAKEFLAGS) $$local_target) \
+	  || eval $$failcom; \
+	done; \
+	if test "$$dot_seen" = "no"; then \
+	  $(MAKE) $(AM_MAKEFLAGS) "$$target-am" || exit 1; \
+	fi; test -z "$$fail"
+
+mostlyclean-recursive clean-recursive distclean-recursive \
+maintainer-clean-recursive:
+	@failcom='exit 1'; \
+	for f in x $$MAKEFLAGS; do \
+	  case $$f in \
+	    *=* | --[!k]*);; \
+	    *k*) failcom='fail=yes';; \
+	  esac; \
+	done; \
+	dot_seen=no; \
+	case "$@" in \
+	  distclean-* | maintainer-clean-*) list='$(DIST_SUBDIRS)' ;; \
+	  *) list='$(SUBDIRS)' ;; \
+	esac; \
+	rev=''; for subdir in $$list; do \
+	  if test "$$subdir" = "."; then :; else \
+	    rev="$$subdir $$rev"; \
+	  fi; \
+	done; \
+	rev="$$rev ."; \
+	target=`echo $@ | sed s/-recursive//`; \
+	for subdir in $$rev; do \
+	  echo "Making $$target in $$subdir"; \
+	  if test "$$subdir" = "."; then \
+	    local_target="$$target-am"; \
+	  else \
+	    local_target="$$target"; \
+	  fi; \
+	  (cd $$subdir && $(MAKE) $(AM_MAKEFLAGS) $$local_target) \
+	  || eval $$failcom; \
+	done && test -z "$$fail"
+tags-recursive:
+	list='$(SUBDIRS)'; for subdir in $$list; do \
+	  test "$$subdir" = . || (cd $$subdir && $(MAKE) $(AM_MAKEFLAGS) tags); \
+	done
+ctags-recursive:
+	list='$(SUBDIRS)'; for subdir in $$list; do \
+	  test "$$subdir" = . || (cd $$subdir && $(MAKE) $(AM_MAKEFLAGS) ctags); \
+	done
+
+ID: $(HEADERS) $(SOURCES) $(LISP) $(TAGS_FILES)
+	list='$(SOURCES) $(HEADERS) $(LISP) $(TAGS_FILES)'; \
+	unique=`for i in $$list; do \
+	    if test -f "$$i"; then echo $$i; else echo $(srcdir)/$$i; fi; \
+	  done | \
+	  $(AWK) '    { files[$$0] = 1; } \
+	       END { for (i in files) print i; }'`; \
+	mkid -fID $$unique
+tags: TAGS
+
+TAGS: tags-recursive $(HEADERS) $(SOURCES)  $(TAGS_DEPENDENCIES) \
+		$(TAGS_FILES) $(LISP)
+	tags=; \
+	here=`pwd`; \
+	if ($(ETAGS) --etags-include --version) >/dev/null 2>&1; then \
+	  include_option=--etags-include; \
+	  empty_fix=.; \
+	else \
+	  include_option=--include; \
+	  empty_fix=; \
+	fi; \
+	list='$(SUBDIRS)'; for subdir in $$list; do \
+	  if test "$$subdir" = .; then :; else \
+	    test ! -f $$subdir/TAGS || \
+	      tags="$$tags $$include_option=$$here/$$subdir/TAGS"; \
+	  fi; \
+	done; \
+	list='$(SOURCES) $(HEADERS)  $(LISP) $(TAGS_FILES)'; \
+	unique=`for i in $$list; do \
+	    if test -f "$$i"; then echo $$i; else echo $(srcdir)/$$i; fi; \
+	  done | \
+	  $(AWK) '    { files[$$0] = 1; } \
+	       END { for (i in files) print i; }'`; \
+	if test -z "$(ETAGS_ARGS)$$tags$$unique"; then :; else \
+	  test -n "$$unique" || unique=$$empty_fix; \
+	  $(ETAGS) $(ETAGSFLAGS) $(AM_ETAGSFLAGS) $(ETAGS_ARGS) \
+	    $$tags $$unique; \
+	fi
+ctags: CTAGS
+CTAGS: ctags-recursive $(HEADERS) $(SOURCES)  $(TAGS_DEPENDENCIES) \
+		$(TAGS_FILES) $(LISP)
+	tags=; \
+	here=`pwd`; \
+	list='$(SOURCES) $(HEADERS)  $(LISP) $(TAGS_FILES)'; \
+	unique=`for i in $$list; do \
+	    if test -f "$$i"; then echo $$i; else echo $(srcdir)/$$i; fi; \
+	  done | \
+	  $(AWK) '    { files[$$0] = 1; } \
+	       END { for (i in files) print i; }'`; \
+	test -z "$(CTAGS_ARGS)$$tags$$unique" \
+	  || $(CTAGS) $(CTAGSFLAGS) $(AM_CTAGSFLAGS) $(CTAGS_ARGS) \
+	     $$tags $$unique
+
+GTAGS:
+	here=`$(am__cd) $(top_builddir) && pwd` \
+	  && cd $(top_srcdir) \
+	  && gtags -i $(GTAGS_ARGS) $$here
+
+distclean-tags:
+	-rm -f TAGS ID GTAGS GRTAGS GSYMS GPATH tags
+
+distdir: $(DISTFILES)
+	@srcdirstrip=`echo "$(srcdir)" | sed 's|.|.|g'`; \
+	topsrcdirstrip=`echo "$(top_srcdir)" | sed 's|.|.|g'`; \
+	list='$(DISTFILES)'; for file in $$list; do \
+	  case $$file in \
+	    $(srcdir)/*) file=`echo "$$file" | sed "s|^$$srcdirstrip/||"`;; \
+	    $(top_srcdir)/*) file=`echo "$$file" | sed "s|^$$topsrcdirstrip/|$(top_builddir)/|"`;; \
+	  esac; \
+	  if test -f $$file || test -d $$file; then d=.; else d=$(srcdir); fi; \
+	  dir=`echo "$$file" | sed -e 's,/[^/]*$$,,'`; \
+	  if test "$$dir" != "$$file" && test "$$dir" != "."; then \
+	    dir="/$$dir"; \
+	    $(mkdir_p) "$(distdir)$$dir"; \
+	  else \
+	    dir=''; \
+	  fi; \
+	  if test -d $$d/$$file; then \
+	    if test -d $(srcdir)/$$file && test $$d != $(srcdir); then \
+	      cp -pR $(srcdir)/$$file $(distdir)$$dir || exit 1; \
+	    fi; \
+	    cp -pR $$d/$$file $(distdir)$$dir || exit 1; \
+	  else \
+	    test -f $(distdir)/$$file \
+	    || cp -p $$d/$$file $(distdir)/$$file \
+	    || exit 1; \
+	  fi; \
+	done
+	list='$(DIST_SUBDIRS)'; for subdir in $$list; do \
+	  if test "$$subdir" = .; then :; else \
+	    test -d "$(distdir)/$$subdir" \
+	    || $(mkdir_p) "$(distdir)/$$subdir" \
+	    || exit 1; \
+	    distdir=`$(am__cd) $(distdir) && pwd`; \
+	    top_distdir=`$(am__cd) $(top_distdir) && pwd`; \
+	    (cd $$subdir && \
+	      $(MAKE) $(AM_MAKEFLAGS) \
+	        top_distdir="$$top_distdir" \
+	        distdir="$$distdir/$$subdir" \
+	        distdir) \
+	      || exit 1; \
+	  fi; \
+	done
+check-am: all-am
+check: check-recursive
+all-am: Makefile $(LIBRARIES)
+installdirs: installdirs-recursive
+installdirs-am:
+install: install-recursive
+install-exec: install-exec-recursive
+install-data: install-data-recursive
+uninstall: uninstall-recursive
+
+install-am: all-am
+	@$(MAKE) $(AM_MAKEFLAGS) install-exec-am install-data-am
+
+installcheck: installcheck-recursive
+install-strip:
+	$(MAKE) $(AM_MAKEFLAGS) INSTALL_PROGRAM="$(INSTALL_STRIP_PROGRAM)" \
+	  install_sh_PROGRAM="$(INSTALL_STRIP_PROGRAM)" INSTALL_STRIP_FLAG=-s \
+	  `test -z '$(STRIP)' || \
+	    echo "INSTALL_PROGRAM_ENV=STRIPPROG='$(STRIP)'"` install
+mostlyclean-generic:
+
+clean-generic:
+
+distclean-generic:
+	-test -z "$(CONFIG_CLEAN_FILES)" || rm -f $(CONFIG_CLEAN_FILES)
+
+maintainer-clean-generic:
+	@echo "This command is intended for maintainers to use"
+	@echo "it deletes files that may require special tools to rebuild."
+clean: clean-recursive
+
+clean-am: clean-generic clean-noinstLIBRARIES mostlyclean-am
+
+distclean: distclean-recursive
+	-rm -rf ./$(DEPDIR)
+	-rm -f Makefile
+distclean-am: clean-am distclean-compile distclean-generic \
+	distclean-tags
+
+dvi: dvi-recursive
+
+dvi-am:
+
+html: html-recursive
+
+info: info-recursive
+
+info-am:
+
+install-data-am:
+
+install-exec-am:
+
+install-info: install-info-recursive
+
+install-man:
+
+installcheck-am:
+
+maintainer-clean: maintainer-clean-recursive
+	-rm -rf ./$(DEPDIR)
+	-rm -f Makefile
+maintainer-clean-am: distclean-am maintainer-clean-generic
+
+mostlyclean: mostlyclean-recursive
+
+mostlyclean-am: mostlyclean-compile mostlyclean-generic
+
+pdf: pdf-recursive
+
+pdf-am:
+
+ps: ps-recursive
+
+ps-am:
+
+uninstall-am: uninstall-info-am
+
+uninstall-info: uninstall-info-recursive
+
+.PHONY: $(RECURSIVE_TARGETS) CTAGS GTAGS all all-am check check-am \
+	clean clean-generic clean-noinstLIBRARIES clean-recursive \
+	ctags ctags-recursive distclean distclean-compile \
+	distclean-generic distclean-recursive distclean-tags distdir \
+	dvi dvi-am html html-am info info-am install install-am \
+	install-data install-data-am install-exec install-exec-am \
+	install-info install-info-am install-man install-strip \
+	installcheck installcheck-am installdirs installdirs-am \
+	maintainer-clean maintainer-clean-generic \
+	maintainer-clean-recursive mostlyclean mostlyclean-compile \
+	mostlyclean-generic mostlyclean-recursive pdf pdf-am ps ps-am \
+	tags tags-recursive uninstall uninstall-am uninstall-info-am
+
+# Tell versions [3.59,3.63) of GNU make to not export all variables.
+# Otherwise a system limit (for SysV at least) may be exceeded.
+.NOEXPORT:

ccstruct/blckerr.h

@@ -0,0 +1,29 @@
+/**********************************************************************
+ * File:        blckerr.h  (Formerly blockerr.h)
+ * Description: Error codes for the page block classes.
+ * Author:					Ray Smith
+ * Created:					Tue Mar 19 17:43:30 GMT 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           BLCKERR_H
+#define           BLCKERR_H
+
+#include          "errcode.h"
+
+const ERRCODE BADBLOCKLINE = "Y coordinate in block out of bounds";
+const ERRCODE LOSTBLOCKLINE = "Can't find rectangle for line";
+const ERRCODE ILLEGAL_GRADIENT = "Gradient wrong side of edge step!";
+const ERRCODE WRONG_WORD = "Word doesn't have blobs of that type";
+#endif

ccstruct/blobbox.cpp

@@ -0,0 +1,778 @@
+/**********************************************************************
+ * File:        blobbox.cpp  (Formerly blobnbox.c)
+ * Description: Code for the textord blob class.
+ * Author:					Ray Smith
+ * Created:					Thu Jul 30 09:08:51 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          "blobbox.h"
+
+#define PROJECTION_MARGIN 10     //arbitrary
+#define EXTERN
+
+EXTERN double_VAR (textord_error_weight, 3,
+"Weighting for error in believability");
+EXTERN BOOL_VAR (pitsync_projection_fix, TRUE,
+"Fix bug in projection profile");
+
+ELISTIZE (BLOBNBOX) ELIST2IZE (TO_ROW) ELISTIZE (TO_BLOCK)
+/**********************************************************************
+ * BLOBNBOX::merge
+ *
+ * Merge this blob with the given blob, which should be after this.
+ **********************************************************************/
+void BLOBNBOX::merge(                    //merge blobs
+                     BLOBNBOX *nextblob  //blob to join with
+                    ) {
+  box += nextblob->box;          //merge boxes
+  nextblob->joined = TRUE;
+}
+
+
+/**********************************************************************
+ * BLOBNBOX::chop
+ *
+ * Chop this blob into equal sized pieces using the x height as a guide.
+ * The blob is not actually chopped. Instead, fake blobs are inserted
+ * with the relevant bounding boxes.
+ **********************************************************************/
+
+void BLOBNBOX::chop(                        //chop blobs
+                    BLOBNBOX_IT *start_it,  //location of this
+                    BLOBNBOX_IT *end_it,    //iterator
+                    FCOORD rotation,        //for landscape
+                    float xheight           //of line
+                   ) {
+  INT16 blobcount;               //no of blobs
+  BLOBNBOX *newblob;             //fake blob
+  BLOBNBOX *blob;                //current blob
+  INT16 blobindex;               //number of chop
+  INT16 leftx;                   //left edge of blob
+  float blobwidth;               //width of each
+  float rightx;                  //right edge to scan
+  float ymin, ymax;              //limits of new blob
+  float test_ymin, test_ymax;    //limits of part blob
+  ICOORD bl, tr;                 //corners of box
+  BLOBNBOX_IT blob_it;           //blob iterator
+
+                                 //get no of chops
+  blobcount = (INT16) floor (box.width () / xheight);
+  if (blobcount > 1 && (blob_ptr != NULL || cblob_ptr != NULL)) {
+                                 //width of each
+    blobwidth = (float) (box.width () + 1) / blobcount;
+    for (blobindex = blobcount - 1, rightx = box.right ();
+    blobindex >= 0; blobindex--, rightx -= blobwidth) {
+      ymin = (float) MAX_INT32;
+      ymax = (float) -MAX_INT32;
+      blob_it = *start_it;
+      do {
+        blob = blob_it.data ();
+        if (blob->blob_ptr != NULL)
+          find_blob_limits (blob->blob_ptr, rightx - blobwidth, rightx,
+            rotation, test_ymin, test_ymax);
+        else
+          find_cblob_vlimits (blob->cblob_ptr, rightx - blobwidth,
+            rightx,
+            /*rotation, */ test_ymin, test_ymax);
+        blob_it.forward ();
+        if (test_ymin < ymin)
+          ymin = test_ymin;
+        if (test_ymax > ymax)
+          ymax = test_ymax;
+      }
+      while (blob != end_it->data ());
+      if (ymin < ymax) {
+        leftx = (INT16) floor (rightx - blobwidth);
+        if (leftx < box.left ())
+          leftx = box.left ();   //clip to real box
+        bl = ICOORD (leftx, (INT16) floor (ymin));
+        tr = ICOORD ((INT16) ceil (rightx), (INT16) ceil (ymax));
+        if (blobindex == 0)
+          box = BOX (bl, tr);    //change box
+        else {
+          newblob = new BLOBNBOX;
+                                 //box is all it has
+          newblob->box = BOX (bl, tr);
+                                 //stay on current
+          end_it->add_after_stay_put (newblob);
+        }
+      }
+    }
+  }
+}
+
+
+/**********************************************************************
+ * find_blob_limits
+ *
+ * Scan the outlines of the blob to locate the y min and max
+ * between the given x limits.
+ **********************************************************************/
+
+void find_blob_limits(                  //get y limits
+                      PBLOB *blob,      //blob to search
+                      float leftx,      //x limits
+                      float rightx,
+                      FCOORD rotation,  //for landscape
+                      float &ymin,      //output y limits
+                      float &ymax) {
+  float testy;                   //y intercept
+  FCOORD pos;                    //rotated
+  FCOORD vec;
+  POLYPT *polypt;                //current point
+                                 //outlines
+  OUTLINE_IT out_it = blob->out_list ();
+  POLYPT_IT poly_it;             //outline pts
+
+  ymin = (float) MAX_INT32;
+  ymax = (float) -MAX_INT32;
+  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
+                                 //get points
+    poly_it.set_to_list (out_it.data ()->polypts ());
+    for (poly_it.mark_cycle_pt (); !poly_it.cycled_list ();
+    poly_it.forward ()) {
+      polypt = poly_it.data ();
+      pos = polypt->pos;
+      pos.rotate (rotation);
+      vec = polypt->vec;
+      vec.rotate (rotation);
+      if (pos.x () < leftx && pos.x () + vec.x () > leftx
+      || pos.x () > leftx && pos.x () + vec.x () < leftx) {
+        testy = pos.y () + vec.y () * (leftx - pos.x ()) / vec.x ();
+        //intercept of boundary
+        if (testy < ymin)
+          ymin = testy;
+        if (testy > ymax)
+          ymax = testy;
+      }
+      if (pos.x () >= leftx && pos.x () <= rightx) {
+        if (pos.y () > ymax)
+          ymax = pos.y ();
+        if (pos.y () < ymin)
+          ymin = pos.y ();
+      }
+      if (pos.x () > rightx && pos.x () + vec.x () < rightx
+      || pos.x () < rightx && pos.x () + vec.x () > rightx) {
+        testy = pos.y () + vec.y () * (rightx - pos.x ()) / vec.x ();
+        //intercept of boundary
+        if (testy < ymin)
+          ymin = testy;
+        if (testy > ymax)
+          ymax = testy;
+      }
+    }
+  }
+}
+
+
+/**********************************************************************
+ * find_cblob_limits
+ *
+ * Scan the outlines of the cblob to locate the y min and max
+ * between the given x limits.
+ **********************************************************************/
+
+void find_cblob_limits(                  //get y limits
+                       C_BLOB *blob,     //blob to search
+                       float leftx,      //x limits
+                       float rightx,
+                       FCOORD rotation,  //for landscape
+                       float &ymin,      //output y limits
+                       float &ymax) {
+  INT16 stepindex;               //current point
+  ICOORD pos;                    //current coords
+  ICOORD vec;                    //rotated step
+  C_OUTLINE *outline;            //current outline
+                                 //outlines
+  C_OUTLINE_IT out_it = blob->out_list ();
+
+  ymin = (float) MAX_INT32;
+  ymax = (float) -MAX_INT32;
+  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
+    outline = out_it.data ();
+    pos = outline->start_pos (); //get coords
+    pos.rotate (rotation);
+    for (stepindex = 0; stepindex < outline->pathlength (); stepindex++) {
+                                 //inside
+      if (pos.x () >= leftx && pos.x () <= rightx) {
+        if (pos.y () > ymax)
+          ymax = pos.y ();
+        if (pos.y () < ymin)
+          ymin = pos.y ();
+      }
+      vec = outline->step (stepindex);
+      vec.rotate (rotation);
+      pos += vec;                //move to next
+    }
+  }
+}
+
+
+/**********************************************************************
+ * find_cblob_vlimits
+ *
+ * Scan the outlines of the cblob to locate the y min and max
+ * between the given x limits.
+ **********************************************************************/
+
+void find_cblob_vlimits(               //get y limits
+                        C_BLOB *blob,  //blob to search
+                        float leftx,   //x limits
+                        float rightx,
+                        float &ymin,   //output y limits
+                        float &ymax) {
+  INT16 stepindex;               //current point
+  ICOORD pos;                    //current coords
+  ICOORD vec;                    //rotated step
+  C_OUTLINE *outline;            //current outline
+                                 //outlines
+  C_OUTLINE_IT out_it = blob->out_list ();
+
+  ymin = (float) MAX_INT32;
+  ymax = (float) -MAX_INT32;
+  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
+    outline = out_it.data ();
+    pos = outline->start_pos (); //get coords
+    for (stepindex = 0; stepindex < outline->pathlength (); stepindex++) {
+                                 //inside
+      if (pos.x () >= leftx && pos.x () <= rightx) {
+        if (pos.y () > ymax)
+          ymax = pos.y ();
+        if (pos.y () < ymin)
+          ymin = pos.y ();
+      }
+      vec = outline->step (stepindex);
+      pos += vec;                //move to next
+    }
+  }
+}
+
+
+/**********************************************************************
+ * find_cblob_hlimits
+ *
+ * Scan the outlines of the cblob to locate the x min and max
+ * between the given y limits.
+ **********************************************************************/
+
+void find_cblob_hlimits(                //get x limits
+                        C_BLOB *blob,   //blob to search
+                        float bottomy,  //y limits
+                        float topy,
+                        float &xmin,    //output x limits
+                        float &xmax) {
+  INT16 stepindex;               //current point
+  ICOORD pos;                    //current coords
+  ICOORD vec;                    //rotated step
+  C_OUTLINE *outline;            //current outline
+                                 //outlines
+  C_OUTLINE_IT out_it = blob->out_list ();
+
+  xmin = (float) MAX_INT32;
+  xmax = (float) -MAX_INT32;
+  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
+    outline = out_it.data ();
+    pos = outline->start_pos (); //get coords
+    for (stepindex = 0; stepindex < outline->pathlength (); stepindex++) {
+                                 //inside
+      if (pos.y () >= bottomy && pos.y () <= topy) {
+        if (pos.x () > xmax)
+          xmax = pos.x ();
+        if (pos.x () < xmin)
+          xmin = pos.x ();
+      }
+      vec = outline->step (stepindex);
+      pos += vec;                //move to next
+    }
+  }
+}
+
+
+/**********************************************************************
+ * rotate_blob
+ *
+ * Poly copy the blob and rotate the copy by the given vector.
+ **********************************************************************/
+
+PBLOB *rotate_blob(                 //get y limits
+                   PBLOB *blob,     //blob to search
+                   FCOORD rotation  //vector to rotate by
+                  ) {
+  PBLOB *copy;                   //copy of blob
+  POLYPT *polypt;                //current point
+  OUTLINE_IT out_it;
+  POLYPT_IT poly_it;             //outline pts
+
+  copy = new PBLOB;
+  *copy = *blob;                 //deep copy
+  out_it.set_to_list (copy->out_list ());
+  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
+                                 //get points
+    poly_it.set_to_list (out_it.data ()->polypts ());
+    for (poly_it.mark_cycle_pt (); !poly_it.cycled_list ();
+    poly_it.forward ()) {
+      polypt = poly_it.data ();
+                                 //rotate it
+      polypt->pos.rotate (rotation);
+      polypt->vec.rotate (rotation);
+    }
+    out_it.data ()->compute_bb ();
+  }
+  return copy;
+}
+
+
+/**********************************************************************
+ * rotate_cblob
+ *
+ * Poly copy the blob and rotate the copy by the given vector.
+ **********************************************************************/
+
+PBLOB *rotate_cblob(                 //rotate it
+                    C_BLOB *blob,    //blob to search
+                    float xheight,   //for poly approx
+                    FCOORD rotation  //for landscape
+                   ) {
+  PBLOB *copy;                   //copy of blob
+  POLYPT *polypt;                //current point
+  OUTLINE_IT out_it;
+  POLYPT_IT poly_it;             //outline pts
+
+  copy = new PBLOB (blob, xheight);
+  out_it.set_to_list (copy->out_list ());
+  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
+                                 //get points
+    poly_it.set_to_list (out_it.data ()->polypts ());
+    for (poly_it.mark_cycle_pt (); !poly_it.cycled_list ();
+    poly_it.forward ()) {
+      polypt = poly_it.data ();
+                                 //rotate it
+      polypt->pos.rotate (rotation);
+      polypt->vec.rotate (rotation);
+    }
+    out_it.data ()->compute_bb ();
+  }
+  return copy;
+}
+
+
+/**********************************************************************
+ * crotate_cblob
+ *
+ * Rotate the copy by the given vector and return a C_BLOB.
+ **********************************************************************/
+
+C_BLOB *crotate_cblob(                 //rotate it
+                      C_BLOB *blob,    //blob to search
+                      FCOORD rotation  //for landscape
+                     ) {
+  C_OUTLINE_LIST out_list;       //output outlines
+                                 //input outlines
+  C_OUTLINE_IT in_it = blob->out_list ();
+                                 //output outlines
+  C_OUTLINE_IT out_it = &out_list;
+
+  for (in_it.mark_cycle_pt (); !in_it.cycled_list (); in_it.forward ()) {
+    out_it.add_after_then_move (new C_OUTLINE (in_it.data (), rotation));
+  }
+  return new C_BLOB (&out_list);
+}
+
+
+/**********************************************************************
+ * box_next
+ *
+ * Compute the bounding box of this blob with merging of x overlaps
+ * but no pre-chopping.
+ * Then move the iterator on to the start of the next blob.
+ **********************************************************************/
+
+BOX box_next(                 //get bounding box
+             BLOBNBOX_IT *it  //iterator to blobds
+            ) {
+  BLOBNBOX *blob;                //current blob
+  BOX result;                    //total box
+
+  blob = it->data ();
+  result = blob->bounding_box ();
+  do {
+    it->forward ();
+    blob = it->data ();
+    if (blob->blob () == NULL && blob->cblob () == NULL)
+                                 //was pre-chopped
+      result += blob->bounding_box ();
+  }
+                                 //until next real blob
+  while (blob->blob () == NULL && blob->cblob () == NULL || blob->joined_to_prev ());
+  return result;
+}
+
+
+/**********************************************************************
+ * box_next_pre_chopped
+ *
+ * Compute the bounding box of this blob with merging of x overlaps
+ * but WITH pre-chopping.
+ * Then move the iterator on to the start of the next pre-chopped blob.
+ **********************************************************************/
+
+BOX box_next_pre_chopped(                 //get bounding box
+                         BLOBNBOX_IT *it  //iterator to blobds
+                        ) {
+  BLOBNBOX *blob;                //current blob
+  BOX result;                    //total box
+
+  blob = it->data ();
+  result = blob->bounding_box ();
+  do {
+    it->forward ();
+    blob = it->data ();
+  }
+                                 //until next real blob
+  while (blob->joined_to_prev ());
+  return result;
+}
+
+
+/**********************************************************************
+ * TO_ROW::TO_ROW
+ *
+ * Constructor to make a row from a blob.
+ **********************************************************************/
+
+TO_ROW::TO_ROW (                 //constructor
+BLOBNBOX * blob,                 //first blob
+float top,                       //corrected top
+float bottom,                    //of row
+float row_size                   //ideal
+):y_min (bottom), y_max (top), initial_y_min (bottom) {
+  float diff;                    //in size
+  BLOBNBOX_IT it = &blobs;       //list of blobs
+
+  it.add_to_end (blob);
+  diff = top - bottom - row_size;
+  if (diff > 0) {
+    y_max -= diff / 2;
+    y_min += diff / 2;
+  }
+                                 //very small object
+  else if ((top - bottom) * 3 < row_size) {
+    diff = row_size / 3 + bottom - top;
+    y_max += diff / 2;
+    y_min -= diff / 2;
+  }
+}
+
+
+/**********************************************************************
+ * TO_ROW:add_blob
+ *
+ * Add the blob to the end of the row.
+ **********************************************************************/
+
+void TO_ROW::add_blob(                 //constructor
+                      BLOBNBOX *blob,  //first blob
+                      float top,       //corrected top
+                      float bottom,    //of row
+                      float row_size   //ideal
+                     ) {
+  float allowed;                 //allowed expansion
+  float available;               //expansion
+  BLOBNBOX_IT it = &blobs;       //list of blobs
+
+  it.add_to_end (blob);
+  allowed = row_size + y_min - y_max;
+  if (allowed > 0) {
+    available = top > y_max ? top - y_max : 0;
+    if (bottom < y_min)
+                                 //total available
+        available += y_min - bottom;
+    if (available > 0) {
+      available += available;    //do it gradually
+      if (available < allowed)
+        available = allowed;
+      if (bottom < y_min)
+        y_min -= (y_min - bottom) * allowed / available;
+      if (top > y_max)
+        y_max += (top - y_max) * allowed / available;
+    }
+  }
+}
+
+
+/**********************************************************************
+ * TO_ROW:insert_blob
+ *
+ * Add the blob to the row in the correct position.
+ **********************************************************************/
+
+void TO_ROW::insert_blob(                //constructor
+                         BLOBNBOX *blob  //first blob
+                        ) {
+  BLOBNBOX_IT it = &blobs;       //list of blobs
+
+  if (it.empty ())
+    it.add_before_then_move (blob);
+  else {
+    it.mark_cycle_pt ();
+    while (!it.cycled_list ()
+      && it.data ()->bounding_box ().left () <=
+      blob->bounding_box ().left ())
+      it.forward ();
+    if (it.cycled_list ())
+      it.add_to_end (blob);
+    else
+      it.add_before_stay_put (blob);
+  }
+}
+
+
+/**********************************************************************
+ * TO_ROW::compute_vertical_projection
+ *
+ * Compute the vertical projection of a TO_ROW from its blobs.
+ **********************************************************************/
+
+void TO_ROW::compute_vertical_projection() {  //project whole row
+  BOX row_box;                   //bound of row
+  BLOBNBOX *blob;                //current blob
+  BOX blob_box;                  //bounding box
+  BLOBNBOX_IT blob_it = blob_list ();
+
+  if (blob_it.empty ())
+    return;
+  row_box = blob_it.data ()->bounding_box ();
+  for (blob_it.mark_cycle_pt (); !blob_it.cycled_list (); blob_it.forward ())
+    row_box += blob_it.data ()->bounding_box ();
+
+  projection.set_range (row_box.left () - PROJECTION_MARGIN,
+    row_box.right () + PROJECTION_MARGIN);
+  projection_left = row_box.left () - PROJECTION_MARGIN;
+  projection_right = row_box.right () + PROJECTION_MARGIN;
+  for (blob_it.mark_cycle_pt (); !blob_it.cycled_list (); blob_it.forward ()) {
+    blob = blob_it.data ();
+    if (blob->blob () != NULL)
+      vertical_blob_projection (blob->blob (), &projection);
+    else if (blob->cblob () != NULL)
+      vertical_cblob_projection (blob->cblob (), &projection);
+  }
+}
+
+
+/**********************************************************************
+ * vertical_blob_projection
+ *
+ * Compute the vertical projection of a blob from its outlines
+ * and add to the given STATS.
+ **********************************************************************/
+
+void vertical_blob_projection(              //project outlines
+                              PBLOB *blob,  //blob to project
+                              STATS *stats  //output
+                             ) {
+                                 //outlines of blob
+  OUTLINE_IT out_it = blob->out_list ();
+
+  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
+    vertical_outline_projection (out_it.data (), stats);
+  }
+}
+
+
+/**********************************************************************
+ * vertical_outline_projection
+ *
+ * Compute the vertical projection of a outline from its outlines
+ * and add to the given STATS.
+ **********************************************************************/
+
+void vertical_outline_projection(                   //project outlines
+                                 OUTLINE *outline,  //outline to project
+                                 STATS *stats       //output
+                                ) {
+  POLYPT *polypt;                //current point
+  INT32 xcoord;                  //current pixel coord
+  float end_x;                   //end of vec
+  POLYPT_IT poly_it = outline->polypts ();
+  OUTLINE_IT out_it = outline->child ();
+  float ymean;                   //amount to add
+  float width;                   //amount of x
+
+  for (poly_it.mark_cycle_pt (); !poly_it.cycled_list (); poly_it.forward ()) {
+    polypt = poly_it.data ();
+    end_x = polypt->pos.x () + polypt->vec.x ();
+    if (polypt->vec.x () > 0) {
+      for (xcoord = (INT32) floor (polypt->pos.x ());
+      xcoord < end_x; xcoord++) {
+        if (polypt->pos.x () < xcoord) {
+          width = (float) xcoord;
+          ymean =
+            polypt->vec.y () * (xcoord -
+            polypt->pos.x ()) / polypt->vec.x () +
+            polypt->pos.y ();
+        }
+        else {
+          width = polypt->pos.x ();
+          ymean = polypt->pos.y ();
+        }
+        if (end_x > xcoord + 1) {
+          width -= xcoord + 1;
+          ymean +=
+            polypt->vec.y () * (xcoord + 1 -
+            polypt->pos.x ()) / polypt->vec.x () +
+            polypt->pos.y ();
+        }
+        else {
+          width -= end_x;
+          ymean += polypt->pos.y () + polypt->vec.y ();
+        }
+        ymean = ymean * width / 2;
+        stats->add (xcoord, (INT32) floor (ymean + 0.5));
+      }
+    }
+    else if (polypt->vec.x () < 0) {
+      for (xcoord = (INT32) floor (end_x);
+      xcoord < polypt->pos.x (); xcoord++) {
+        if (polypt->pos.x () > xcoord + 1) {
+          width = xcoord + 1.0f;
+          ymean =
+            polypt->vec.y () * (xcoord + 1 -
+            polypt->pos.x ()) / polypt->vec.x () +
+            polypt->pos.y ();
+        }
+        else {
+          width = polypt->pos.x ();
+          ymean = polypt->pos.y ();
+        }
+        if (end_x < xcoord) {
+          width -= xcoord;
+          ymean +=
+            polypt->vec.y () * (xcoord -
+            polypt->pos.x ()) / polypt->vec.x () +
+            polypt->pos.y ();
+        }
+        else {
+          width -= end_x;
+          ymean += polypt->pos.y () + polypt->vec.y ();
+        }
+        ymean = ymean * width / 2;
+        stats->add (xcoord, (INT32) floor (ymean + 0.5));
+      }
+    }
+  }
+
+  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
+    vertical_outline_projection (out_it.data (), stats);
+  }
+}
+
+
+/**********************************************************************
+ * vertical_cblob_projection
+ *
+ * Compute the vertical projection of a cblob from its outlines
+ * and add to the given STATS.
+ **********************************************************************/
+
+void vertical_cblob_projection(               //project outlines
+                               C_BLOB *blob,  //blob to project
+                               STATS *stats   //output
+                              ) {
+                                 //outlines of blob
+  C_OUTLINE_IT out_it = blob->out_list ();
+
+  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
+    vertical_coutline_projection (out_it.data (), stats);
+  }
+}
+
+
+/**********************************************************************
+ * vertical_coutline_projection
+ *
+ * Compute the vertical projection of a outline from its outlines
+ * and add to the given STATS.
+ **********************************************************************/
+
+void vertical_coutline_projection(                     //project outlines
+                                  C_OUTLINE *outline,  //outline to project
+                                  STATS *stats         //output
+                                 ) {
+  ICOORD pos;                    //current point
+  ICOORD step;                   //edge step
+  INT32 length;                  //of outline
+  INT16 stepindex;               //current step
+  C_OUTLINE_IT out_it = outline->child ();
+
+  pos = outline->start_pos ();
+  length = outline->pathlength ();
+  for (stepindex = 0; stepindex < length; stepindex++) {
+    step = outline->step (stepindex);
+    if (step.x () > 0) {
+      if (pitsync_projection_fix)
+        stats->add (pos.x (), -pos.y ());
+      else
+        stats->add (pos.x (), pos.y ());
+    }
+    else if (step.x () < 0) {
+      if (pitsync_projection_fix)
+        stats->add (pos.x () - 1, pos.y ());
+      else
+        stats->add (pos.x () - 1, -pos.y ());
+    }
+    pos += step;
+  }
+
+  for (out_it.mark_cycle_pt (); !out_it.cycled_list (); out_it.forward ()) {
+    vertical_coutline_projection (out_it.data (), stats);
+  }
+}
+
+
+/**********************************************************************
+ * TO_BLOCK::TO_BLOCK
+ *
+ * Constructor to make a TO_BLOCK from a real block.
+ **********************************************************************/
+
+TO_BLOCK::TO_BLOCK(                  //make a block
+                   BLOCK *src_block  //real block
+                  ) {
+  block = src_block;
+}
+
+static void clear_blobnboxes(BLOBNBOX_LIST* boxes) {
+  BLOBNBOX_IT it = boxes;
+  // A BLOBNBOX generally doesn't own its blobs, so if they do, you
+  // have to delete them explicitly.
+  for (it.mark_cycle_pt(); !it.cycled_list(); it.forward()) {
+    BLOBNBOX* box = it.data();
+    if (box->blob() != NULL)
+      delete box->blob();
+    if (box->cblob() != NULL)
+      delete box->cblob();
+  }
+}
+
+TO_BLOCK::~TO_BLOCK() {
+  // Any residual BLOBNBOXes at this stage own their blobs, so delete them.
+  clear_blobnboxes(&blobs);
+  clear_blobnboxes(&underlines);
+  clear_blobnboxes(&noise_blobs);
+  clear_blobnboxes(&small_blobs);
+  clear_blobnboxes(&large_blobs);
+}
+

ccstruct/blobbox.h

@@ -0,0 +1,381 @@
+/**********************************************************************
+ * File:        blobbox.h  (Formerly blobnbox.h)
+ * Description: Code for the textord blob class.
+ * Author:					Ray Smith
+ * Created:					Thu Jul 30 09:08:51 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           BLOBBOX_H
+#define           BLOBBOX_H
+
+#include          "varable.h"
+#include          "clst.h"
+#include          "elst2.h"
+#include          "werd.h"
+#include          "ocrblock.h"
+#include          "statistc.h"
+
+extern double_VAR_H (textord_error_weight, 3,
+"Weighting for error in believability");
+
+enum PITCH_TYPE
+{
+  PITCH_DUNNO,                   //insufficient data
+  PITCH_DEF_FIXED,               //definitely fixed
+  PITCH_MAYBE_FIXED,             //could be
+  PITCH_DEF_PROP,
+  PITCH_MAYBE_PROP,
+  PITCH_CORR_FIXED,
+  PITCH_CORR_PROP
+};
+
+class BLOBNBOX;
+ELISTIZEH (BLOBNBOX)
+class BLOBNBOX:public ELIST_LINK
+{
+  public:
+    BLOBNBOX() {  //empty
+      blob_ptr = NULL;
+      cblob_ptr = NULL;
+      joined = FALSE;
+      reduced = FALSE;
+      area = 0;
+    }
+    BLOBNBOX(  //constructor
+             PBLOB *srcblob) {
+      blob_ptr = srcblob;
+      cblob_ptr = NULL;
+      box = srcblob->bounding_box ();
+      joined = FALSE;
+      reduced = FALSE;
+      area = (int) srcblob->area ();
+    }
+    BLOBNBOX(  //constructor
+             C_BLOB *srcblob) {
+      blob_ptr = NULL;
+      cblob_ptr = srcblob;
+      box = srcblob->bounding_box ();
+      joined = FALSE;
+      reduced = FALSE;
+      area = (int) srcblob->area ();
+    }
+
+                                 //get bounding box
+    const BOX &bounding_box() const {
+      return box;
+    }
+                                 //get bounding box
+    const BOX &reduced_box() const {
+      return red_box;
+    }
+    void set_reduced_box(  //set other box
+                         BOX new_box) {
+      red_box = new_box;
+      reduced = TRUE;
+    }
+    INT32 enclosed_area() const {  //get area
+      return area;
+    }
+
+    void rotate_box(  //just box
+                    FCOORD vec) {
+      box.rotate (vec);
+    }
+
+    BOOL8 joined_to_prev() const {  //access function
+      return joined != 0;
+    }
+    BOOL8 red_box_set() const {  //access function
+      return reduced != 0;
+    }
+    void merge(  //merge with next
+               BLOBNBOX *nextblob);
+    void chop(                        //fake chop blob
+              BLOBNBOX_IT *start_it,  //location of this
+              BLOBNBOX_IT *blob_it,   //iterator
+              FCOORD rotation,        //for landscape
+              float xheight);         //line height
+
+    PBLOB *blob() {  //access function
+      return blob_ptr;
+    }
+    C_BLOB *cblob() {  //access function
+      return cblob_ptr;
+    }
+
+#ifndef GRAPHICS_DISABLED
+    void plot(                        //draw one
+              WINDOW window,          //window to draw in
+              COLOUR blob_colour,     //for outer bits
+              COLOUR child_colour) {  //for holes
+      if (blob_ptr != NULL)
+        blob_ptr->plot (window, blob_colour, child_colour);
+      if (cblob_ptr != NULL)
+        cblob_ptr->plot (window, blob_colour, child_colour);
+    }
+#endif
+
+    NEWDELETE2 (BLOBNBOX) private:
+    int area:30;                 //enclosed area
+    int joined:1;                //joined to prev
+    int reduced:1;               //reduced box set
+    BOX box;                     //bounding box
+    BOX red_box;                 //bounding box
+    PBLOB *blob_ptr;             //poly blob
+    C_BLOB *cblob_ptr;           //edgestep blob
+};
+
+class TO_ROW:public ELIST2_LINK
+{
+  public:
+    TO_ROW() {
+    }                            //empty
+    TO_ROW(                 //constructor
+           BLOBNBOX *blob,  //from first blob
+           float top,       //of row //target height
+           float bottom,
+           float row_size);
+
+    float max_y() const {  //access function
+      return y_max;
+    }
+    float min_y() const {
+      return y_min;
+    }
+    float mean_y() const {
+      return (y_min + y_max) / 2.0f;
+    }
+    float initial_min_y() const {
+      return initial_y_min;
+    }
+    float line_m() const {  //access to line fit
+      return m;
+    }
+    float line_c() const {
+      return c;
+    }
+    float line_error() const {
+      return error;
+    }
+    float parallel_c() const {
+      return para_c;
+    }
+    float parallel_error() const {
+      return para_error;
+    }
+    float believability() const {  //baseline goodness
+      return credibility;
+    }
+    float intercept() const {  //real parallel_c
+      return y_origin;
+    }
+    void add_blob(                 //put in row
+                  BLOBNBOX *blob,  //blob to add
+                  float top,       //of row //target height
+                  float bottom,
+                  float row_size);
+    void insert_blob(  //put in row in order
+                     BLOBNBOX *blob);
+
+    BLOBNBOX_LIST *blob_list() {  //get list
+      return &blobs;
+    }
+
+    void set_line(              //set line spec
+                  float new_m,  //line to set
+                  float new_c,
+                  float new_error) {
+      m = new_m;
+      c = new_c;
+      error = new_error;
+    }
+    void set_parallel_line(                 //set fixed gradient line
+                           float gradient,  //page gradient
+                           float new_c,
+                           float new_error) {
+      para_c = new_c;
+      para_error = new_error;
+      credibility =
+        (float) (blobs.length () - textord_error_weight * new_error);
+      y_origin = (float) (new_c / sqrt (1 + gradient * gradient));
+      //real intercept
+    }
+    void set_limits(                  //set min,max
+                    float new_min,    //bottom and
+                    float new_max) {  //top of row
+      y_min = new_min;
+      y_max = new_max;
+    }
+    void compute_vertical_projection();
+    //get projection
+
+                                 //true when dead
+    NEWDELETE2 (TO_ROW) BOOL8 merged;
+    BOOL8 all_caps;              //had no ascenders
+    BOOL8 used_dm_model;         //in guessing pitch
+    INT16 projection_left;       //start of projection
+    INT16 projection_right;      //start of projection
+    PITCH_TYPE pitch_decision;   //how strong is decision
+    float fixed_pitch;           //pitch or 0
+    float fp_space;              //sp if fixed pitch
+    float fp_nonsp;              //nonsp if fixed pitch
+    float pr_space;              //sp if prop
+    float pr_nonsp;              //non sp if prop
+    float spacing;               //to "next" row
+    float xheight;               //of line
+    float ascrise;               //ascenders
+    float descdrop;              //descenders
+    INT32 min_space;             //min size for real space
+    INT32 max_nonspace;          //max size of non-space
+    INT32 space_threshold;       //space vs nonspace
+    float kern_size;             //average non-space
+    float space_size;            //average space
+    WERD_LIST rep_words;         //repeated chars
+    ICOORDELT_LIST char_cells;   //fixed pitch cells
+    QSPLINE baseline;            //curved baseline
+    STATS projection;            //vertical projection
+
+  private:
+    BLOBNBOX_LIST blobs;         //blobs in row
+    float y_min;                 //coords
+    float y_max;
+    float initial_y_min;
+    float m, c;                  //line spec
+    float error;                 //line error
+    float para_c;                //constrained fit
+    float para_error;
+    float y_origin;              //rotated para_c;
+    float credibility;           //baseline believability
+};
+
+ELIST2IZEH (TO_ROW)
+class TO_BLOCK:public ELIST_LINK
+{
+  public:
+    TO_BLOCK() {
+    }                            //empty
+    TO_BLOCK(                    //constructor
+             BLOCK *src_block);  //real block
+    ~TO_BLOCK();
+
+    TO_ROW_LIST *get_rows() {  //access function
+      return &row_list;
+    }
+
+    void print_rows() {  //debug info
+      TO_ROW_IT row_it = &row_list;
+      TO_ROW *row;
+
+      for (row_it.mark_cycle_pt (); !row_it.cycled_list ();
+      row_it.forward ()) {
+        row = row_it.data ();
+        printf ("Row range (%g,%g), para_c=%g, blobcount=" INT32FORMAT
+          "\n", row->min_y (), row->max_y (), row->parallel_c (),
+          row->blob_list ()->length ());
+      }
+    }
+
+    BLOBNBOX_LIST blobs;         //medium size
+    BLOBNBOX_LIST underlines;    //underline blobs
+    BLOBNBOX_LIST noise_blobs;   //very small
+    BLOBNBOX_LIST small_blobs;   //fairly small
+    BLOBNBOX_LIST large_blobs;   //big blobs
+    BLOCK *block;                //real block
+    PITCH_TYPE pitch_decision;   //how strong is decision
+    float line_spacing;          //estimate
+    float line_size;             //estimate
+    float max_blob_size;         //line assignment limit
+    float baseline_offset;       //phase shift
+    float xheight;               //median blob size
+    float fixed_pitch;           //pitch or 0
+    float kern_size;             //average non-space
+    float space_size;            //average space
+    INT32 min_space;             //min definite space
+    INT32 max_nonspace;          //max definite
+    float fp_space;              //sp if fixed pitch
+    float fp_nonsp;              //nonsp if fixed pitch
+    float pr_space;              //sp if prop
+    float pr_nonsp;              //non sp if prop
+    TO_ROW *key_row;             //starting row
+
+    NEWDELETE2 (TO_BLOCK) private:
+    TO_ROW_LIST row_list;        //temporary rows
+};
+
+ELISTIZEH (TO_BLOCK)
+extern double_VAR_H (textord_error_weight, 3,
+"Weighting for error in believability");
+void find_blob_limits(                  //get y limits
+                      PBLOB *blob,      //blob to search
+                      float leftx,      //x limits
+                      float rightx,
+                      FCOORD rotation,  //for landscape
+                      float &ymin,      //output y limits
+                      float &ymax);
+void find_cblob_limits(                  //get y limits
+                       C_BLOB *blob,     //blob to search
+                       float leftx,      //x limits
+                       float rightx,
+                       FCOORD rotation,  //for landscape
+                       float &ymin,      //output y limits
+                       float &ymax);
+void find_cblob_vlimits(               //get y limits
+                        C_BLOB *blob,  //blob to search
+                        float leftx,   //x limits
+                        float rightx,
+                        float &ymin,   //output y limits
+                        float &ymax);
+void find_cblob_hlimits(                //get x limits
+                        C_BLOB *blob,   //blob to search
+                        float bottomy,  //y limits
+                        float topy,
+                        float &xmin,    //output x limits
+                        float &xymax);
+PBLOB *rotate_blob(                 //get y limits
+                   PBLOB *blob,     //blob to search
+                   FCOORD rotation  //vector to rotate by
+                  );
+PBLOB *rotate_cblob(                 //rotate it
+                    C_BLOB *blob,    //blob to search
+                    float xheight,   //for poly approx
+                    FCOORD rotation  //for landscape
+                   );
+C_BLOB *crotate_cblob(                 //rotate it
+                      C_BLOB *blob,    //blob to search
+                      FCOORD rotation  //for landscape
+                     );
+BOX box_next(                 //get bounding box
+             BLOBNBOX_IT *it  //iterator to blobds
+            );
+BOX box_next_pre_chopped(                 //get bounding box
+                         BLOBNBOX_IT *it  //iterator to blobds
+                        );
+void vertical_blob_projection(              //project outlines
+                              PBLOB *blob,  //blob to project
+                              STATS *stats  //output
+                             );
+                                 //project outlines
+void vertical_outline_projection(OUTLINE *outline,  //outline to project
+                                 STATS *stats       //output
+                                );
+void vertical_cblob_projection(               //project outlines
+                               C_BLOB *blob,  //blob to project
+                               STATS *stats   //output
+                              );
+void vertical_coutline_projection(                     //project outlines
+                                  C_OUTLINE *outline,  //outline to project
+                                  STATS *stats         //output
+                                 );
+#endif

ccstruct/blobs.cpp

@@ -0,0 +1,247 @@
+/* -*-C-*-
+ ********************************************************************************
+ *
+ * File:        blobs.c  (Formerly blobs.c)
+ * Description:  Blob definition
+ * Author:       Mark Seaman, OCR Technology
+ * Created:      Fri Oct 27 15:39:52 1989
+ * Modified:     Thu Mar 28 15:33:26 1991 (Mark Seaman) marks@hpgrlt
+ * Language:     C
+ * Package:      N/A
+ * Status:       Experimental (Do Not Distribute)
+ *
+ * (c) Copyright 1989, Hewlett-Packard Company.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ *********************************************************************************/
+
+/*----------------------------------------------------------------------
+              I n c l u d e s
+----------------------------------------------------------------------*/
+#include "mfcpch.h"
+#include "blobs.h"
+#include "cutil.h"
+#include "emalloc.h"
+#include "structures.h"
+
+/*----------------------------------------------------------------------
+              F u n c t i o n s
+----------------------------------------------------------------------*/
+/**********************************************************************
+ * blob_origin
+ *
+ * Compute the origin of a compound blob, define to be the centre
+ * of the bounding box.
+ **********************************************************************/
+void blob_origin(TBLOB *blob,       /*blob to compute on */
+                 TPOINT *origin) {  /*return value */
+  TPOINT topleft;                /*bounding box */
+  TPOINT botright;
+
+                                 /*find bounding box */
+  blob_bounding_box(blob, &topleft, &botright); 
+                                 /*centre of box */
+  origin->x = (topleft.x + botright.x) / 2;
+  origin->y = (topleft.y + botright.y) / 2;
+}
+
+
+/**********************************************************************
+ * blob_bounding_box
+ *
+ * Compute the bounding_box of a compound blob, define to be the
+ * max coordinate value of the bounding boxes of all the top-level
+ * outlines in the box.
+ **********************************************************************/
+void blob_bounding_box(TBLOB *blob,               /*blob to compute on */
+                       register TPOINT *topleft,  /*bounding box */
+                       register TPOINT *botright) {
+  register TESSLINE *outline;    /*current outline */
+
+  if (blob == NULL || blob->outlines == NULL) {
+    topleft->x = topleft->y = 0;
+    *botright = *topleft;        /*default value */
+  }
+  else {
+    outline = blob->outlines;
+    *topleft = outline->topleft;
+    *botright = outline->botright;
+    for (outline = outline->next; outline != NULL; outline = outline->next) {
+      if (outline->topleft.x < topleft->x)
+                                 /*find extremes */
+        topleft->x = outline->topleft.x;
+      if (outline->botright.x > botright->x)
+                                 /*find extremes */
+        botright->x = outline->botright.x;
+      if (outline->topleft.y > topleft->y)
+                                 /*find extremes */
+        topleft->y = outline->topleft.y;
+      if (outline->botright.y < botright->y)
+                                 /*find extremes */
+        botright->y = outline->botright.y;
+    }
+  }
+}
+
+
+/**********************************************************************
+ * blobs_bounding_box
+ *
+ * Return the smallest extreme point that contain this word.
+ **********************************************************************/
+void blobs_bounding_box(TBLOB *blobs, TPOINT *topleft, TPOINT *botright) { 
+  TPOINT tl;
+  TPOINT br;
+  TBLOB *blob;
+  /* Start with first blob */
+  blob_bounding_box(blobs, topleft, botright); 
+
+  iterate_blobs(blob, blobs) { 
+    blob_bounding_box(blob, &tl, &br); 
+
+    if (tl.x < topleft->x)
+      topleft->x = tl.x;
+    if (tl.y > topleft->y)
+      topleft->y = tl.y;
+    if (br.x > botright->x)
+      botright->x = br.x;
+    if (br.y < botright->y)
+      botright->y = br.y;
+  }
+}
+
+
+/**********************************************************************
+ * blobs_origin
+ *
+ * Compute the origin of a compound blob, define to be the centre
+ * of the bounding box.
+ **********************************************************************/
+void blobs_origin(TBLOB *blobs,      /*blob to compute on */
+                  TPOINT *origin) {  /*return value */
+  TPOINT topleft;                /*bounding box */
+  TPOINT botright;
+
+                                 /*find bounding box */
+  blobs_bounding_box(blobs, &topleft, &botright); 
+                                 /*center of box */
+  origin->x = (topleft.x + botright.x) / 2;
+  origin->y = (topleft.y + botright.y) / 2;
+}
+
+
+/**********************************************************************
+ * blobs_widths
+ *
+ * Compute the widths of a list of blobs. Return an array of the widths
+ * and gaps.
+ **********************************************************************/
+WIDTH_RECORD *blobs_widths(TBLOB *blobs) {  /*blob to compute on */
+  WIDTH_RECORD *width_record;
+  TPOINT topleft;                /*bounding box */
+  TPOINT botright;
+  TBLOB *blob;                   /*blob to compute on */
+  int i = 0;
+  int blob_end;
+  int num_blobs = count_blobs (blobs);
+
+  /* Get memory */
+  width_record = (WIDTH_RECORD *) memalloc (sizeof (int) * num_blobs * 2);
+  width_record->num_chars = num_blobs;
+
+  blob_bounding_box(blobs, &topleft, &botright); 
+  width_record->widths[i++] = botright.x - topleft.x;
+  /* First width */
+  blob_end = botright.x;
+
+  iterate_blobs (blob, blobs->next) {
+    blob_bounding_box(blob, &topleft, &botright); 
+    width_record->widths[i++] = topleft.x - blob_end;
+    width_record->widths[i++] = botright.x - topleft.x;
+    blob_end = botright.x;
+  }
+  return (width_record);
+}
+
+
+/**********************************************************************
+ * count_blobs
+ *
+ * Return a count of the number of blobs attached to this one.
+ **********************************************************************/
+int count_blobs(TBLOB *blobs) { 
+  TBLOB *b;
+  int x = 0;
+
+  iterate_blobs (b, blobs) x++;
+  return (x);
+}
+
+
+/**********************************************************************
+ * delete_word
+ *
+ * Reclaim the memory taken by this word structure and all of its
+ * lower level structures.
+ **********************************************************************/
+void delete_word(TWERD *word) { 
+  TBLOB *blob;
+  TBLOB *nextblob;
+  TESSLINE *outline;
+  TESSLINE *nextoutline;
+  TESSLINE *child;
+  TESSLINE *nextchild;
+
+  for (blob = word->blobs; blob; blob = nextblob) {
+    nextblob = blob->next;
+
+    for (outline = blob->outlines; outline; outline = nextoutline) {
+      nextoutline = outline->next;
+
+      delete_edgepts (outline->loop);
+
+      for (child = outline->child; child; child = nextchild) {
+        nextchild = child->next;
+
+        delete_edgepts (child->loop);
+
+        oldoutline(child); 
+      }
+      oldoutline(outline); 
+    }
+    oldblob(blob); 
+  }
+  if (word->correct != NULL)
+    strfree (word->correct);     /* Reclaim memory */
+  oldword(word); 
+}
+
+
+/**********************************************************************
+ * delete_edgepts
+ *
+ * Delete a list of EDGEPT structures.
+ **********************************************************************/
+void delete_edgepts(register EDGEPT *edgepts) { 
+  register EDGEPT *this_edge;
+  register EDGEPT *next_edge;
+
+  if (edgepts == NULL)
+    return;
+
+  this_edge = edgepts;
+  do {
+    next_edge = this_edge->next;
+    oldedgept(this_edge); 
+    this_edge = next_edge;
+  }
+  while (this_edge != edgepts);
+}

ccstruct/blobs.h

@@ -0,0 +1,119 @@
+/* -*-C-*-
+ ********************************************************************************
+ *
+ * File:        blobs.h  (Formerly blobs.h)
+ * Description:  Blob definition
+ * Author:       Mark Seaman, OCR Technology
+ * Created:      Fri Oct 27 15:39:52 1989
+ * Modified:     Thu Mar 28 15:33:38 1991 (Mark Seaman) marks@hpgrlt
+ * Language:     C
+ * Package:      N/A
+ * Status:       Experimental (Do Not Distribute)
+ *
+ * (c) Copyright 1989, Hewlett-Packard Company.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ *********************************************************************************/
+
+#ifndef BLOBS_H
+#define BLOBS_H
+
+/*----------------------------------------------------------------------
+              I n c l u d e s
+----------------------------------------------------------------------*/
+#include                   "vecfuncs.h"
+#include  "tessclas.h"
+
+/*----------------------------------------------------------------------
+              T y p e s
+----------------------------------------------------------------------*/
+typedef struct
+{                                /* Widths of pieces */
+  int num_chars;
+  int widths[1];
+} WIDTH_RECORD;
+
+/*----------------------------------------------------------------------
+              M a c r o s
+----------------------------------------------------------------------*/
+/**********************************************************************
+ * free_widths
+ *
+ * Free the memory taken up by a width array.
+ **********************************************************************/
+#define free_widths(w)  \
+if (w) memfree (w)
+
+/*----------------------------------------------------------------------
+              F u n c t i o n s
+----------------------------------------------------------------------*/
+void blob_origin(TBLOB *blob,      /*blob to compute on */
+                 TPOINT *origin);  /*return value */
+
+                                 /*blob to compute on */
+void blob_bounding_box(TBLOB *blob,
+                       register TPOINT *topleft,  /*bounding box */
+                       register TPOINT *botright);
+
+void blobs_bounding_box(TBLOB *blobs, TPOINT *topleft, TPOINT *botright); 
+
+void blobs_origin(TBLOB *blobs,     /*blob to compute on */
+                  TPOINT *origin);  /*return value */
+
+                                 /*blob to compute on */
+WIDTH_RECORD *blobs_widths(TBLOB *blobs); 
+
+int count_blobs(TBLOB *blobs); 
+
+void delete_word(TWERD *word); 
+
+void delete_edgepts(register EDGEPT *edgepts); 
+
+/*
+#if defined(__STDC__) || defined(__cplusplus)
+# define	_ARGS(s) s
+#else
+# define	_ARGS(s) ()
+#endif*/
+
+/* blobs.c
+void blob_origin
+  _ARGS((BLOB *blob,
+  TPOINT *origin));
+
+void blob_bounding_box
+  _ARGS((BLOB *blob,
+  TPOINT *topleft,
+  TPOINT *botright));
+
+void blobs_bounding_box
+  _ARGS((BLOB *blobs,
+  TPOINT *topleft,
+  TPOINT *botright));
+
+void blobs_origin
+  _ARGS((BLOB *blobs,
+  TPOINT *origin));
+
+WIDTH_RECORD *blobs_widths
+  _ARGS((BLOB *blobs));
+
+int count_blobs
+  _ARGS((BLOB *blobs));
+
+void delete_word
+  _ARGS((TWERD *word));
+
+void delete_edgepts
+  _ARGS((EDGEPT *edgepts));
+#undef _ARGS
+*/
+#endif

ccstruct/blread.cpp

@@ -0,0 +1,537 @@
+/**********************************************************************
+ * File:        blread.cpp  (Formerly pdread.c)
+ * Description: Friend function of BLOCK to read the uscan pd file.
+ * Author:		Ray Smith
+ * Created:		Mon Mar 18 14:39:00 GMT 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          <stdlib.h>
+#ifdef __UNIX__
+#include          <assert.h>
+#endif
+#include          "scanutils.h"
+#include          "fileerr.h"
+#include          "imgtiff.h"
+#include          "pdclass.h"
+#include          "rwpoly.h"
+#include          "blread.h"
+
+#define PD_EXT        ".pd"
+#define VEC_EXT       ".vec"     //accupage file
+#define HPD_EXT       ".bl"      //hand pd file
+                                 //unlv zone file
+#define UNLV_EXT            ".uzn"
+#define BLOCK_EXPANSION   8      //boundary expansion
+#define EXTERN
+
+EXTERN BOOL_EVAR (ignore_weird_blocks, TRUE, "Don't read weird blocks");
+
+static BOX convert_vec_block(                        //make non-rect block
+                             VEC_ENTRY *entries,     //vectors
+                             UINT16 entry_count,     //no of entries
+                             INT32 ysize,            //image size
+                             ICOORDELT_IT *left_it,  //block sides
+                             ICOORDELT_IT *right_it);
+
+/**********************************************************************
+ * BLOCK::read_pd_file
+ *
+ * Read a whole pd file to make a list of blocks, or use the whole page.
+ **********************************************************************/
+
+BOOL8 read_pd_file(                    //print list of sides
+                   STRING name,        //basename of file
+                   INT32 xsize,        //image size
+                   INT32 ysize,        //image size
+                   BLOCK_LIST *blocks  //output list
+                  ) {
+  FILE *pdfp;                    //file pointer
+  BLOCK *block;                  //current block
+  INT32 block_count;             //no of blocks
+  INT32 junk_count;              //no of junks to read
+  INT32 junks[4];                //junk elements
+  INT32 vertex_count;            //boundary vertices
+  INT32 xcoord;                  //current coords
+  INT32 ycoord;
+  INT32 prevx;                   //previous coords
+  INT32 prevy;
+  BLOCK_IT block_it = blocks;    //block iterator
+  ICOORDELT_LIST dummy;          //for constructor
+  ICOORDELT_IT left_it = &dummy; //iterator
+  ICOORDELT_IT right_it = &dummy;//iterator
+
+  if (read_hpd_file (name, xsize, ysize, blocks))
+    return TRUE;                 //succeeded
+  if (read_vec_file (name, xsize, ysize, blocks))
+    return TRUE;                 //succeeded
+  if (read_unlv_file (name, xsize, ysize, blocks))
+    return TRUE;                 //succeeded
+  name += PD_EXT;                //add extension
+  if ((pdfp = fopen (name.string (), "r")) == NULL) {
+                                 //make rect block
+    block = new BLOCK (name.string (), TRUE, 0, 0, 0, 0, xsize, ysize);
+    block_it.add_to_end (block); //on end of list
+    return FALSE;                //didn't read one
+  }
+  else {
+    if (fread (&block_count, sizeof (block_count), 1, pdfp) != 1)
+      READFAILED.error ("read_pd_file", EXIT, "Block count");
+    tprintf ("%d blocks in .pd file.\n", block_count);
+    while (block_count > 0) {
+      if (fread (&junk_count, sizeof (junk_count), 1, pdfp) != 1)
+        READFAILED.error ("read_pd_file", EXIT, "Junk count");
+      if (fread (&vertex_count, sizeof (vertex_count), 1, pdfp) != 1)
+        READFAILED.error ("read_pd_file", EXIT, "Vertex count");
+      block = new BLOCK;         //make a block
+                                 //on end of list
+      block_it.add_to_end (block);
+      left_it.set_to_list (&block->leftside);
+      right_it.set_to_list (&block->rightside);
+
+                                 //read a pair
+      get_pd_vertex (pdfp, xsize, ysize, &block->box, xcoord, ycoord);
+      vertex_count -= 2;         //count read ones
+      prevx = xcoord;
+      do {
+        if (xcoord == prevx) {
+          if (!right_it.empty ()) {
+            if (right_it.data ()->x () <= xcoord + BLOCK_EXPANSION)
+              right_it.data ()->set_y (right_it.data ()->y () +
+                BLOCK_EXPANSION);
+            else
+              right_it.data ()->set_y (right_it.data ()->y () -
+                BLOCK_EXPANSION);
+          }
+          right_it.
+            add_before_then_move (new
+            ICOORDELT (xcoord + BLOCK_EXPANSION,
+            ycoord));
+        }
+        prevx = xcoord;          //remember previous
+        prevy = ycoord;
+        get_pd_vertex (pdfp, xsize, ysize, &block->box, xcoord, ycoord);
+        vertex_count -= 2;       //count read ones
+      }
+      while (ycoord <= prevy);
+      right_it.data ()->set_y (right_it.data ()->y () - BLOCK_EXPANSION);
+
+                                 //start of left
+      left_it.add_to_end (new ICOORDELT (prevx - BLOCK_EXPANSION, prevy - BLOCK_EXPANSION));
+
+      do {
+        prevx = xcoord;          //remember previous
+        get_pd_vertex (pdfp, xsize, ysize, &block->box, xcoord, ycoord);
+        vertex_count -= 2;
+        if (xcoord != prevx && vertex_count > 0) {
+          if (xcoord > prevx)
+            left_it.
+              add_to_end (new
+              ICOORDELT (xcoord - BLOCK_EXPANSION,
+              ycoord + BLOCK_EXPANSION));
+          else
+            left_it.
+              add_to_end (new
+              ICOORDELT (xcoord - BLOCK_EXPANSION,
+              ycoord - BLOCK_EXPANSION));
+        }
+        else if (vertex_count == 0)
+          left_it.add_to_end (new ICOORDELT (prevx - BLOCK_EXPANSION,
+              ycoord + BLOCK_EXPANSION));
+      }
+      while (vertex_count > 0);  //until all read
+
+      while (junk_count > 0) {
+        if (fread (junks, sizeof (INT32), 4, pdfp) != 4)
+          READFAILED.error ("read_pd_file", EXIT, "Junk coords");
+        junk_count--;
+      }
+      block_count--;             //count read blocks
+    }
+  }
+  fclose(pdfp);
+  return TRUE;                   //read one
+}
+
+
+/**********************************************************************
+ * get_pd_vertex
+ *
+ * Read a pair of coords, invert the y and clip to image limits.
+ * Also update the bounding box.
+ *
+ * Read a whole pd file to make a list of blocks, or use the whole page.
+ **********************************************************************/
+
+void get_pd_vertex(                //get new vertex
+                   FILE *pdfp,     //file to read
+                   INT32 xsize,    //image size
+                   INT32 ysize,    //image size
+                   BOX *box,       //bounding box
+                   INT32 &xcoord,  //output coords
+                   INT32 &ycoord) {
+  BOX new_coord;                 //expansion box
+
+                                 //get new coords
+  if (fread (&xcoord, sizeof (xcoord), 1, pdfp) != 1)
+    READFAILED.error ("read_pd_file", EXIT, "Xcoord");
+  if (fread (&ycoord, sizeof (ycoord), 1, pdfp) != 1)
+    READFAILED.error ("read_pd_file", EXIT, "Xcoord");
+  ycoord = ysize - ycoord;       //invert y
+  if (xcoord < BLOCK_EXPANSION)
+    xcoord = BLOCK_EXPANSION;    //clip to limits
+  if (xcoord > xsize - BLOCK_EXPANSION)
+    xcoord = xsize - BLOCK_EXPANSION;
+  if (ycoord < BLOCK_EXPANSION)
+    ycoord = BLOCK_EXPANSION;
+  if (ycoord > ysize - BLOCK_EXPANSION)
+    ycoord = ysize - BLOCK_EXPANSION;
+
+  new_coord =
+    BOX (ICOORD (xcoord - BLOCK_EXPANSION, ycoord - BLOCK_EXPANSION),
+    ICOORD (xcoord + BLOCK_EXPANSION, ycoord + BLOCK_EXPANSION));
+  (*box) += new_coord;
+}
+
+
+/**********************************************************************
+ * BLOCK::read_hpd_file
+ *
+ * Read a whole hpd file to make a list of blocks.
+ * Return FALSE if the .vec fiel cannot be found
+ **********************************************************************/
+
+BOOL8 read_hpd_file(                    //print list of sides
+                    STRING name,        //basename of file
+                    INT32 xsize,        //image size
+                    INT32 ysize,        //image size
+                    BLOCK_LIST *blocks  //output list
+                   ) {
+  FILE *pdfp;                    //file pointer
+  PAGE_BLOCK_LIST *page_blocks;
+  INT32 block_no;                //no of blocks
+  BLOCK_IT block_it = blocks;    //block iterator
+
+  name += HPD_EXT;               //add extension
+  if ((pdfp = fopen (name.string (), "r")) == NULL) {
+    return FALSE;                //can't find it
+  }
+  fclose(pdfp);
+  page_blocks = read_poly_blocks (name.string ());
+  block_no = 0;
+  scan_hpd_blocks (name.string (), page_blocks, block_no, &block_it);
+  tprintf ("Text region count=%d\n", block_no);
+  return TRUE;                   //read one
+}
+
+
+/**********************************************************************
+ * BLOCK::scan_hpd_blocks
+ *
+ * Read a whole hpd file to make a list of blocks.
+ * Return FALSE if the .vec fiel cannot be found
+ **********************************************************************/
+
+void scan_hpd_blocks(                               //print list of sides
+                     const char *name,              //block label
+                     PAGE_BLOCK_LIST *page_blocks,  //head of full pag
+                     INT32 &block_no,               //no of blocks
+                     BLOCK_IT *block_it             //block iterator
+                    ) {
+  BLOCK *block;                  //current block
+                                 //page blocks
+  PAGE_BLOCK_IT pb_it = page_blocks;
+  PAGE_BLOCK *current_block;
+  TEXT_REGION_IT tr_it;
+  TEXT_BLOCK *tb;
+  TEXT_REGION *tr;
+  BOX *block_box;                //from text region
+
+  for (pb_it.mark_cycle_pt (); !pb_it.cycled_list (); pb_it.forward ()) {
+    current_block = pb_it.data ();
+    if (current_block->type () == PB_TEXT) {
+      tb = (TEXT_BLOCK *) current_block;
+      if (!tb->regions ()->empty ()) {
+        tr_it.set_to_list (tb->regions ());
+        for (tr_it.mark_cycle_pt ();
+        !tr_it.cycled_list (); tr_it.forward ()) {
+          block_no++;
+          tr = tr_it.data ();
+          block_box = tr->bounding_box ();
+          block = new BLOCK (name, TRUE, 0, 0,
+            block_box->left (), block_box->bottom (),
+            block_box->right (), block_box->top ());
+          block->hand_block = tr;
+          block->hand_poly = tr;
+          block_it->add_after_then_move (block);
+        }
+      }
+    }
+    else if (current_block->type () == PB_WEIRD
+      && !ignore_weird_blocks
+    && ((WEIRD_BLOCK *) current_block)->id_no () > 0) {
+      block_no++;
+      block_box = current_block->bounding_box ();
+      block = new BLOCK (name, TRUE, 0, 0,
+        block_box->left (), block_box->bottom (),
+        block_box->right (), block_box->top ());
+      block->hand_block = NULL;
+      block->hand_poly = current_block;
+      block_it->add_after_then_move (block);
+    }
+    if (!current_block->child ()->empty ())
+      scan_hpd_blocks (name, current_block->child (), block_no, block_it);
+  }
+}
+
+
+
+
+
+/**********************************************************************
+ * BLOCK::read_vec_file
+ *
+ * Read a whole vec file to make a list of blocks.
+ * Return FALSE if the .vec fiel cannot be found
+ **********************************************************************/
+
+BOOL8 read_vec_file(                    //print list of sides
+                    STRING name,        //basename of file
+                    INT32 xsize,        //image size
+                    INT32 ysize,        //image size
+                    BLOCK_LIST *blocks  //output list
+                   ) {
+  FILE *pdfp;                    //file pointer
+  BLOCK *block;                  //current block
+  INT32 block_no;                //no of blocks
+  INT32 block_index;             //current blocks
+  INT32 vector_count;            //total vectors
+  VEC_HEADER header;             //file header
+  BLOCK_HEADER *vec_blocks;      //blocks from file
+  VEC_ENTRY *vec_entries;        //vectors from file
+  BLOCK_IT block_it = blocks;    //block iterator
+  ICOORDELT_IT left_it;          //iterators
+  ICOORDELT_IT right_it;
+
+  name += VEC_EXT;               //add extension
+  if ((pdfp = fopen (name.string (), "r")) == NULL) {
+    return FALSE;                //can't find it
+  }
+  if (fread (&header, sizeof (header), 1, pdfp) != 1)
+    READFAILED.error ("read_vec_file", EXIT, "Header");
+                                 //from intel
+  header.filesize = reverse32 (header.filesize);
+  header.bytesize = reverse16 (header.bytesize);
+  header.arraysize = reverse16 (header.arraysize);
+  header.width = reverse16 (header.width);
+  header.height = reverse16 (header.height);
+  header.res = reverse16 (header.res);
+  header.bpp = reverse16 (header.bpp);
+  tprintf ("%d blocks in %s file:", header.arraysize, VEC_EXT);
+  vector_count = header.filesize - header.arraysize * sizeof (BLOCK_HEADER);
+  vector_count /= sizeof (VEC_ENTRY);
+  vec_blocks =
+    (BLOCK_HEADER *) alloc_mem (header.arraysize * sizeof (BLOCK_HEADER));
+  vec_entries = (VEC_ENTRY *) alloc_mem (vector_count * sizeof (VEC_ENTRY));
+  xsize = header.width;          //real image size
+  ysize = header.height;
+  if (fread (vec_blocks, sizeof (BLOCK_HEADER), header.arraysize, pdfp)
+    != static_cast<size_t>(header.arraysize))
+    READFAILED.error ("read_vec_file", EXIT, "Blocks");
+  if (fread (vec_entries, sizeof (VEC_ENTRY), vector_count, pdfp)
+    != static_cast<size_t>(vector_count))
+    READFAILED.error ("read_vec_file", EXIT, "Vectors");
+  for (block_index = 0; block_index < header.arraysize; block_index++) {
+    vec_blocks[block_index].offset =
+      reverse16 (vec_blocks[block_index].offset);
+    vec_blocks[block_index].order =
+      reverse16 (vec_blocks[block_index].order);
+    vec_blocks[block_index].entries =
+      reverse16 (vec_blocks[block_index].entries);
+    vec_blocks[block_index].charsize =
+      reverse16 (vec_blocks[block_index].charsize);
+  }
+  for (block_index = 0; block_index < vector_count; block_index++) {
+    vec_entries[block_index].start =
+      ICOORD (reverse16 (vec_entries[block_index].start.x ()),
+      reverse16 (vec_entries[block_index].start.y ()));
+    vec_entries[block_index].end =
+      ICOORD (reverse16 (vec_entries[block_index].end.x ()),
+      reverse16 (vec_entries[block_index].end.y ()));
+  }
+  for (block_no = 1; block_no <= header.arraysize; block_no++) {
+    for (block_index = 0; block_index < header.arraysize; block_index++) {
+      if (vec_blocks[block_index].order == block_no
+      && vec_blocks[block_index].valid) {
+        block = new BLOCK;
+        left_it.set_to_list (&block->leftside);
+        right_it.set_to_list (&block->rightside);
+        block->box =
+          convert_vec_block (&vec_entries
+          [vec_blocks[block_index].offset],
+          vec_blocks[block_index].entries, ysize,
+          &left_it, &right_it);
+        block->set_xheight (vec_blocks[block_index].charsize);
+                                 //on end of list
+        block_it.add_to_end (block);
+        //                              tprintf("Block at (%d,%d)->(%d,%d) has index %d and order %d\n",
+        //                                      block->box.left(),
+        //                                      block->box.bottom(),
+        //                                      block->box.right(),
+        //                                      block->box.top(),
+        //                                      block_index,vec_blocks[block_index].order);
+      }
+    }
+  }
+  free_mem(vec_blocks);
+  free_mem(vec_entries);
+  tprintf ("%d valid\n", block_it.length ());
+  fclose(pdfp);
+  return TRUE;                   //read one
+}
+
+
+/**********************************************************************
+ * BLOCK::convert_vec_block
+ *
+ * Read a whole vec file to make a list of blocks.
+ * Return FALSE if the .vec fiel cannot be found
+ **********************************************************************/
+
+static BOX convert_vec_block(                        //make non-rect block
+                             VEC_ENTRY *entries,     //vectors
+                             UINT16 entry_count,     //no of entries
+                             INT32 ysize,            //image size
+                             ICOORDELT_IT *left_it,  //block sides
+                             ICOORDELT_IT *right_it) {
+  BOX block_box;                 //bounding box
+  BOX vec_box;                   //box of vec
+  ICOORD box_point;              //expanded coord
+  ICOORD shift_vec;              //for box expansion
+  ICOORD prev_pt;                //previous coord
+  ICOORD end_pt;                 //end of vector
+  INT32 vertex_index;            //boundary vertices
+
+  for (vertex_index = 0; vertex_index < entry_count; vertex_index++) {
+    entries[vertex_index].start = ICOORD (entries[vertex_index].start.x (),
+      ysize - 1 -
+      entries[vertex_index].start.y ());
+    entries[vertex_index].end =
+      ICOORD (entries[vertex_index].end.x (),
+      ysize - 1 - entries[vertex_index].end.y ());
+    vec_box = BOX (entries[vertex_index].start, entries[vertex_index].end);
+    block_box += vec_box;        //find total bounds
+  }
+
+  for (vertex_index = 0; vertex_index < entry_count
+    && (entries[vertex_index].start.y () != block_box.bottom ()
+    || entries[vertex_index].end.y () != block_box.bottom ());
+    vertex_index++);
+  ASSERT_HOST (vertex_index < entry_count);
+  prev_pt = entries[vertex_index].start;
+  end_pt = entries[vertex_index].end;
+  do {
+    for (vertex_index = 0; vertex_index < entry_count
+      && entries[vertex_index].start != end_pt; vertex_index++);
+                                 //found start of vertical
+    ASSERT_HOST (vertex_index < entry_count);
+    box_point = entries[vertex_index].start;
+    if (box_point.x () <= prev_pt.x ())
+      shift_vec = ICOORD (-BLOCK_EXPANSION, -BLOCK_EXPANSION);
+    else
+      shift_vec = ICOORD (-BLOCK_EXPANSION, BLOCK_EXPANSION);
+    left_it->add_to_end (new ICOORDELT (box_point + shift_vec));
+    prev_pt = box_point;
+    for (vertex_index = 0; vertex_index < entry_count
+      && entries[vertex_index].start != end_pt; vertex_index++);
+                                 //found horizontal
+    ASSERT_HOST (vertex_index < entry_count);
+    end_pt = entries[vertex_index].end;
+  }
+  while (end_pt.y () < block_box.top ());
+  shift_vec = ICOORD (-BLOCK_EXPANSION, BLOCK_EXPANSION);
+  left_it->add_to_end (new ICOORDELT (end_pt + shift_vec));
+
+  for (vertex_index = 0; vertex_index < entry_count
+    && (entries[vertex_index].start.y () != block_box.top ()
+    || entries[vertex_index].end.y () != block_box.top ());
+    vertex_index++);
+  ASSERT_HOST (vertex_index < entry_count);
+  prev_pt = entries[vertex_index].start;
+  end_pt = entries[vertex_index].end;
+  do {
+    for (vertex_index = 0; vertex_index < entry_count
+      && entries[vertex_index].start != end_pt; vertex_index++);
+                                 //found start of vertical
+    ASSERT_HOST (vertex_index < entry_count);
+    box_point = entries[vertex_index].start;
+    if (box_point.x () < prev_pt.x ())
+      shift_vec = ICOORD (BLOCK_EXPANSION, -BLOCK_EXPANSION);
+    else
+      shift_vec = ICOORD (BLOCK_EXPANSION, BLOCK_EXPANSION);
+    right_it->add_before_then_move (new ICOORDELT (box_point + shift_vec));
+    prev_pt = box_point;
+    for (vertex_index = 0; vertex_index < entry_count
+      && entries[vertex_index].start != end_pt; vertex_index++);
+                                 //found horizontal
+    ASSERT_HOST (vertex_index < entry_count);
+    end_pt = entries[vertex_index].end;
+  }
+  while (end_pt.y () > block_box.bottom ());
+  shift_vec = ICOORD (BLOCK_EXPANSION, -BLOCK_EXPANSION);
+  right_it->add_before_then_move (new ICOORDELT (end_pt + shift_vec));
+
+  shift_vec = ICOORD (BLOCK_EXPANSION, BLOCK_EXPANSION);
+  box_point = block_box.botleft () - shift_vec;
+  end_pt = block_box.topright () + shift_vec;
+  return BOX (box_point, end_pt);
+}
+
+
+/**********************************************************************
+ * read_unlv_file
+ *
+ * Read a whole unlv zone file to make a list of blocks.
+ **********************************************************************/
+
+BOOL8 read_unlv_file(                    //print list of sides
+                     STRING name,        //basename of file
+                     INT32 xsize,        //image size
+                     INT32 ysize,        //image size
+                     BLOCK_LIST *blocks  //output list
+                    ) {
+  FILE *pdfp;                    //file pointer
+  BLOCK *block;                  //current block
+  int x;                         //current top-down coords
+  int y;
+  int width;                     //of current block
+  int height;
+  BLOCK_IT block_it = blocks;    //block iterator
+
+  name += UNLV_EXT;              //add extension
+  if ((pdfp = fopen (name.string (), "r")) == NULL) {
+    return FALSE;                //didn't read one
+  }
+  else {
+    while (fscanf (pdfp, "%d %d %d %d %*s", &x, &y, &width, &height) >= 4) {
+                                 //make rect block
+      block = new BLOCK (name.string (), TRUE, 0, 0, (INT16) x, (INT16) (ysize - 1 - y - height), (INT16) (x + width), (INT16) (ysize - 1 - y));
+                                 //on end of list
+      block_it.add_to_end (block);
+    }
+    fclose(pdfp);
+  }
+  return true;
+}

ccstruct/blread.h

@@ -0,0 +1,63 @@
+/**********************************************************************
+ * File:        blread.h  (Formerly pdread.h)
+ * Description: Friend function of BLOCK to read the uscan pd file.
+ * Author:		Ray Smith
+ * Created:		Mon Mar 18 14:39:00 GMT 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           BLREAD_H
+#define           BLREAD_H
+
+#include          "varable.h"
+#include          "ocrblock.h"
+
+BOOL8 read_pd_file(                    //print list of sides
+                   STRING name,        //basename of file
+                   INT32 xsize,        //image size
+                   INT32 ysize,        //image size
+                   BLOCK_LIST *blocks  //output list
+                  );
+void get_pd_vertex(                //get new vertex
+                   FILE *pdfp,     //file to read
+                   INT32 xsize,    //image size
+                   INT32 ysize,    //image size
+                   BOX *box,       //bounding box
+                   INT32 &xcoord,  //output coords
+                   INT32 &ycoord);
+BOOL8 read_hpd_file(                    //print list of sides
+                    STRING name,        //basename of file
+                    INT32 xsize,        //image size
+                    INT32 ysize,        //image size
+                    BLOCK_LIST *blocks  //output list
+                   );
+void scan_hpd_blocks(                               //print list of sides
+                     const char *name,              //block label
+                     PAGE_BLOCK_LIST *page_blocks,  //head of full pag
+                     INT32 &block_no,               //no of blocks
+                     BLOCK_IT *block_it             //block iterator
+                    );
+BOOL8 read_vec_file(                    //print list of sides
+                    STRING name,        //basename of file
+                    INT32 xsize,        //image size
+                    INT32 ysize,        //image size
+                    BLOCK_LIST *blocks  //output list
+                   );
+BOOL8 read_unlv_file(                    //print list of sides
+                     STRING name,        //basename of file
+                     INT32 xsize,        //image size
+                     INT32 ysize,        //image size
+                     BLOCK_LIST *blocks  //output list
+                    );
+#endif

ccstruct/callcpp.cpp

@@ -0,0 +1,270 @@
+/**********************************************************************
+ * File:        callcpp.cpp
+ * Description: extern C interface calling C++ from C.
+ * Author:		Ray Smith
+ * Created:		Sun Feb 04 20:39:23 MST 1996
+ *
+ * (C) Copyright 1996, Hewlett-Packard Co.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          "errcode.h"
+#ifdef __UNIX__
+#include          <assert.h>
+#include <stdarg.h>
+#endif
+#include          <time.h>
+#include          "memry.h"
+#include          "grphics.h"
+#include          "evnts.h"
+#include          "varable.h"
+#include          "callcpp.h"
+#include          "tprintf.h"
+//#include                                      "strace.h"
+#include          "host.h"
+
+//extern "C" {
+
+INT_VAR (tess_cp_mapping0, 0, "Mappings for class pruner distance");
+INT_VAR (tess_cp_mapping1, 1, "Mappings for class pruner distance");
+INT_VAR (tess_cp_mapping2, 2, "Mappings for class pruner distance");
+INT_VAR (tess_cp_mapping3, 3, "Mappings for class pruner distance");
+INT_VAR (stopper_numbers_on, 0, "Allow numbers to be acceptable choices");
+INT_VAR (config_pruner_enabled, 0, "Turn on config pruner");
+INT_VAR (feature_prune_percentile, 0, "Percent of features to use");
+INT_VAR (newcp_ratings_on, 0, "Use new class pruner normalisation");
+INT_VAR (record_matcher_output, 0, "Record detailed matcher info");
+INT_VAR (il1_adaption_test, 0, "Dont adapt to i/I at beginning of word");
+double_VAR (permuter_pending_threshold, 0.0,
+"Worst conf for using pending dictionary");
+double_VAR (newcp_duff_rating, 0.30, "Worst rating for calling real matcher");
+double_VAR (newcp_prune_threshold, 1.2, "Ratio of best to prune");
+double_VAR (tessedit_cp_ratio, 0.0, "Ratio from best to prune");
+//Global matcher info from the class pruner.
+INT32 cp_classes;
+INT32 cp_bestindex;
+INT32 cp_bestrating;
+INT32 cp_bestconf;
+char cp_chars[2];
+INT32 cp_ratings[2];
+INT32 cp_confs[2];
+INT32 cp_maps[4];
+//Global info to control writes of matcher info
+INT32 blob_type;                 //write control
+char blob_answer;                //correct char
+char *word_answer;               //correct word
+INT32 matcher_pass;              //pass in chopper.c
+INT32 bits_in_states;            //no of bits in states
+
+#ifndef __UNIX__
+/**********************************************************************
+ * assert
+ *
+ * A version of assert for C on NT.
+ **********************************************************************/
+
+void assert(             //recog one owrd
+            int testing  //assert fail if false
+           ) {
+  ASSERT_HOST(testing); 
+}
+#endif
+
+void setup_cp_maps() { 
+  cp_maps[0] = tess_cp_mapping0;
+  cp_maps[1] = tess_cp_mapping1;
+  cp_maps[2] = tess_cp_mapping2;
+  cp_maps[3] = tess_cp_mapping3;
+}
+
+
+void trace_stack() {  //Trace current stack
+}
+
+
+void
+cprintf (                        //Trace printf
+const char *format, ...          //special message
+) {
+  va_list args;                  //variable args
+  char msg[1000];
+
+  va_start(args, format);  //variable list
+  vsprintf(msg, format, args);  //Format into msg
+  va_end(args); 
+
+  tprintf ("%s", msg);
+}
+
+
+char *c_alloc_string(             //allocate string
+                     INT32 count  //no of chars required
+                    ) {
+  return alloc_string (count);
+}
+
+
+void c_free_string(              //free a string
+                   char *string  //string to free
+                  ) {
+  free_string(string); 
+}
+
+
+void *c_alloc_struct(                  //allocate memory
+                     INT32 count,      //no of chars required
+                     const char *name  //class name
+                    ) {
+  return alloc_struct (count, name);
+}
+
+
+void c_free_struct(                   //free a structure
+                   void *deadstruct,  //structure to free
+                   INT32 count,       //no of bytes
+                   const char *name   //class name
+                  ) {
+  free_struct(deadstruct, count, name); 
+}
+
+
+void *c_alloc_mem_p(             //allocate permanent space
+                    INT32 count  //block size to allocate
+                   ) {
+  return alloc_mem_p (count);
+}
+
+
+void *c_alloc_mem(             //get some memory
+                  INT32 count  //no of bytes to get
+                 ) {
+  return alloc_mem (count);
+}
+
+
+void c_free_mem(                //free mem from alloc_mem
+                void *oldchunk  //chunk to free
+               ) {
+  free_mem(oldchunk); 
+}
+
+
+void c_check_mem(                     //check consistency
+                 const char *string,  //context message
+                 INT8 level           //level of check
+                ) {
+  check_mem(string, level); 
+}
+
+#ifndef GRAPHICS_DISABLED
+void *c_create_window(                   /*create a window */
+                      const char *name,  /*name/title of window */
+                      INT16 xpos,        /*coords of window */
+                      INT16 ypos,        /*coords of window */
+                      INT16 xsize,       /*size of window */
+                      INT16 ysize,       /*size of window */
+                      double xmin,       /*scrolling limits */
+                      double xmax,       /*to stop users */
+                      double ymin,       /*getting lost in */
+                      double ymax        /*empty space */
+                     ) {
+  return create_window (name, SCROLLINGWIN, xpos, ypos, xsize, ysize,
+    xmin, xmax, ymin, ymax, TRUE, FALSE, FALSE, TRUE);
+}
+
+
+void c_line_color_index(  /*set color */
+                        void *win,
+                        C_COL index) {
+  WINDOW window = (WINDOW) win;
+
+  //      ASSERT_HOST(index>=0 && index<=48);
+  if (index < 0 || index > 48)
+    index = (C_COL) 1;
+  window->Line_color_index ((COLOUR) index);
+}
+
+
+void c_move(  /*move pen */
+            void *win,
+            double x,
+            double y) {
+  WINDOW window = (WINDOW) win;
+
+  window->Move2d (x, y);
+}
+
+
+void c_draw(  /*move pen */
+            void *win,
+            double x,
+            double y) {
+  WINDOW window = (WINDOW) win;
+
+  window->Draw2d (x, y);
+}
+
+
+void c_make_current(  /*move pen */
+                    void *win) {
+  WINDOW window = (WINDOW) win;
+
+  window->Make_picture_current ();
+}
+
+
+void c_clear_window(  /*move pen */
+                    void *win) {
+  WINDOW window = (WINDOW) win;
+
+  window->Clear_view_surface ();
+}
+
+
+char window_wait(  /*move pen */
+                 void *win) {
+  WINDOW window = (WINDOW) win;
+  GRAPHICS_EVENT event;
+
+  await_event(window, TRUE, ANY_EVENT, &event); 
+  if (event.type == KEYPRESS_EVENT)
+    return event.key;
+  else
+    return '\0';
+}
+#endif
+
+void reverse32(void *ptr) { 
+  char tmp;
+  char *cptr = (char *) ptr;
+
+  tmp = *cptr;
+  *cptr = *(cptr + 3);
+  *(cptr + 3) = tmp;
+  tmp = *(cptr + 1);
+  *(cptr + 1) = *(cptr + 2);
+  *(cptr + 2) = tmp;
+}
+
+
+void reverse16(void *ptr) { 
+  char tmp;
+  char *cptr = (char *) ptr;
+
+  tmp = *cptr;
+  *cptr = *(cptr + 1);
+  *(cptr + 1) = tmp;
+}
+
+
+//};

ccstruct/coutln.cpp

@@ -0,0 +1,604 @@
+/**********************************************************************
+ * File:        coutln.c  (Formerly coutline.c)
+ * Description: Code for the C_OUTLINE class.
+ * Author:					Ray Smith
+ * Created:					Mon Oct 07 16:01:57 BST 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          <string.h>
+#ifdef __UNIX__
+#include          <assert.h>
+#endif
+#include          "coutln.h"
+
+ELISTIZE_S (C_OUTLINE)
+ICOORD C_OUTLINE::step_coords[4] = {
+  ICOORD (-1, 0), ICOORD (0, -1), ICOORD (1, 0), ICOORD (0, 1)
+};
+
+/**********************************************************************
+ * C_OUTLINE::C_OUTLINE
+ *
+ * Constructor to build a C_OUTLINE from a CRACKEDGE LOOP.
+ **********************************************************************/
+
+C_OUTLINE::C_OUTLINE (
+//constructor
+CRACKEDGE * startpt,             //outline to convert
+ICOORD bot_left,                 //bounding box
+ICOORD top_right, INT16 length   //length of loop
+):box (bot_left, top_right), start (startpt->pos) {
+  INT16 stepindex;               //index to step
+  CRACKEDGE *edgept;             //current point
+
+  stepcount = length;            //no of steps
+                                 //get memory
+  steps = (UINT8 *) alloc_mem (step_mem());
+  memset(steps, 0, step_mem());
+  edgept = startpt;
+
+  for (stepindex = 0; stepindex < length; stepindex++) {
+                                 //set compact step
+    set_step (stepindex, edgept->stepdir);
+    edgept = edgept->next;
+  }
+}
+
+
+/**********************************************************************
+ * C_OUTLINE::C_OUTLINE
+ *
+ * Constructor to build a C_OUTLINE from a C_OUTLINE_FRAG.
+ **********************************************************************/
+C_OUTLINE::C_OUTLINE (
+//constructor
+                                 //steps to copy
+ICOORD startpt, DIR128 * new_steps,
+INT16 length                     //length of loop
+):start (startpt) {
+  INT8 dirdiff;                  //direction difference
+  DIR128 prevdir;                //previous direction
+  DIR128 dir;                    //current direction
+  DIR128 lastdir;                //dir of last step
+  BOX new_box;                   //easy bounding
+  INT16 stepindex;               //index to step
+  INT16 srcindex;                //source steps
+  ICOORD pos;                    //current position
+
+  pos = startpt;
+  stepcount = length;            //no of steps
+                                 //get memory
+  steps = (UINT8 *) alloc_mem (step_mem());
+  memset(steps, 0, step_mem());
+
+  lastdir = new_steps[length - 1];
+  prevdir = lastdir;
+  for (stepindex = 0, srcindex = 0; srcindex < length;
+  stepindex++, srcindex++) {
+    new_box = BOX (pos, pos);
+    box += new_box;
+                                 //copy steps
+    dir = new_steps[srcindex];
+    set_step(stepindex, dir);
+    dirdiff = dir - prevdir;
+    pos += step (stepindex);
+    if ((dirdiff == 64 || dirdiff == -64) && stepindex > 0) {
+      stepindex -= 2;            //cancel there-and-back
+      prevdir = stepindex >= 0 ? step_dir (stepindex) : lastdir;
+    }
+    else
+      prevdir = dir;
+  }
+  ASSERT_HOST (pos.x () == startpt.x () && pos.y () == startpt.y ());
+  do {
+    dirdiff = step_dir (stepindex - 1) - step_dir (0);
+    if (dirdiff == 64 || dirdiff == -64) {
+      start += step (0);
+      stepindex -= 2;            //cancel there-and-back
+      for (int i = 0; i < stepindex; ++i)
+        set_step(i, step_dir(i + 1));
+    }
+  }
+  while (stepindex > 1 && (dirdiff == 64 || dirdiff == -64));
+  stepcount = stepindex;
+  ASSERT_HOST (stepcount >= 4);
+}
+
+/**********************************************************************
+ * C_OUTLINE::C_OUTLINE
+ *
+ * Constructor to build a C_OUTLINE from a rotation of a C_OUTLINE.
+ **********************************************************************/
+
+C_OUTLINE::C_OUTLINE(                     //constructor
+                     C_OUTLINE *srcline,  //outline to
+                     FCOORD rotation      //rotate
+                    ) {
+  BOX new_box;                   //easy bounding
+  INT16 stepindex;               //index to step
+  INT16 dirdiff;                 //direction change
+  ICOORD pos;                    //current position
+  ICOORD prevpos;                //previous dest point
+
+  ICOORD destpos;                //destination point
+  INT16 destindex;               //index to step
+  DIR128 dir;                    //coded direction
+  UINT8 new_step;
+
+  stepcount = srcline->stepcount * 2;
+                                 //get memory
+  steps = (UINT8 *) alloc_mem (step_mem());
+  memset(steps, 0, step_mem());
+
+  for (int iteration = 0; iteration < 2; ++iteration) {
+    DIR128 round1 = iteration == 0 ? 32 : 0;
+    DIR128 round2 = iteration != 0 ? 32 : 0;
+    pos = srcline->start;
+    prevpos = pos;
+    prevpos.rotate (rotation);
+    start = prevpos;
+    box = BOX (start, start);
+    destindex = 0;
+    for (stepindex = 0; stepindex < srcline->stepcount; stepindex++) {
+      pos += srcline->step (stepindex);
+      destpos = pos;
+      destpos.rotate (rotation);
+      if (destpos.x () != prevpos.x () || destpos.y () != prevpos.y ()) {
+        dir = DIR128 (FCOORD (destpos - prevpos));
+        dir += 64;                 //turn to step style
+        new_step = dir.get_dir ();
+        if (new_step & 31) {
+          set_step(destindex++, dir + round1);
+          if (destindex < 2
+            || (dirdiff =
+            step_dir (destindex - 1) - step_dir (destindex - 2)) !=
+            -64 && dirdiff != 64)
+            set_step(destindex++, dir + round2);
+          else {
+            set_step(destindex - 1, dir + round2);
+            set_step(destindex++, dir + round1);
+          }
+        }
+        else {
+          set_step(destindex++, dir);
+          if (destindex >= 2
+            &&
+            ((dirdiff =
+            step_dir (destindex - 1) - step_dir (destindex - 2)) ==
+            -64 || dirdiff == 64))
+            destindex -= 2;        // Forget u turn
+        }
+        prevpos = destpos;
+        new_box = BOX (destpos, destpos);
+        box += new_box;
+      }
+    }
+    ASSERT_HOST (destpos.x () == start.x () && destpos.y () == start.y ());
+    dirdiff = step_dir (destindex - 1) - step_dir (0);
+    while ((dirdiff == 64 || dirdiff == -64) && destindex > 1) {
+      start += step (0);
+      destindex -= 2;
+      for (int i = 0; i < destindex; ++i)
+        set_step(i, step_dir(i + 1));
+      dirdiff = step_dir (destindex - 1) - step_dir (0);
+    }
+    if (destindex >= 4)
+      break;
+  }
+  stepcount = destindex;
+  destpos = start;
+  for (stepindex = 0; stepindex < stepcount; stepindex++) {
+    destpos += step (stepindex);
+  }
+  ASSERT_HOST (destpos.x () == start.x () && destpos.y () == start.y ());
+}
+
+
+/**********************************************************************
+ * C_OUTLINE::area
+ *
+ * Compute the area of the outline.
+ **********************************************************************/
+
+INT32 C_OUTLINE::area() {  //winding number
+  int stepindex;                 //current step
+  INT32 total_steps;             //steps to do
+  INT32 total;                   //total area
+  ICOORD pos;                    //position of point
+  ICOORD next_step;              //step to next pix
+  C_OUTLINE_IT it = child ();
+
+  pos = start_pos ();
+  total_steps = pathlength ();
+  total = 0;
+  for (stepindex = 0; stepindex < total_steps; stepindex++) {
+                                 //all intersected
+    next_step = step (stepindex);
+    if (next_step.x () < 0)
+      total += pos.y ();
+    else if (next_step.x () > 0)
+      total -= pos.y ();
+    pos += next_step;
+  }
+  for (it.mark_cycle_pt (); !it.cycled_list (); it.forward ())
+    total += it.data ()->area ();//add areas of children
+
+  return total;
+}
+
+
+/**********************************************************************
+ * C_OUTLINE::outer_area
+ *
+ * Compute the area of the outline.
+ **********************************************************************/
+
+INT32 C_OUTLINE::outer_area() {  //winding number
+  int stepindex;                 //current step
+  INT32 total_steps;             //steps to do
+  INT32 total;                   //total area
+  ICOORD pos;                    //position of point
+  ICOORD next_step;              //step to next pix
+
+  pos = start_pos ();
+  total_steps = pathlength ();
+  total = 0;
+  for (stepindex = 0; stepindex < total_steps; stepindex++) {
+                                 //all intersected
+    next_step = step (stepindex);
+    if (next_step.x () < 0)
+      total += pos.y ();
+    else if (next_step.x () > 0)
+      total -= pos.y ();
+    pos += next_step;
+  }
+
+  return total;
+}
+
+
+/**********************************************************************
+ * C_OUTLINE::count_transitions
+ *
+ * Compute the number of x and y maxes and mins in the outline.
+ **********************************************************************/
+
+INT32 C_OUTLINE::count_transitions(                 //winding number
+                                   INT32 threshold  //on size
+                                  ) {
+  BOOL8 first_was_max_x;         //what was first
+  BOOL8 first_was_max_y;
+  BOOL8 looking_for_max_x;       //what is next
+  BOOL8 looking_for_min_x;
+  BOOL8 looking_for_max_y;       //what is next
+  BOOL8 looking_for_min_y;
+  int stepindex;                 //current step
+  INT32 total_steps;             //steps to do
+                                 //current limits
+  INT32 max_x, min_x, max_y, min_y;
+  INT32 initial_x, initial_y;    //initial limits
+  INT32 total;                   //total changes
+  ICOORD pos;                    //position of point
+  ICOORD next_step;              //step to next pix
+
+  pos = start_pos ();
+  total_steps = pathlength ();
+  total = 0;
+  max_x = min_x = pos.x ();
+  max_y = min_y = pos.y ();
+  looking_for_max_x = TRUE;
+  looking_for_min_x = TRUE;
+  looking_for_max_y = TRUE;
+  looking_for_min_y = TRUE;
+  first_was_max_x = FALSE;
+  first_was_max_y = FALSE;
+  initial_x = pos.x ();
+  initial_y = pos.y ();          //stop uninit warning
+  for (stepindex = 0; stepindex < total_steps; stepindex++) {
+                                 //all intersected
+    next_step = step (stepindex);
+    pos += next_step;
+    if (next_step.x () < 0) {
+      if (looking_for_max_x && pos.x () < min_x)
+        min_x = pos.x ();
+      if (looking_for_min_x && max_x - pos.x () > threshold) {
+        if (looking_for_max_x) {
+          initial_x = max_x;
+          first_was_max_x = FALSE;
+        }
+        total++;
+        looking_for_max_x = TRUE;
+        looking_for_min_x = FALSE;
+        min_x = pos.x ();        //reset min
+      }
+    }
+    else if (next_step.x () > 0) {
+      if (looking_for_min_x && pos.x () > max_x)
+        max_x = pos.x ();
+      if (looking_for_max_x && pos.x () - min_x > threshold) {
+        if (looking_for_min_x) {
+          initial_x = min_x;     //remember first min
+          first_was_max_x = TRUE;
+        }
+        total++;
+        looking_for_max_x = FALSE;
+        looking_for_min_x = TRUE;
+        max_x = pos.x ();
+      }
+    }
+    else if (next_step.y () < 0) {
+      if (looking_for_max_y && pos.y () < min_y)
+        min_y = pos.y ();
+      if (looking_for_min_y && max_y - pos.y () > threshold) {
+        if (looking_for_max_y) {
+          initial_y = max_y;     //remember first max
+          first_was_max_y = FALSE;
+        }
+        total++;
+        looking_for_max_y = TRUE;
+        looking_for_min_y = FALSE;
+        min_y = pos.y ();        //reset min
+      }
+    }
+    else {
+      if (looking_for_min_y && pos.y () > max_y)
+        max_y = pos.y ();
+      if (looking_for_max_y && pos.y () - min_y > threshold) {
+        if (looking_for_min_y) {
+          initial_y = min_y;     //remember first min
+          first_was_max_y = TRUE;
+        }
+        total++;
+        looking_for_max_y = FALSE;
+        looking_for_min_y = TRUE;
+        max_y = pos.y ();
+      }
+    }
+
+  }
+  if (first_was_max_x && looking_for_min_x) {
+    if (max_x - initial_x > threshold)
+      total++;
+    else
+      total--;
+  }
+  else if (!first_was_max_x && looking_for_max_x) {
+    if (initial_x - min_x > threshold)
+      total++;
+    else
+      total--;
+  }
+  if (first_was_max_y && looking_for_min_y) {
+    if (max_y - initial_y > threshold)
+      total++;
+    else
+      total--;
+  }
+  else if (!first_was_max_y && looking_for_max_y) {
+    if (initial_y - min_y > threshold)
+      total++;
+    else
+      total--;
+  }
+
+  return total;
+}
+
+
+/**********************************************************************
+ * C_OUTLINE::operator<
+ *
+ * Return TRUE if the left operand is inside the right one.
+ **********************************************************************/
+
+BOOL8
+C_OUTLINE::operator< (           //winding number
+const C_OUTLINE & other          //other outline
+) const
+{
+  INT16 count = 0;               //winding count
+  ICOORD pos;                    //position of point
+  INT32 stepindex;               //index to cstep
+
+  if (!box.overlap (other.box))
+    return FALSE;                //can't be contained
+
+  pos = start;
+  for (stepindex = 0; stepindex < stepcount
+    && (count = other.winding_number (pos)) == INTERSECTING; stepindex++)
+    pos += step (stepindex);     //try all points
+  if (count == INTERSECTING) {
+                                 //all intersected
+    pos = other.start;
+    for (stepindex = 0; stepindex < other.stepcount
+      && (count = winding_number (pos)) == INTERSECTING; stepindex++)
+                                 //try other way round
+      pos += other.step (stepindex);
+    return count == INTERSECTING || count == 0;
+  }
+  return count != 0;
+}
+
+
+/**********************************************************************
+ * C_OUTLINE::winding_number
+ *
+ * Return the winding number of the outline around the given point.
+ **********************************************************************/
+
+INT16 C_OUTLINE::winding_number(              //winding number
+                                ICOORD point  //point to wind around
+                               ) const {
+  INT16 stepindex;               //index to cstep
+  INT16 count;                   //winding count
+  ICOORD vec;                    //to current point
+  ICOORD stepvec;                //step vector
+  INT32 cross;                   //cross product
+
+  vec = start - point;           //vector to it
+  count = 0;
+  for (stepindex = 0; stepindex < stepcount; stepindex++) {
+    stepvec = step (stepindex);  //get the step
+                                 //crossing the line
+    if (vec.y () <= 0 && vec.y () + stepvec.y () > 0) {
+      cross = vec * stepvec;     //cross product
+      if (cross > 0)
+        count++;                 //crossing right half
+      else if (cross == 0)
+        return INTERSECTING;     //going through point
+    }
+    else if (vec.y () > 0 && vec.y () + stepvec.y () <= 0) {
+      cross = vec * stepvec;
+      if (cross < 0)
+        count--;                 //crossing back
+      else if (cross == 0)
+        return INTERSECTING;     //illegal
+    }
+    vec += stepvec;              //sum vectors
+  }
+  return count;                  //winding number
+}
+
+
+/**********************************************************************
+ * C_OUTLINE::turn_direction
+ *
+ * Return the sum direction delta of the outline.
+ **********************************************************************/
+
+INT16 C_OUTLINE::turn_direction() const {  //winding number
+  DIR128 prevdir;                //previous direction
+  DIR128 dir;                    //current direction
+  INT16 stepindex;               //index to cstep
+  INT8 dirdiff;                  //direction difference
+  INT16 count;                   //winding count
+
+  count = 0;
+  prevdir = step_dir (stepcount - 1);
+  for (stepindex = 0; stepindex < stepcount; stepindex++) {
+    dir = step_dir (stepindex);
+    dirdiff = dir - prevdir;
+    ASSERT_HOST (dirdiff == 0 || dirdiff == 32 || dirdiff == -32);
+    count += dirdiff;
+    prevdir = dir;
+  }
+  ASSERT_HOST (count == 128 || count == -128);
+  return count;                  //winding number
+}
+
+
+/**********************************************************************
+ * C_OUTLINE::reverse
+ *
+ * Reverse the direction of an outline.
+ **********************************************************************/
+
+void C_OUTLINE::reverse() {  //reverse drection
+  DIR128 halfturn = MODULUS / 2; //amount to shift
+  DIR128 stepdir;                //direction of step
+  INT16 stepindex;               //index to cstep
+  INT16 farindex;                //index to other side
+  INT16 halfsteps;               //half of stepcount
+
+  halfsteps = (stepcount + 1) / 2;
+  for (stepindex = 0; stepindex < halfsteps; stepindex++) {
+    farindex = stepcount - stepindex - 1;
+    stepdir = step_dir (stepindex);
+    set_step (stepindex, step_dir (farindex) + halfturn);
+    set_step (farindex, stepdir + halfturn);
+  }
+}
+
+
+/**********************************************************************
+ * C_OUTLINE::move
+ *
+ * Move C_OUTLINE by vector
+ **********************************************************************/
+
+void C_OUTLINE::move(                  // reposition OUTLINE
+                     const ICOORD vec  // by vector
+                    ) {
+  C_OUTLINE_IT it(&children);  // iterator
+
+  box.move (vec);
+  start += vec;
+
+  for (it.mark_cycle_pt (); !it.cycled_list (); it.forward ())
+    it.data ()->move (vec);      // move child outlines
+}
+
+
+/**********************************************************************
+ * C_OUTLINE::plot
+ *
+ * Draw the outline in the given colour.
+ **********************************************************************/
+
+#ifndef GRAPHICS_DISABLED
+void C_OUTLINE::plot(                //draw it
+                     WINDOW window,  //window to draw in
+                     COLOUR colour   //colour to draw in
+                    ) const {
+  INT16 stepindex;               //index to cstep
+  ICOORD pos;                    //current position
+  DIR128 stepdir;                //direction of step
+  DIR128 oldstepdir;             //previous stepdir
+
+  pos = start;                   //current position
+  line_color_index(window, colour);
+  move2d (window, pos.x (), pos.y ());
+  stepindex = 0;
+  stepdir = step_dir (0);        //get direction
+  while (stepindex < stepcount) {
+    do {
+      pos += step (stepindex);   //step to next
+      stepindex++;               //count steps
+      oldstepdir = stepdir;
+                                 //new direction
+      stepdir = step_dir (stepindex);
+    }
+    while (stepindex < stepcount
+      && oldstepdir.get_dir () == stepdir.get_dir ());
+    //merge straight lines
+    draw2d (window, pos.x (), pos.y ());
+  }
+}
+#endif
+
+
+/**********************************************************************
+ * C_OUTLINE::operator=
+ *
+ * Assignment - deep copy data
+ **********************************************************************/
+
+                                 //assignment
+C_OUTLINE & C_OUTLINE::operator= (
+const C_OUTLINE & source         //from this
+) {
+  box = source.box;
+  start = source.start;
+  if (steps != NULL)
+    free_mem(steps);
+  stepcount = source.stepcount;
+  steps = (UINT8 *) alloc_mem (step_mem());
+  memmove (steps, source.steps, step_mem());
+  if (!children.empty ())
+    children.clear ();
+  children.deep_copy (&source.children);
+  return *this;
+}

ccstruct/coutln.h

@@ -0,0 +1,176 @@
+/**********************************************************************
+ * File:						coutln.c      (Formerly:  coutline.c)
+ * Description: Code for the C_OUTLINE class.
+ * Author:					Ray Smith
+ * Created:					Mon Oct 07 16:01:57 BST 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           COUTLN_H
+#define           COUTLN_H
+
+#include          "grphics.h"
+#include          "crakedge.h"
+#include          "mod128.h"
+#include          "bits16.h"
+#include          "rect.h"
+#include          "blckerr.h"
+
+#define INTERSECTING    MAX_INT16//no winding number
+
+                                 //mask to get step
+#define STEP_MASK       3
+
+enum C_OUTLINE_FLAGS
+{
+  COUT_INVERSE                   //White on black blob
+};
+
+class DLLSYM C_OUTLINE;          //forward declaration
+
+ELISTIZEH_S (C_OUTLINE)
+class DLLSYM C_OUTLINE:public ELIST_LINK
+{
+  public:
+    C_OUTLINE() {  //empty constructor
+      steps = NULL;
+    }
+    C_OUTLINE(                     //constructor
+              CRACKEDGE *startpt,  //from edge detector
+              ICOORD bot_left,     //bounding box //length of loop
+              ICOORD top_right,
+              INT16 length);
+    C_OUTLINE(ICOORD startpt,    //start of loop
+              DIR128 *new_steps,  //steps in loop
+              INT16 length);     //length of loop
+                                 //outline to copy
+    C_OUTLINE(C_OUTLINE *srcline, FCOORD rotation);  //and rotate
+    ~C_OUTLINE () {              //destructor
+      if (steps != NULL)
+        free_mem(steps);
+      steps = NULL;
+    }
+
+    BOOL8 flag(                               //test flag
+               C_OUTLINE_FLAGS mask) const {  //flag to test
+      return flags.bit (mask);
+    }
+    void set_flag(                       //set flag value
+                  C_OUTLINE_FLAGS mask,  //flag to test
+                  BOOL8 value) {         //value to set
+      flags.set_bit (mask, value);
+    }
+
+    C_OUTLINE_LIST *child() {  //get child list
+      return &children;
+    }
+
+                                 //access function
+    const BOX &bounding_box() const {
+      return box;
+    }
+    void set_step(                    //set a step
+                  INT16 stepindex,    //index of step
+                  INT8 stepdir) {     //chain code
+      int shift = stepindex%4 * 2;
+      UINT8 mask = 3 << shift;
+      steps[stepindex/4] = ((stepdir << shift) & mask) |
+                           (steps[stepindex/4] & ~mask);
+      //squeeze 4 into byte
+    }
+    void set_step(                    //set a step
+                  INT16 stepindex,    //index of step
+                  DIR128 stepdir) {   //direction
+                                 //clean it
+      INT8 chaindir = stepdir.get_dir() >> (DIRBITS - 2);
+                                 //difference
+      set_step(stepindex, chaindir);
+      //squeeze 4 into byte
+    }
+
+                                 //get start position
+    const ICOORD &start_pos() const {
+      return start;
+    }
+    INT32 pathlength() const {  //get path length
+      return stepcount;
+    }
+    // Return step at a given index as a DIR128.
+    DIR128 step_dir(INT16 index) const {
+      return DIR128((INT16)(((steps[index/4] >> (index%4 * 2)) & STEP_MASK) <<
+                      (DIRBITS - 2)));
+    }
+    // Return the step vector for the given outline position.
+    ICOORD step(INT16 index) const { //index of step
+      return step_coords[(steps[index/4] >> (index%4 * 2)) & STEP_MASK];
+    }
+
+    INT32 area();  //return area
+    INT32 outer_area();  //return area
+    INT32 count_transitions(                   //count maxima
+                            INT32 threshold);  //size threshold
+
+    BOOL8 operator< (            //containment test
+      const C_OUTLINE & other) const;
+    BOOL8 operator> (            //containment test
+      C_OUTLINE & other) const
+    {
+      return other < *this;      //use the < to do it
+    }
+    INT16 winding_number(                       //get winding number
+                         ICOORD testpt) const;  //around this point
+                                 //get direction
+    INT16 turn_direction() const;
+    void reverse();  //reverse direction
+
+    void move(                    // reposition outline
+              const ICOORD vec);  // by vector
+
+    void plot(                       //draw one
+              WINDOW window,         //window to draw in
+              COLOUR colour) const;  //colour to draw it
+
+    void prep_serialise() {  //set ptrs to counts
+      children.prep_serialise ();
+    }
+
+    void dump(  //write external bits
+              FILE *f) {
+                                 //stepcount = # bytes
+      serialise_bytes (f, (void *) steps, step_mem());
+      children.dump (f);
+    }
+
+    void de_dump(  //read external bits
+                 FILE *f) {
+      steps = (UINT8 *) de_serialise_bytes (f, step_mem());
+      children.de_dump (f);
+    }
+
+                                 //assignment
+    make_serialise (C_OUTLINE) C_OUTLINE & operator= (
+      const C_OUTLINE & source); //from this
+
+  private:
+    int step_mem() const { return (stepcount+3) / 4; }
+
+    BOX box;                     //boudning box
+    ICOORD start;                //start coord
+    UINT8 *steps;                //step array
+    INT16 stepcount;             //no of steps
+    BITS16 flags;                //flags about outline
+    C_OUTLINE_LIST children;     //child elements
+    static ICOORD step_coords[4];
+};
+#endif

ccstruct/crakedge.h

@@ -0,0 +1,39 @@
+/**********************************************************************
+ * File:						crakedge.h      (Formerly: crkedge.h)
+ * Description: Sturctures for the Crack following edge detector.
+ * Author:					Ray Smith
+ * Created:					Fri Mar 22 16:06:38 GMT 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           CRAKEDGE_H
+#define           CRAKEDGE_H
+
+#include          "points.h"
+#include          "mod128.h"
+
+class CRACKEDGE
+{
+  public:
+    ICOORD pos;                  /*position of crack */
+    INT8 stepx;                  //edge step
+    INT8 stepy;
+    INT8 stepdir;                //chaincode
+    CRACKEDGE *prev;             /*previous point */
+    CRACKEDGE *next;             /*next point */
+
+    NEWDELETE2 (CRACKEDGE) CRACKEDGE () {
+    }                            //empty constructor
+};
+#endif

ccstruct/genblob.cpp

@@ -0,0 +1,133 @@
+/**********************************************************************
+ * File:        genblob.cpp  (Formerly gblob.c)
+ * Description: Generic Blob processing routines
+ * Author:      Phil Cheatle
+ * Created:     Mon Nov 25 10:53:26 GMT 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          "stepblob.h"
+#include          "polyblob.h"
+#include          "genblob.h"
+
+/**********************************************************************
+ *  blob_comparator()
+ *
+ *  Blob comparator used to sort a blob list so that blobs are in increasing
+ *  order of left edge.
+ **********************************************************************/
+
+int blob_comparator(                     //sort blobs
+                    const void *blob1p,  //ptr to ptr to blob1
+                    const void *blob2p   //ptr to ptr to blob2
+                   ) {
+  PBLOB *blob1 = *(PBLOB **) blob1p;
+  PBLOB *blob2 = *(PBLOB **) blob2p;
+
+  return blob1->bounding_box ().left () - blob2->bounding_box ().left ();
+}
+
+
+/**********************************************************************
+ *  c_blob_comparator()
+ *
+ *  Blob comparator used to sort a blob list so that blobs are in increasing
+ *  order of left edge.
+ **********************************************************************/
+
+int c_blob_comparator(                     //sort blobs
+                      const void *blob1p,  //ptr to ptr to blob1
+                      const void *blob2p   //ptr to ptr to blob2
+                     ) {
+  C_BLOB *blob1 = *(C_BLOB **) blob1p;
+  C_BLOB *blob2 = *(C_BLOB **) blob2p;
+
+  return blob1->bounding_box ().left () - blob2->bounding_box ().left ();
+}
+
+
+/**********************************************************************
+ *  gblob_bounding_box()
+ *
+ *  Return the bounding box of a generic blob.
+ **********************************************************************/
+
+BOX gblob_bounding_box(                 //Get bounding box
+                       PBLOB *blob,     //generic blob
+                       BOOL8 polygonal  //is blob polygonal?
+                      ) {
+  if (polygonal)
+    return blob->bounding_box ();
+  else
+    return ((C_BLOB *) blob)->bounding_box ();
+}
+
+
+/**********************************************************************
+ *  gblob_sort_list()
+ *
+ *  Sort a generic blob list into order of bounding box left edge
+ **********************************************************************/
+
+void gblob_sort_list(                        //Sort a gblob list
+                     PBLOB_LIST *blob_list,  //generic blob list
+                     BOOL8 polygonal         //is list polygonal?
+                    ) {
+  PBLOB_IT b_it;
+  C_BLOB_IT c_it;
+
+  if (polygonal) {
+    b_it.set_to_list (blob_list);
+    b_it.sort (blob_comparator);
+  }
+  else {
+    c_it.set_to_list ((C_BLOB_LIST *) blob_list);
+    c_it.sort (c_blob_comparator);
+  }
+}
+
+
+/**********************************************************************
+ *  gblob_out_list()
+ *
+ *  Return the generic outline list of a generic blob.
+ **********************************************************************/
+
+OUTLINE_LIST *gblob_out_list(                 //Get outline list
+                             PBLOB *blob,     //generic blob
+                             BOOL8 polygonal  //is blob polygonal?
+                            ) {
+  if (polygonal)
+    return blob->out_list ();
+  else
+    return (OUTLINE_LIST *) ((C_BLOB *) blob)->out_list ();
+}
+
+
+/**********************************************************************
+ *  goutline_bounding_box()
+ *
+ *  Return the bounding box of a generic outline.
+ **********************************************************************/
+
+BOX goutline_bounding_box(                   //Get bounding box
+                          OUTLINE *outline,  //generic outline
+                          BOOL8 polygonal    //is outline polygonal?
+                         ) {
+  if (polygonal)
+    return outline->bounding_box ();
+  else
+    return ((C_OUTLINE *) outline)->bounding_box ();
+}

ccstruct/genblob.h

@@ -0,0 +1,52 @@
+/**********************************************************************
+ * File:        genblob.h  (Formerly gblob.h)
+ * Description: Generic Blob processing routines
+ * Author:      Phil Cheatle
+ * Created:     Mon Nov 25 10:53:26 GMT 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           GENBLOB_H
+#define           GENBLOB_H
+
+#include          "polyblob.h"
+#include          "hosthplb.h"
+#include          "rect.h"
+#include          "notdll.h"
+
+int blob_comparator(                     //sort blobs
+                    const void *blob1p,  //ptr to ptr to blob1
+                    const void *blob2p   //ptr to ptr to blob2
+                   );
+int c_blob_comparator(                     //sort blobs
+                      const void *blob1p,  //ptr to ptr to blob1
+                      const void *blob2p   //ptr to ptr to blob2
+                     );
+BOX gblob_bounding_box(                 //Get bounding box
+                       PBLOB *blob,     //generic blob
+                       BOOL8 polygonal  //is blob polygonal?
+                      );
+void gblob_sort_list(                        //Sort a gblob list
+                     PBLOB_LIST *blob_list,  //generic blob list
+                     BOOL8 polygonal         //is list polygonal?
+                    );
+OUTLINE_LIST *gblob_out_list(                 //Get outline list
+                             PBLOB *blob,     //generic blob
+                             BOOL8 polygonal  //is blob polygonal?
+                            );
+BOX goutline_bounding_box(                   //Get bounding box
+                          OUTLINE *outline,  //generic outline
+                          BOOL8 polygonal    //is outline polygonal?
+                         );
+#endif

ccstruct/hpddef.h

@@ -0,0 +1,39 @@
+/**********************************************************************
+ * File:        hpddef.h
+ * Description: Defines for dll symbols for handpd.dll.
+ * Author:					Ray Smith
+ * Created:					Tue Apr 30 17:15:01 MDT 1996
+ *
+ * (C) Copyright 1996, Hewlett-Packard Co.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+//This file does NOT use the usual single inclusion code as it
+//is necessary to allow it to be executed every time it is included.
+//#ifndef                                                       HPDDEF_H
+//#define                                                       HPDDEF_H
+
+#undef DLLSYM
+#ifndef __IPEDLL
+#             define DLLSYM
+#else
+#             ifdef __BUILDING_HANDPD__
+#             define DLLSYM DLLEXPORT
+#             else
+#             define DLLSYM DLLIMPORT
+#             endif
+#endif
+#if defined(__CFM68K__) && !defined(__USING_STATIC_LIBS__)
+#             pragma import on
+#endif
+
+//#endif

ccstruct/hpdsizes.h

@@ -0,0 +1,8 @@
+#ifndef           HPDSIZES_H
+#define           HPDSIZES_H
+
+#define NUM_TEXT_ATTR   10
+#define NUM_BLOCK_ATTR   7
+#define MAXLENGTH 128
+#define NUM_BACKGROUNDS   8
+#endif

ccstruct/ipoints.h

@@ -0,0 +1,479 @@
+/**********************************************************************
+ * File:        ipoints.h  (Formerly icoords.h)
+ * Description: Inline functions for coords.h.
+ * Author:					Ray Smith
+ * Created:					Fri Jun 21 15:14:21 BST 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           IPOINTS_H
+#define           IPOINTS_H
+
+#include          <math.h>
+
+/**********************************************************************
+ * operator!
+ *
+ * Rotate an ICOORD 90 degrees anticlockwise.
+ **********************************************************************/
+
+inline ICOORD
+operator! (                      //rotate 90 deg anti
+const ICOORD & src               //thing to rotate
+) {
+  ICOORD result;                 //output
+
+  result.xcoord = -src.ycoord;
+  result.ycoord = src.xcoord;
+  return result;
+}
+
+
+/**********************************************************************
+ * operator-
+ *
+ * Unary minus of an ICOORD.
+ **********************************************************************/
+
+inline ICOORD
+operator- (                      //unary minus
+const ICOORD & src               //thing to minus
+) {
+  ICOORD result;                 //output
+
+  result.xcoord = -src.xcoord;
+  result.ycoord = -src.ycoord;
+  return result;
+}
+
+
+/**********************************************************************
+ * operator+
+ *
+ * Add 2 ICOORDS.
+ **********************************************************************/
+
+inline ICOORD
+operator+ (                      //sum vectors
+const ICOORD & op1,              //operands
+const ICOORD & op2) {
+  ICOORD sum;                    //result
+
+  sum.xcoord = op1.xcoord + op2.xcoord;
+  sum.ycoord = op1.ycoord + op2.ycoord;
+  return sum;
+}
+
+
+/**********************************************************************
+ * operator+=
+ *
+ * Add 2 ICOORDS.
+ **********************************************************************/
+
+inline ICOORD &
+operator+= (                     //sum vectors
+ICOORD & op1,                    //operands
+const ICOORD & op2) {
+  op1.xcoord += op2.xcoord;
+  op1.ycoord += op2.ycoord;
+  return op1;
+}
+
+
+/**********************************************************************
+ * operator-
+ *
+ * Subtract 2 ICOORDS.
+ **********************************************************************/
+
+inline ICOORD
+operator- (                      //subtract vectors
+const ICOORD & op1,              //operands
+const ICOORD & op2) {
+  ICOORD sum;                    //result
+
+  sum.xcoord = op1.xcoord - op2.xcoord;
+  sum.ycoord = op1.ycoord - op2.ycoord;
+  return sum;
+}
+
+
+/**********************************************************************
+ * operator-=
+ *
+ * Subtract 2 ICOORDS.
+ **********************************************************************/
+
+inline ICOORD &
+operator-= (                     //sum vectors
+ICOORD & op1,                    //operands
+const ICOORD & op2) {
+  op1.xcoord -= op2.xcoord;
+  op1.ycoord -= op2.ycoord;
+  return op1;
+}
+
+
+/**********************************************************************
+ * operator%
+ *
+ * Scalar product of 2 ICOORDS.
+ **********************************************************************/
+
+inline INT32
+operator% (                      //scalar product
+const ICOORD & op1,              //operands
+const ICOORD & op2) {
+  return op1.xcoord * op2.xcoord + op1.ycoord * op2.ycoord;
+}
+
+
+/**********************************************************************
+ * operator*
+ *
+ * Cross product of 2 ICOORDS.
+ **********************************************************************/
+
+inline INT32 operator *(                    //cross product
+                        const ICOORD &op1,  //operands
+                        const ICOORD &op2) {
+  return op1.xcoord * op2.ycoord - op1.ycoord * op2.xcoord;
+}
+
+
+/**********************************************************************
+ * operator*
+ *
+ * Scalar multiply of an ICOORD.
+ **********************************************************************/
+
+inline ICOORD operator *(                    //scalar multiply
+                         const ICOORD &op1,  //operands
+                         INT16 scale) {
+  ICOORD result;                 //output
+
+  result.xcoord = op1.xcoord * scale;
+  result.ycoord = op1.ycoord * scale;
+  return result;
+}
+
+
+inline ICOORD operator *(                   //scalar multiply
+                         INT16 scale,
+                         const ICOORD &op1  //operands
+                        ) {
+  ICOORD result;                 //output
+
+  result.xcoord = op1.xcoord * scale;
+  result.ycoord = op1.ycoord * scale;
+  return result;
+}
+
+
+/**********************************************************************
+ * operator*=
+ *
+ * Scalar multiply of an ICOORD.
+ **********************************************************************/
+
+inline ICOORD &
+operator*= (                     //scalar multiply
+ICOORD & op1,                    //operands
+INT16 scale) {
+  op1.xcoord *= scale;
+  op1.ycoord *= scale;
+  return op1;
+}
+
+
+/**********************************************************************
+ * operator/
+ *
+ * Scalar divide of an ICOORD.
+ **********************************************************************/
+
+inline ICOORD
+operator/ (                      //scalar divide
+const ICOORD & op1,              //operands
+INT16 scale) {
+  ICOORD result;                 //output
+
+  result.xcoord = op1.xcoord / scale;
+  result.ycoord = op1.ycoord / scale;
+  return result;
+}
+
+
+/**********************************************************************
+ * operator/=
+ *
+ * Scalar divide of an ICOORD.
+ **********************************************************************/
+
+inline ICOORD &
+operator/= (                     //scalar divide
+ICOORD & op1,                    //operands
+INT16 scale) {
+  op1.xcoord /= scale;
+  op1.ycoord /= scale;
+  return op1;
+}
+
+
+/**********************************************************************
+ * ICOORD::rotate
+ *
+ * Rotate an ICOORD by the given (normalized) (cos,sin) vector.
+ **********************************************************************/
+
+inline void ICOORD::rotate(  //rotate by vector
+                           const FCOORD& vec) {
+  INT16 tmp;
+
+  tmp = (INT16) floor (xcoord * vec.x () - ycoord * vec.y () + 0.5);
+  ycoord = (INT16) floor (ycoord * vec.x () + xcoord * vec.y () + 0.5);
+  xcoord = tmp;
+}
+
+
+/**********************************************************************
+ * operator!
+ *
+ * Rotate an FCOORD 90 degrees anticlockwise.
+ **********************************************************************/
+
+inline FCOORD
+operator! (                      //rotate 90 deg anti
+const FCOORD & src               //thing to rotate
+) {
+  FCOORD result;                 //output
+
+  result.xcoord = -src.ycoord;
+  result.ycoord = src.xcoord;
+  return result;
+}
+
+
+/**********************************************************************
+ * operator-
+ *
+ * Unary minus of an FCOORD.
+ **********************************************************************/
+
+inline FCOORD
+operator- (                      //unary minus
+const FCOORD & src               //thing to minus
+) {
+  FCOORD result;                 //output
+
+  result.xcoord = -src.xcoord;
+  result.ycoord = -src.ycoord;
+  return result;
+}
+
+
+/**********************************************************************
+ * operator+
+ *
+ * Add 2 FCOORDS.
+ **********************************************************************/
+
+inline FCOORD
+operator+ (                      //sum vectors
+const FCOORD & op1,              //operands
+const FCOORD & op2) {
+  FCOORD sum;                    //result
+
+  sum.xcoord = op1.xcoord + op2.xcoord;
+  sum.ycoord = op1.ycoord + op2.ycoord;
+  return sum;
+}
+
+
+/**********************************************************************
+ * operator+=
+ *
+ * Add 2 FCOORDS.
+ **********************************************************************/
+
+inline FCOORD &
+operator+= (                     //sum vectors
+FCOORD & op1,                    //operands
+const FCOORD & op2) {
+  op1.xcoord += op2.xcoord;
+  op1.ycoord += op2.ycoord;
+  return op1;
+}
+
+
+/**********************************************************************
+ * operator-
+ *
+ * Subtract 2 FCOORDS.
+ **********************************************************************/
+
+inline FCOORD
+operator- (                      //subtract vectors
+const FCOORD & op1,              //operands
+const FCOORD & op2) {
+  FCOORD sum;                    //result
+
+  sum.xcoord = op1.xcoord - op2.xcoord;
+  sum.ycoord = op1.ycoord - op2.ycoord;
+  return sum;
+}
+
+
+/**********************************************************************
+ * operator-=
+ *
+ * Subtract 2 FCOORDS.
+ **********************************************************************/
+
+inline FCOORD &
+operator-= (                     //sum vectors
+FCOORD & op1,                    //operands
+const FCOORD & op2) {
+  op1.xcoord -= op2.xcoord;
+  op1.ycoord -= op2.ycoord;
+  return op1;
+}
+
+
+/**********************************************************************
+ * operator%
+ *
+ * Scalar product of 2 FCOORDS.
+ **********************************************************************/
+
+inline float
+operator% (                      //scalar product
+const FCOORD & op1,              //operands
+const FCOORD & op2) {
+  return op1.xcoord * op2.xcoord + op1.ycoord * op2.ycoord;
+}
+
+
+/**********************************************************************
+ * operator*
+ *
+ * Cross product of 2 FCOORDS.
+ **********************************************************************/
+
+inline float operator *(                    //cross product
+                        const FCOORD &op1,  //operands
+                        const FCOORD &op2) {
+  return op1.xcoord * op2.ycoord - op1.ycoord * op2.xcoord;
+}
+
+
+/**********************************************************************
+ * operator*
+ *
+ * Scalar multiply of an FCOORD.
+ **********************************************************************/
+
+inline FCOORD operator *(                    //scalar multiply
+                         const FCOORD &op1,  //operands
+                         float scale) {
+  FCOORD result;                 //output
+
+  result.xcoord = op1.xcoord * scale;
+  result.ycoord = op1.ycoord * scale;
+  return result;
+}
+
+
+inline FCOORD operator *(                   //scalar multiply
+                         float scale,
+                         const FCOORD &op1  //operands
+                        ) {
+  FCOORD result;                 //output
+
+  result.xcoord = op1.xcoord * scale;
+  result.ycoord = op1.ycoord * scale;
+  return result;
+}
+
+
+/**********************************************************************
+ * operator*=
+ *
+ * Scalar multiply of an FCOORD.
+ **********************************************************************/
+
+inline FCOORD &
+operator*= (                     //scalar multiply
+FCOORD & op1,                    //operands
+float scale) {
+  op1.xcoord *= scale;
+  op1.ycoord *= scale;
+  return op1;
+}
+
+
+/**********************************************************************
+ * operator/
+ *
+ * Scalar divide of an FCOORD.
+ **********************************************************************/
+
+inline FCOORD
+operator/ (                      //scalar divide
+const FCOORD & op1,              //operands
+float scale) {
+  FCOORD result;                 //output
+
+  if (scale != 0) {
+    result.xcoord = op1.xcoord / scale;
+    result.ycoord = op1.ycoord / scale;
+  }
+  return result;
+}
+
+
+/**********************************************************************
+ * operator/=
+ *
+ * Scalar divide of an FCOORD.
+ **********************************************************************/
+
+inline FCOORD &
+operator/= (                     //scalar divide
+FCOORD & op1,                    //operands
+float scale) {
+  if (scale != 0) {
+    op1.xcoord /= scale;
+    op1.ycoord /= scale;
+  }
+  return op1;
+}
+
+
+/**********************************************************************
+ * rotate
+ *
+ * Rotate an FCOORD by the given (normalized) (cos,sin) vector.
+ **********************************************************************/
+
+inline void FCOORD::rotate(  //rotate by vector
+                           const FCOORD vec) {
+  float tmp;
+
+  tmp = xcoord * vec.x () - ycoord * vec.y ();
+  ycoord = ycoord * vec.x () + xcoord * vec.y ();
+  xcoord = tmp;
+}
+#endif

ccstruct/labls.cpp

@@ -0,0 +1,188 @@
+/**********************************************************************
+ * File:        labls.c  (Formerly labels.c)
+ * Description: Attribute definition tables
+ * Author:					Sheelagh Lloyd?
+ * Created:
+ *
+ * (C) Copyright 1993, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          "hpdsizes.h"
+#include          "labls.h"
+
+/******************************************************************************
+ * TEXT REGIONS
+ *****************************************************************************/
+DLLSYM INT32 tn[NUM_TEXT_ATTR] = {
+  3,                             //T_HORIZONTAL
+  4,                             //T_TEXT
+  2,                             //T_SERIF
+  2,                             //T_PROPORTIONAL
+  2,                             //T_NORMAL
+  2,                             //T_UPRIGHT
+  2,                             //T_SOLID
+  3,                             //T_BLACK
+  2,                             //T_NOTUNDER
+  2,                             //T_NOTDROP
+};
+
+DLLSYM char tlabel[NUM_TEXT_ATTR][4][MAXLENGTH] = { {
+                                 //T_HORIZONTAL
+    "Horizontal",
+    "Vertical",
+    "Skew",
+    ""
+  },
+  {                              //T_TEXT
+    "Text",
+    "Table",
+    "Form",
+    "Mixed"
+  },
+  {                              //T_SERIF
+    "Serif",
+    "Sans-serif",
+    "",
+    ""
+  },
+  {                              //T_PROPORTIONAL
+    "Proportional",
+    "Fixed pitch",
+    "",
+    ""
+  },
+  {                              //T_NORMAL
+    "Normal",
+    "Bold",
+    "",
+    ""
+  },
+  {                              //T_UPRIGHT
+    "Upright",
+    "Italic",
+    "",
+    ""
+  },
+  {                              //T_SOLID
+    "Solid",
+    "Outline",
+    "",
+    ""
+  },
+  {                              //T_BLACK
+    "Black",
+    "White",
+    "Coloured",
+    ""
+  },
+  {                              //T_NOTUNDER
+    "Not underlined",
+    "Underlined",
+    "",
+    ""
+  },
+  {                              //T_NOTDROP
+    "Not drop caps",
+    "Drop Caps",
+    "",
+    ""
+  }
+};
+
+DLLSYM INT32 bn[NUM_BLOCK_ATTR] = {
+  4,                             //G_MONOCHROME
+  2,                             //I_MONOCHROME
+  2,                             //I_SMOOTH
+  3,                             //R_SINGLE
+  3,                             //R_BLACK
+  3,                             //S_BLACK
+  2                              //W_TEXT
+};
+
+DLLSYM INT32 tvar[NUM_TEXT_ATTR];
+DLLSYM INT32 bvar[NUM_BLOCK_ATTR];
+DLLSYM char blabel[NUM_BLOCK_ATTR][4][MAXLENGTH] = { {
+                                 //G_MONOCHROME
+
+  /****************************************************************************
+   *  GRAPHICS
+   ***************************************************************************/
+    "Monochrome ",
+    "Two colour ",
+    "Spot colour",
+    "Multicolour"
+  },
+
+  /****************************************************************************
+   *  IMAGE
+   ***************************************************************************/
+  {                              //I_MONOCHROME
+    "Monochrome ",
+    "Colour     ",
+    "",
+    ""
+  },
+  {                              //I_SMOOTH
+    "Smooth     ",
+    "Grainy     ",
+    "",
+    ""
+  },
+
+  /****************************************************************************
+   *  RULES
+   ***************************************************************************/
+  {                              //R_SINGLE
+    "Single  ",
+    "Double  ",
+    "Multiple",
+    ""
+  },
+  {                              //R_BLACK
+    "Black   ",
+    "White   ",
+    "Coloured",
+    ""
+  },
+
+  /****************************************************************************
+   *  SCRIBBLE
+   ***************************************************************************/
+  {                              //S_BLACK
+    "Black   ",
+    "White   ",
+    "Coloured",
+    ""
+  },
+  /****************************************************************************
+   *  WEIRD
+   ***************************************************************************/
+  {                              //W_TEXT
+    "No text      ",
+    "Contains text",
+    "",
+    ""
+  }
+};
+
+DLLSYM char backlabel[NUM_BACKGROUNDS][MAXLENGTH] = {
+  "White",                       //B_WHITE
+  "Black",                       //B_BLACK
+  "Coloured",                    //B_COLOURED
+  "Textured",                    //B_TEXTURED
+  "Patterned",                   //B_PATTERNED
+  "Gradient fill",               //B_GRADIENTFILL
+  "Image",                       //B_IMAGE
+  "Text"                         //B_TEXT
+};

ccstruct/labls.h

@@ -0,0 +1,38 @@
+/**********************************************************************
+ * File:        labls.h  (Formerly labels.h)
+ * Description: Attribute definition tables
+ * Author:					Sheelagh Lloyd?
+ * Created:
+ *
+ * (C) Copyright 1993, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+#ifndef           LABLS_H
+#define           LABLS_H
+
+#include          "host.h"
+#include          "hpdsizes.h"
+
+#include          "hpddef.h"     //must be last (handpd.dll)
+
+extern DLLSYM INT32 tn[NUM_TEXT_ATTR];
+
+extern DLLSYM char tlabel[NUM_TEXT_ATTR][4][MAXLENGTH];
+
+extern DLLSYM INT32 bn[NUM_BLOCK_ATTR];
+
+extern DLLSYM INT32 tvar[NUM_TEXT_ATTR];
+extern DLLSYM INT32 bvar[NUM_BLOCK_ATTR];
+extern DLLSYM char blabel[NUM_BLOCK_ATTR][4][MAXLENGTH];
+
+extern DLLSYM char backlabel[NUM_BACKGROUNDS][MAXLENGTH];
+#endif

ccstruct/linlsq.cpp

@@ -0,0 +1,249 @@
+/**********************************************************************
+ * File:        linlsq.cpp  (Formerly llsq.c)
+ * Description: Linear Least squares fitting code.
+ * Author:		Ray Smith
+ * Created:		Thu Sep 12 08:44:51 BST 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          <stdio.h>
+#include          <math.h>
+#include          "errcode.h"
+#include          "linlsq.h"
+
+#ifndef __UNIX__
+#define M_PI        3.14159265359
+#endif
+
+const ERRCODE EMPTY_LLSQ = "Can't delete from an empty LLSQ";
+
+#define EXTERN
+
+EXTERN double_VAR (pdlsq_posdir_ratio, 4e-6, "Mult of dir to cf pos");
+EXTERN double_VAR (pdlsq_threshold_angleavg, 0.1666666,
+"Frac of pi for simple fit");
+
+/**********************************************************************
+ * LLSQ::clear
+ *
+ * Function to initialize a LLSQ.
+ **********************************************************************/
+
+void LLSQ::clear() {  //initialize
+  n = 0;                         //no elements
+  sigx = 0;                      //update accumulators
+  sigy = 0;
+  sigxx = 0;
+  sigxy = 0;
+  sigyy = 0;
+}
+
+
+/**********************************************************************
+ * LLSQ::add
+ *
+ * Add an element to the accumulator.
+ **********************************************************************/
+
+void LLSQ::add(           //add an element
+               double x,  //xcoord
+               double y   //ycoord
+              ) {
+  n++;                           //count elements
+  sigx += x;                     //update accumulators
+  sigy += y;
+  sigxx += x * x;
+  sigxy += x * y;
+  sigyy += y * y;
+}
+
+
+/**********************************************************************
+ * LLSQ::remove
+ *
+ * Delete an element from the acculuator.
+ **********************************************************************/
+
+void LLSQ::remove(           //delete an element
+                  double x,  //xcoord
+                  double y   //ycoord
+                 ) {
+  if (n <= 0)
+                                 //illegal
+    EMPTY_LLSQ.error ("LLSQ::remove", ABORT, NULL);
+  n--;                           //count elements
+  sigx -= x;                     //update accumulators
+  sigy -= y;
+  sigxx -= x * x;
+  sigxy -= x * y;
+  sigyy -= y * y;
+}
+
+
+/**********************************************************************
+ * LLSQ::m
+ *
+ * Return the gradient of the line fit.
+ **********************************************************************/
+
+double LLSQ::m() {  //get gradient
+  if (n > 1)
+    return (sigxy - sigx * sigy / n) / (sigxx - sigx * sigx / n);
+  else
+    return 0;                    //too little
+}
+
+
+/**********************************************************************
+ * LLSQ::c
+ *
+ * Return the constant of the line fit.
+ **********************************************************************/
+
+double LLSQ::c(          //get constant
+               double m  //gradient to fit with
+              ) {
+  if (n > 0)
+    return (sigy - m * sigx) / n;
+  else
+    return 0;                    //too little
+}
+
+
+/**********************************************************************
+ * LLSQ::rms
+ *
+ * Return the rms error of the fit.
+ **********************************************************************/
+
+double LLSQ::rms(           //get error
+                 double m,  //gradient to fit with
+                 double c   //constant to fit with
+                ) {
+  double error;                  //total error
+
+  if (n > 0) {
+    error =
+      sigyy + m * (m * sigxx + 2 * (c * sigx - sigxy)) + c * (n * c -
+      2 * sigy);
+    if (error >= 0)
+      error = sqrt (error / n);  //sqrt of mean
+    else
+      error = 0;
+  }
+  else
+    error = 0;                   //too little
+  return error;
+}
+
+
+/**********************************************************************
+ * LLSQ::spearman
+ *
+ * Return the spearman correlation coefficient.
+ **********************************************************************/
+
+double LLSQ::spearman() {  //get error
+  double error;                  //total error
+
+  if (n > 1) {
+    error = (sigxx - sigx * sigx / n) * (sigyy - sigy * sigy / n);
+    if (error > 0) {
+      error = (sigxy - sigx * sigy / n) / sqrt (error);
+    }
+    else
+      error = 1;
+  }
+  else
+    error = 1;                   //too little
+  return error;
+}
+
+
+/**********************************************************************
+ * PDLSQ::fit
+ *
+ * Return all the parameters of the fit to pos/dir.
+ * The return value is the rms error.
+ **********************************************************************/
+
+float PDLSQ::fit(                 //get fit
+                 DIR128 &ang,     //output angle
+                 float &sin_ang,  //r,theta parameterisation
+                 float &cos_ang,
+                 float &r) {
+  double a, b;                   //itermediates
+  double angle;                  //resulting angle
+  double avg_angle;              //simple average
+  double error;                  //total error
+  double sinx, cosx;             //return values
+
+  if (pos.n > 0) {
+    a = pos.sigxy - pos.sigx * pos.sigy / pos.n
+      + pdlsq_posdir_ratio * dir.sigxy;
+    b =
+      pos.sigxx - pos.sigyy + (pos.sigy * pos.sigy -
+      pos.sigx * pos.sigx) / pos.n +
+      pdlsq_posdir_ratio * (dir.sigxx - dir.sigyy);
+    if (dir.sigy != 0 || dir.sigx != 0)
+      avg_angle = atan2 (dir.sigy, dir.sigx);
+    else
+      avg_angle = 0;
+    if ((a != 0 || b != 0) && pos.n > 1)
+      angle = atan2 (2 * a, b) / 2;
+    else
+      angle = avg_angle;
+    error = avg_angle - angle;
+    if (error > M_PI / 2) {
+      error -= M_PI;
+      angle += M_PI;
+    }
+    if (error < -M_PI / 2) {
+      error += M_PI;
+      angle -= M_PI;
+    }
+    if (error > M_PI * pdlsq_threshold_angleavg
+      || error < -M_PI * pdlsq_threshold_angleavg)
+      angle = avg_angle;         //go simple
+                                 //convert direction
+    ang = (INT16) (angle * MODULUS / (2 * M_PI));
+    sinx = sin (angle);
+    cosx = cos (angle);
+    r = (sinx * pos.sigx - cosx * pos.sigy) / pos.n;
+    //              tprintf("x=%g, y=%g, xx=%g, xy=%g, yy=%g, a=%g, b=%g, ang=%g, r=%g\n",
+    //                      pos.sigx,pos.sigy,pos.sigxx,pos.sigxy,pos.sigyy,
+    //                      a,b,angle,r);
+    error = dir.sigxx * sinx * sinx + dir.sigyy * cosx * cosx
+      - 2 * dir.sigxy * sinx * cosx;
+    error *= pdlsq_posdir_ratio;
+    error += sinx * sinx * pos.sigxx + cosx * cosx * pos.sigyy
+      - 2 * sinx * cosx * pos.sigxy
+      - 2 * r * (sinx * pos.sigx - cosx * pos.sigy) + r * r * pos.n;
+    if (error >= 0)
+                                 //rms value
+        error = sqrt (error / pos.n);
+    else
+      error = 0;                 //-0
+    sin_ang = sinx;
+    cos_ang = cosx;
+  }
+  else {
+    sin_ang = 0.0f;
+    cos_ang = 0.0f;
+    ang = 0;
+    error = 0;                   //too little
+  }
+  return error;
+}

ccstruct/linlsq.h

@@ -0,0 +1,102 @@
+/**********************************************************************
+ * File:        linlsq.h  (Formerly llsq.h)
+ * Description: Linear Least squares fitting code.
+ * Author:		Ray Smith
+ * Created:		Thu Sep 12 08:44:51 BST 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           LINLSQ_H
+#define           LINLSQ_H
+
+#include          "points.h"
+#include          "mod128.h"
+#include          "varable.h"
+
+class LLSQ
+{
+  friend class PDLSQ;            //pos & direction
+
+  public:
+    LLSQ() {  //constructor
+      clear();  //set to zeros
+    }
+    void clear();  //initialize
+
+    void add(           //add element
+             double x,  //coords to add
+             double y);
+    void remove(           //delete element
+                double x,  //coords to delete
+                double y);
+    INT32 count() {  //no of elements
+      return n;
+    }
+
+    double m();  //get gradient
+    double c(            //get constant
+             double m);  //gradient
+    double rms(            //get error
+               double m,   //gradient
+               double c);  //constant
+    double spearman();  //get error
+
+  private:
+    INT32 n;                     //no of elements
+    double sigx;                 //sum of x
+    double sigy;                 //sum of y
+    double sigxx;                //sum x squared
+    double sigxy;                //sum of xy
+    double sigyy;                //sum y squared
+};
+
+class PDLSQ
+{
+  public:
+    PDLSQ() {  //constructor
+      clear();  //set to zeros
+    }
+    void clear() {  //initialize
+      pos.clear ();              //clear both
+      dir.clear ();
+    }
+
+    void add(                         //add element
+             const ICOORD &addpos,    //position of pt
+             const ICOORD &adddir) {  //dir of pt
+      pos.add (addpos.x (), addpos.y ());
+      dir.add (adddir.x (), adddir.y ());
+    }
+    void remove(                            //remove element
+                const ICOORD &removepos,    //position of pt
+                const ICOORD &removedir) {  //dir of pt
+      pos.remove (removepos.x (), removepos.y ());
+      dir.remove (removedir.x (), removedir.y ());
+    }
+    INT32 count() {  //no of elements
+      return pos.count ();
+    }
+
+    float fit(                 //get fit parameters
+              DIR128 &ang,     //output angle
+              float &sin_ang,  //output components
+              float &cos_ang,
+              float &r);
+
+  private:
+    LLSQ pos;                    //position
+    LLSQ dir;                    //directions
+};
+extern double_VAR_H (pdlsq_posdir_ratio, 0.4e-6, "Mult of dir to cf pos");
+#endif

ccstruct/lmedsq.cpp

@@ -0,0 +1,453 @@
+/**********************************************************************
+ * File:        lmedsq.cpp  (Formerly lms.c)
+ * Description: Code for the LMS class.
+ * Author:		Ray Smith
+ * Created:		Fri Aug  7 09:30:53 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          <stdlib.h>
+#include          "statistc.h"
+#include          "memry.h"
+#include          "statistc.h"
+#include          "lmedsq.h"
+
+#define EXTERN
+
+EXTERN INT_VAR (lms_line_trials, 12, "Number of linew fits to do");
+#define SEED1       0x1234       //default seeds
+#define SEED2       0x5678
+#define SEED3       0x9abc
+#define LMS_MAX_FAILURES  3
+
+#ifndef __UNIX__
+UINT32 nrand48(               //get random number
+               UINT16 *seeds  //seeds to use
+              ) {
+  static UINT32 seed = 0;        //only seed
+
+  if (seed == 0) {
+    seed = seeds[0] ^ (seeds[1] << 8) ^ (seeds[2] << 16);
+    srand(seed); 
+  }
+                                 //make 32 bit one
+  return rand () | (rand () << 16);
+}
+#endif
+
+/**********************************************************************
+ * LMS::LMS
+ *
+ * Construct a LMS class, given the max no of samples to be given
+ **********************************************************************/
+
+LMS::LMS (                       //constructor
+INT32 size                       //samplesize
+):samplesize (size) {
+  samplecount = 0;
+  a = 0;
+  m = 0.0f;
+  c = 0.0f;
+  samples = (FCOORD *) alloc_mem (size * sizeof (FCOORD));
+  errors = (float *) alloc_mem (size * sizeof (float));
+  line_error = 0.0f;
+  fitted = FALSE;
+}
+
+
+/**********************************************************************
+ * LMS::~LMS
+ *
+ * Destruct a LMS class.
+ **********************************************************************/
+
+LMS::~LMS (                      //constructor
+) {
+  free_mem(samples); 
+  free_mem(errors); 
+}
+
+
+/**********************************************************************
+ * LMS::clear
+ *
+ * Clear samples from array.
+ **********************************************************************/
+
+void LMS::clear() {  //clear sample
+  samplecount = 0;
+  fitted = FALSE;
+}
+
+
+/**********************************************************************
+ * LMS::add
+ *
+ * Add another sample. More than the constructed number will be ignored.
+ **********************************************************************/
+
+void LMS::add(               //add sample
+              FCOORD sample  //sample coords
+             ) {
+  if (samplecount < samplesize)
+                                 //save it
+    samples[samplecount++] = sample;
+  fitted = FALSE;
+}
+
+
+/**********************************************************************
+ * LMS::fit
+ *
+ * Fit a line to the given sample points.
+ **********************************************************************/
+
+void LMS::fit(               //fit sample
+              float &out_m,  //output line
+              float &out_c) {
+  INT32 index;                   //of median
+  INT32 trials;                  //no of medians
+  float test_m, test_c;          //candidate line
+  float test_error;              //error of test line
+
+  switch (samplecount) {
+    case 0:
+      m = 0.0f;                  //no info
+      c = 0.0f;
+      line_error = 0.0f;
+      break;
+
+    case 1:
+      m = 0.0f;
+      c = samples[0].y ();       //horiz thru pt
+      line_error = 0.0f;
+      break;
+
+    case 2:
+      if (samples[0].x () != samples[1].x ()) {
+        m = (samples[1].y () - samples[0].y ())
+          / (samples[1].x () - samples[0].x ());
+        c = samples[0].y () - m * samples[0].x ();
+      }
+      else {
+        m = 0.0f;
+        c = (samples[0].y () + samples[1].y ()) / 2;
+      }
+      line_error = 0.0f;
+      break;
+
+    default:
+      pick_line(m, c);  //use pts at random
+      compute_errors(m, c);  //from given line
+      index = choose_nth_item (samplecount / 2, errors, samplecount);
+      line_error = errors[index];
+      for (trials = 1; trials < lms_line_trials; trials++) {
+                                 //random again
+        pick_line(test_m, test_c); 
+        compute_errors(test_m, test_c); 
+        index = choose_nth_item (samplecount / 2, errors, samplecount);
+        test_error = errors[index];
+        if (test_error < line_error) {
+                                 //find least median
+          line_error = test_error;
+          m = test_m;
+          c = test_c;
+        }
+      }
+  }
+  fitted = TRUE;
+  out_m = m;
+  out_c = c;
+  a = 0;
+}
+
+
+/**********************************************************************
+ * LMS::fit_quadratic
+ *
+ * Fit a quadratic to the given sample points.
+ **********************************************************************/
+
+void LMS::fit_quadratic(                          //fit sample
+                        float outlier_threshold,  //min outlier size
+                        double &out_a,            //x squared
+                        float &out_b,             //output line
+                        float &out_c) {
+  INT32 trials;                  //no of medians
+  double test_a;
+  float test_b, test_c;          //candidate line
+  float test_error;              //error of test line
+
+  if (samplecount < 3) {
+    out_a = 0;
+    fit(out_b, out_c); 
+    return;
+  }
+  pick_quadratic(a, m, c); 
+  line_error = compute_quadratic_errors (outlier_threshold, a, m, c);
+  for (trials = 1; trials < lms_line_trials * 2; trials++) {
+    pick_quadratic(test_a, test_b, test_c); 
+    test_error = compute_quadratic_errors (outlier_threshold,
+      test_a, test_b, test_c);
+    if (test_error < line_error) {
+      line_error = test_error;   //find least median
+      a = test_a;
+      m = test_b;
+      c = test_c;
+    }
+  }
+  fitted = TRUE;
+  out_a = a;
+  out_b = m;
+  out_c = c;
+}
+
+
+/**********************************************************************
+ * LMS::constrained_fit
+ *
+ * Fit a line to the given sample points.
+ * The line must have the given gradient.
+ **********************************************************************/
+
+void LMS::constrained_fit(                //fit sample
+                          float fixed_m,  //forced gradient
+                          float &out_c) {
+  INT32 index;                   //of median
+  INT32 trials;                  //no of medians
+  float test_c;                  //candidate line
+  static UINT16 seeds[3] = { SEED1, SEED2, SEED3 };
+  //for nrand
+  float test_error;              //error of test line
+
+  m = fixed_m;
+  switch (samplecount) {
+    case 0:
+      c = 0.0f;
+      line_error = 0.0f;
+      break;
+
+    case 1:
+                                 //horiz thru pt
+      c = samples[0].y () - m * samples[0].x ();
+      line_error = 0.0f;
+      break;
+
+    case 2:
+      c = (samples[0].y () + samples[1].y ()
+        - m * (samples[0].x () + samples[1].x ())) / 2;
+      line_error = m * samples[0].x () + c - samples[0].y ();
+      line_error *= line_error;
+      break;
+
+    default:
+      index = (INT32) nrand48 (seeds) % samplecount;
+                                 //compute line
+      c = samples[index].y () - m * samples[index].x ();
+      compute_errors(m, c);  //from given line
+      index = choose_nth_item (samplecount / 2, errors, samplecount);
+      line_error = errors[index];
+      for (trials = 1; trials < lms_line_trials; trials++) {
+        index = (INT32) nrand48 (seeds) % samplecount;
+        test_c = samples[index].y () - m * samples[index].x ();
+        //compute line
+        compute_errors(m, test_c); 
+        index = choose_nth_item (samplecount / 2, errors, samplecount);
+        test_error = errors[index];
+        if (test_error < line_error) {
+                                 //find least median
+          line_error = test_error;
+          c = test_c;
+        }
+      }
+  }
+  fitted = TRUE;
+  out_c = c;
+  a = 0;
+}
+
+
+/**********************************************************************
+ * LMS::pick_line
+ *
+ * Fit a line to a random pair of sample points.
+ **********************************************************************/
+
+void LMS::pick_line(                //fit sample
+                    float &line_m,  //output gradient
+                    float &line_c) {
+  INT16 trial_count;             //no of attempts
+  static UINT16 seeds[3] = { SEED1, SEED2, SEED3 };
+  //for nrand
+  INT32 index1;                  //picked point
+  INT32 index2;                  //picked point
+
+  trial_count = 0;
+  do {
+    index1 = (INT32) nrand48 (seeds) % samplecount;
+    index2 = (INT32) nrand48 (seeds) % samplecount;
+    line_m = samples[index2].x () - samples[index1].x ();
+    trial_count++;
+  }
+  while (line_m == 0 && trial_count < LMS_MAX_FAILURES);
+  if (line_m == 0) {
+    line_c = (samples[index2].y () + samples[index1].y ()) / 2;
+  }
+  else {
+    line_m = (samples[index2].y () - samples[index1].y ()) / line_m;
+    line_c = samples[index1].y () - samples[index1].x () * line_m;
+  }
+}
+
+
+/**********************************************************************
+ * LMS::pick_quadratic
+ *
+ * Fit a quadratic to a random triplet of sample points.
+ **********************************************************************/
+
+void LMS::pick_quadratic(                 //fit sample
+                         double &line_a,  //x suaread
+                         float &line_m,   //output gradient
+                         float &line_c) {
+  INT16 trial_count;             //no of attempts
+  static UINT16 seeds[3] = { SEED1, SEED2, SEED3 };
+  //for nrand
+  INT32 index1;                  //picked point
+  INT32 index2;                  //picked point
+  INT32 index3;
+  FCOORD x1x2;                   //vector
+  FCOORD x1x3;
+  FCOORD x3x2;
+  double bottom;                 //of a
+
+  trial_count = 0;
+  do {
+    if (trial_count >= LMS_MAX_FAILURES - 1) {
+      index1 = 0;
+      index2 = samplecount / 2;
+      index3 = samplecount - 1;
+    }
+    else {
+      index1 = (INT32) nrand48 (seeds) % samplecount;
+      index2 = (INT32) nrand48 (seeds) % samplecount;
+      index3 = (INT32) nrand48 (seeds) % samplecount;
+    }
+    x1x2 = samples[index2] - samples[index1];
+    x1x3 = samples[index3] - samples[index1];
+    x3x2 = samples[index2] - samples[index3];
+    bottom = x1x2.x () * x1x3.x () * x3x2.x ();
+    trial_count++;
+  }
+  while (bottom == 0 && trial_count < LMS_MAX_FAILURES);
+  if (bottom == 0) {
+    line_a = 0;
+    pick_line(line_m, line_c); 
+  }
+  else {
+    line_a = x1x3 * x1x2 / bottom;
+    line_m = x1x2.y () - line_a * x1x2.x ()
+      * (samples[index2].x () + samples[index1].x ());
+    line_m /= x1x2.x ();
+    line_c = samples[index1].y () - samples[index1].x ()
+      * (samples[index1].x () * line_a + line_m);
+  }
+}
+
+
+/**********************************************************************
+ * LMS::compute_errors
+ *
+ * Compute the squared error from all the points.
+ **********************************************************************/
+
+void LMS::compute_errors(               //fit sample
+                         float line_m,  //input gradient
+                         float line_c) {
+  INT32 index;                   //picked point
+
+  for (index = 0; index < samplecount; index++) {
+    errors[index] =
+      line_m * samples[index].x () + line_c - samples[index].y ();
+    errors[index] *= errors[index];
+  }
+}
+
+
+/**********************************************************************
+ * LMS::compute_quadratic_errors
+ *
+ * Compute the squared error from all the points.
+ **********************************************************************/
+
+float LMS::compute_quadratic_errors(                          //fit sample
+                                    float outlier_threshold,  //min outlier
+                                    double line_a,
+                                    float line_m,             //input gradient
+                                    float line_c) {
+  INT32 outlier_count;           //total outliers
+  INT32 index;                   //picked point
+  INT32 error_count;             //no in total
+  double total_error;            //summed squares
+
+  total_error = 0;
+  outlier_count = 0;
+  error_count = 0;
+  for (index = 0; index < samplecount; index++) {
+    errors[error_count] = line_c + samples[index].x ()
+      * (line_m + samples[index].x () * line_a) - samples[index].y ();
+    errors[error_count] *= errors[error_count];
+    if (errors[error_count] > outlier_threshold) {
+      outlier_count++;
+      errors[samplecount - outlier_count] = errors[error_count];
+    }
+    else {
+      total_error += errors[error_count++];
+    }
+  }
+  if (outlier_count * 3 < error_count)
+    return total_error / error_count;
+  else {
+    index = choose_nth_item (outlier_count / 2,
+      errors + samplecount - outlier_count,
+      outlier_count);
+    //median outlier
+    return errors[samplecount - outlier_count + index];
+  }
+}
+
+
+/**********************************************************************
+ * LMS::plot
+ *
+ * Plot the fitted line of a LMS.
+ **********************************************************************/
+
+#ifndef GRAPHICS_DISABLED
+void LMS::plot(               //plot fit
+               WINDOW win,    //window
+               COLOUR colour  //colour to draw in
+              ) {
+  if (fitted) {
+    line_color_index(win, colour); 
+    move2d (win, samples[0].x (),
+      c + samples[0].x () * (m + samples[0].x () * a));
+    draw2d (win, samples[samplecount - 1].x (),
+      c + samples[samplecount - 1].x () * (m +
+      samples[samplecount -
+      1].x () * a));
+  }
+}
+#endif

ccstruct/lmedsq.h

@@ -0,0 +1,84 @@
+/**********************************************************************
+ * File:        lmedsq.h  (Formerly lms.h)
+ * Description: Code for the LMS class.
+ * Author:		Ray Smith
+ * Created:		Fri Aug  7 09:30:53 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           LMEDSQ_H
+#define           LMEDSQ_H
+
+#include          "points.h"
+#include          "varable.h"
+#include          "grphics.h"
+#include          "notdll.h"
+
+class LMS
+{
+  public:
+    LMS(              //constructor
+        INT32 size);  //no of samples
+    ~LMS ();                     //destructor
+    void clear();  //clear samples
+    void add(                 //add sample
+             FCOORD sample);  //sample coords
+    void fit(           //generate fit
+             float &m,  //output line
+             float &c);
+    void constrained_fit(                //fixed gradient
+                         float fixed_m,  //forced gradient
+                         float &out_c);  //output line
+    void fit_quadratic(                          //easy quadratic
+                       float outlier_threshold,  //min outlier
+                       double &a,                //x squared
+                       float &b,                 //x
+                       float &c);                //constant
+    void plot(                 //plot fit
+              WINDOW win,      //window
+              COLOUR colour);  //colour to draw in
+    float error() {  //get error
+      return fitted ? line_error : -1;
+    }
+
+  private:
+
+    void pick_line(           //random choice
+                   float &m,  //output line
+                   float &c);
+    void pick_quadratic(            //random choice
+                        double &a,  //output curve
+                        float &b,
+                        float &c);
+    void compute_errors(          //find errors
+                        float m,  //from line
+                        float c);
+                                 //find errors
+    float compute_quadratic_errors(float outlier_threshold,  //min outlier
+                                   double a,                 //from curve
+                                   float m,
+                                   float c);
+
+    BOOL8 fitted;                //line parts valid
+    INT32 samplesize;            //max samples
+    INT32 samplecount;           //current sample size
+    FCOORD *samples;             //array of samples
+    float *errors;               //error distances
+    double a;                    //x squared
+    float m;                     //line gradient
+    float c;
+    float line_error;            //error of fit
+};
+extern INT_VAR_H (lms_line_trials, 12, "Number of linew fits to do");
+#endif

ccstruct/mod128.cpp

@@ -0,0 +1,100 @@
+/**********************************************************************
+ * File:        mod128.c  (Formerly dir128.c)
+ * Description: Code to convert a DIR128 to an ICOORD.
+ * Author:					Ray Smith
+ * Created:					Tue Oct 22 11:56:09 BST 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include          "mfcpch.h"     //precompiled headers
+#include          "mod128.h"
+
+static INT16 idirtab[] = {
+  1000, 0, 998, 49, 995, 98, 989, 146,
+  980, 195, 970, 242, 956, 290, 941, 336,
+  923, 382, 903, 427, 881, 471, 857, 514,
+  831, 555, 803, 595, 773, 634, 740, 671,
+  707, 707, 671, 740, 634, 773, 595, 803,
+  555, 831, 514, 857, 471, 881, 427, 903,
+  382, 923, 336, 941, 290, 956, 242, 970,
+  195, 980, 146, 989, 98, 995, 49, 998,
+  0, 1000, -49, 998, -98, 995, -146, 989,
+  -195, 980, -242, 970, -290, 956, -336, 941,
+  -382, 923, -427, 903, -471, 881, -514, 857,
+  -555, 831, -595, 803, -634, 773, -671, 740,
+  -707, 707, -740, 671, -773, 634, -803, 595,
+  -831, 555, -857, 514, -881, 471, -903, 427,
+  -923, 382, -941, 336, -956, 290, -970, 242,
+  -980, 195, -989, 146, -995, 98, -998, 49,
+  -1000, 0, -998, -49, -995, -98, -989, -146,
+  -980, -195, -970, -242, -956, -290, -941, -336,
+  -923, -382, -903, -427, -881, -471, -857, -514,
+  -831, -555, -803, -595, -773, -634, -740, -671,
+  -707, -707, -671, -740, -634, -773, -595, -803,
+  -555, -831, -514, -857, -471, -881, -427, -903,
+  -382, -923, -336, -941, -290, -956, -242, -970,
+  -195, -980, -146, -989, -98, -995, -49, -998,
+  0, -1000, 49, -998, 98, -995, 146, -989,
+  195, -980, 242, -970, 290, -956, 336, -941,
+  382, -923, 427, -903, 471, -881, 514, -857,
+  555, -831, 595, -803, 634, -773, 671, -740,
+  707, -707, 740, -671, 773, -634, 803, -595,
+  831, -555, 857, -514, 881, -471, 903, -427,
+  923, -382, 941, -336, 956, -290, 970, -242,
+  980, -195, 989, -146, 995, -98, 998, -49
+};
+
+static ICOORD *dirtab = (ICOORD *) idirtab;
+
+/**********************************************************************
+ * DIR128::DIR128
+ *
+ * Quantize the direction of an FCOORD to make a DIR128.
+ **********************************************************************/
+
+DIR128::DIR128(                 //from fcoord
+               const FCOORD fc  //vector to quantize
+              ) {
+  int high, low, current;        //binary search
+
+  low = 0;
+  if (fc.y () == 0) {
+    if (fc.x () >= 0)
+      dir = 0;
+    else
+      dir = MODULUS / 2;
+    return;
+  }
+  high = MODULUS;
+  do {
+    current = (high + low) / 2;
+    if (dirtab[current] * fc >= 0)
+      low = current;
+    else
+      high = current;
+  }
+  while (high - low > 1);
+  dir = low;
+}
+
+
+/**********************************************************************
+ * dir_to_gradient
+ *
+ * Convert a direction to a vector.
+ **********************************************************************/
+
+ICOORD DIR128::vector() const {  //convert to vector
+  return dirtab[dir];            //easy really
+}

ccstruct/mod128.h

@@ -0,0 +1,85 @@
+/**********************************************************************
+ * File:        mod128.h  (Formerly dir128.h)
+ * Description: Header for class which implements modulo arithmetic.
+ * Author:					Ray Smith
+ * Created:					Tue Mar 26 17:48:13 GMT 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           MOD128_H
+#define           MOD128_H
+
+#include          "points.h"
+
+#define MODULUS       128        /*range of directions */
+#define DIRBITS       7          //no of bits used
+#define DIRSCALE      1000       //length of vector
+
+class DLLSYM DIR128
+{
+  public:
+    DIR128() { 
+    }                            //empty constructor
+
+    DIR128(                //constructor
+           INT16 value) {  //value to assign
+      value %= MODULUS;          //modulo arithmetic
+      if (value < 0)
+        value += MODULUS;        //done properly
+      dir = (INT8) value;
+    }
+    DIR128(const FCOORD fc);  //quantize vector
+
+    DIR128 & operator= (         //assign of INT16
+    INT16 value) {               //value to assign
+      value %= MODULUS;          //modulo arithmetic
+      if (value < 0)
+        value += MODULUS;        //done properly
+      dir = (INT8) value;
+      return *this;
+    }
+    INT8 operator- (             //subtraction
+      const DIR128 & minus) const//for signed result
+    {
+                                 //result
+      INT16 result = dir - minus.dir;
+
+      if (result > MODULUS / 2)
+        result -= MODULUS;       //get in range
+      else if (result < -MODULUS / 2)
+        result += MODULUS;
+      return (INT8) result;
+    }
+    DIR128 operator+ (           //addition
+      const DIR128 & add) const  //of itself
+    {
+      DIR128 result;             //sum
+
+      result = dir + add.dir;    //let = do the work
+      return result;
+    }
+    DIR128 & operator+= (        //same as +
+    const DIR128 & add) {
+      *this = dir + add.dir;     //let = do the work
+      return *this;
+    }
+    INT8 get_dir() const {  //access function
+      return dir;
+    }
+    ICOORD vector() const;  //turn to vector
+
+  private:
+    INT8 dir;                    //a direction
+};
+#endif

ccstruct/normalis.cpp

@@ -0,0 +1,176 @@
+/**********************************************************************
+ * File:        normalis.cpp  (Formerly denorm.c)
+ * Description: Code for the DENORM class.
+ * Author:		Ray Smith
+ * Created:		Thu Apr 23 09:22:43 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          "werd.h"
+#include          "normalis.h"
+
+/**********************************************************************
+ * DENORM::binary_search_segment
+ *
+ * Find the segment to use for the given x.
+ **********************************************************************/
+
+const DENORM_SEG *DENORM::binary_search_segment(float src_x) const {
+  int bottom, top, middle;       //binary search
+
+  bottom = 0;
+  top = segments;
+  do {
+    middle = (bottom + top) / 2;
+    if (segs[middle].xstart > src_x)
+      top = middle;
+    else
+      bottom = middle;
+  }
+  while (top - bottom > 1);
+  return &segs[bottom];
+}
+
+/**********************************************************************
+ * DENORM::scale_at_x
+ *
+ * Return scaling at a given (normalized) x coord.
+ **********************************************************************/
+
+float DENORM::scale_at_x(float src_x) const {  // In normalized coords.
+  if (segments != 0) {
+    const DENORM_SEG* seg = binary_search_segment(src_x);
+    if (seg->scale_factor > 0.0)
+      return seg->scale_factor;
+  }
+  return scale_factor;
+}
+
+/**********************************************************************
+ * DENORM::yshift_at_x
+ *
+ * Return yshift at a given (normalized) x coord.
+ **********************************************************************/
+
+float DENORM::yshift_at_x(float src_x) const {  // In normalized coords.
+  if (segments != 0) {
+    const DENORM_SEG* seg = binary_search_segment(src_x);
+    if (seg->ycoord == -MAX_INT32) {
+      if (base_is_row)
+        return source_row->base_line(x(src_x)/scale_at_x(src_x) + x_centre);
+      else
+        return m * x(src_x) + c;
+    } else {
+      return seg->ycoord;
+    }
+  }
+  return source_row->base_line (x(src_x)/scale_at_x(src_x) + x_centre);
+}
+
+/**********************************************************************
+ * DENORM::x
+ *
+ * Denormalise an x coordinate.
+ **********************************************************************/
+
+float DENORM::x(             //convert x coord
+                float src_x  //coord to convert
+               ) const {
+  return src_x / scale_at_x(src_x) + x_centre;
+}
+
+
+/**********************************************************************
+ * DENORM::y
+ *
+ * Denormalise a y coordinate.
+ **********************************************************************/
+
+float DENORM::y(                  //convert y coord
+                float src_y,      //coord to convert
+                float src_centre  //x location for base
+               ) const {
+  return (src_y - bln_baseline_offset) / scale_at_x(src_centre)
+    + yshift_at_x(src_centre);
+}
+
+
+DENORM::DENORM(float x,              //from same pieces
+               float scaling,
+               double line_m,        //default line
+               double line_c,
+               INT16 seg_count,      //no of segments
+               DENORM_SEG *seg_pts,  //actual segments
+               BOOL8 using_row,      //as baseline
+               ROW *src) {
+  x_centre = x;                  //just copy
+  scale_factor = scaling;
+  source_row = src;
+  if (seg_count > 0) {
+    segs = new DENORM_SEG[seg_count];
+    for (segments = 0; segments < seg_count; segments++) {
+      // It is possible, if infrequent that the segments may be out of order.
+      // since we are searching with a binary search, keep them in order.
+      if (segments == 0 || segs[segments - 1].xstart <=
+          seg_pts[segments].xstart) {
+        segs[segments] = seg_pts[segments];
+      } else {
+        int i;
+        for (i = 0; i < segments
+             && segs[segments - 1 - i].xstart > seg_pts[segments].xstart;
+             ++i) {
+          segs[segments - i ] = segs[segments - 1 - i];
+        }
+        segs[segments - i] = seg_pts[segments];
+      }
+    }
+  }
+  else {
+    segments = 0;
+    segs = NULL;
+  }
+  base_is_row = using_row;
+  m = line_m;
+  c = line_c;
+}
+
+
+DENORM::DENORM(const DENORM &src) { 
+  segments = 0;
+  segs = NULL;
+  *this = src;
+}
+
+
+DENORM & DENORM::operator= (const DENORM & src) {
+  x_centre = src.x_centre;
+  scale_factor = src.scale_factor;
+  source_row = src.source_row;
+  if (segments > 0)
+    delete[]segs;
+  if (src.segments > 0) {
+    segs = new DENORM_SEG[src.segments];
+    for (segments = 0; segments < src.segments; segments++)
+      segs[segments] = src.segs[segments];
+  }
+  else {
+    segments = 0;
+    segs = NULL;
+  }
+  base_is_row = src.base_is_row;
+  m = src.m;
+  c = src.c;
+  return *this;
+}

ccstruct/normalis.h

@@ -0,0 +1,108 @@
+/**********************************************************************
+ * File:        normalis.h  (Formerly denorm.h)
+ * Description: Code for the DENORM class.
+ * Author:		Ray Smith
+ * Created:		Thu Apr 23 09:22:43 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           NORMALIS_H
+#define           NORMALIS_H
+
+#include          <stdio.h>
+
+class ROW;                       //forward decl
+
+class DENORM_SEG
+{
+  public:
+    DENORM_SEG() { 
+    }                            //empty
+
+    INT32 xstart;                //start of segment
+    INT32 ycoord;                //y at segment
+    float scale_factor;          //for this segment
+};
+
+class DENORM
+{
+  public:
+    DENORM() {  //constructor
+      source_row = NULL;
+      x_centre = 0.0f;
+      scale_factor = 1.0f;
+      segments = 0;
+      segs = NULL;
+      base_is_row = TRUE;
+      m = c = 0;
+    }
+    DENORM(          //constructor
+           float x,  //from same pieces
+           float scaling,
+           ROW *src) {
+      x_centre = x;              //just copy
+      scale_factor = scaling;
+      source_row = src;
+      segments = 0;
+      segs = NULL;
+      base_is_row = TRUE;
+      m = c = 0;
+    }
+    DENORM(                      //constructor
+           float x,              //from same pieces
+           float scaling,
+           double line_m,        //default line //no of segments
+           double line_c,
+           INT16 seg_count,
+           DENORM_SEG *seg_pts,  //actual segments
+           BOOL8 using_row,      //as baseline
+           ROW *src);
+    DENORM(const DENORM &); 
+    DENORM & operator= (const DENORM &);
+    ~DENORM () {
+      if (segments > 0)
+        delete[]segs;
+    }
+
+    float origin() const {  //get x centre
+      return x_centre;
+    }
+    float scale() const {  //get scale
+      return scale_factor;
+    }
+    ROW *row() const {  //get row
+      return source_row;
+    }
+    float x(  //convert an xcoord
+            float src_x) const;
+    float y(                    //convert a ycoord
+            float src_y,        //coord to convert
+            float src_centre) const;  //normed x centre
+    float scale_at_x(  // Return scaling at this coord.
+            float src_x) const;
+    float yshift_at_x(  // Return yshift at this coord.
+            float src_x) const;
+
+  private:
+    const DENORM_SEG *binary_search_segment(float src_x) const;
+
+    BOOL8 base_is_row;           //using row baseline?
+    INT16 segments;              //no of segments
+    double c, m;                 //baseline
+    float x_centre;              //middle of word
+    float scale_factor;          //scaling
+    ROW *source_row;             //row it came from
+    DENORM_SEG *segs;            //array of segments
+};
+#endif

ccstruct/ocrblock.cpp

@@ -0,0 +1,368 @@
+/**********************************************************************
+ * File:        ocrblock.cpp  (Formerly block.c)
+ * Description: BLOCK member functions and iterator functions.
+ * Author:		Ray Smith
+ * Created:		Fri Mar 15 09:41:28 GMT 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          <stdlib.h>
+#include          "blckerr.h"
+#include          "ocrblock.h"
+#include          "tprintf.h"
+
+#define BLOCK_LABEL_HEIGHT  150  //char height of block id
+
+ELISTIZE_S (BLOCK)
+/**********************************************************************
+ * BLOCK::BLOCK
+ *
+ * Constructor for a simple rectangular block.
+ **********************************************************************/
+BLOCK::BLOCK (                   //rectangular block
+const char *name,                //filename
+BOOL8 prop,                      //proportional
+INT16 kern,                      //kerning
+INT16 space,                     //spacing
+INT16 xmin,                      //bottom left
+INT16 ymin, INT16 xmax,          //top right
+INT16 ymax):
+PDBLK (xmin, ymin, xmax, ymax),
+filename(name) {  //box(ICOORD(xmin,ymin),ICOORD(xmax,ymax))
+                                 //boundaries
+  ICOORDELT_IT left_it = &leftside;
+  ICOORDELT_IT right_it = &rightside;
+
+  proportional = prop;
+  kerning = kern;
+  spacing = space;
+  font_class = -1;               //not assigned
+  hand_block = NULL;
+  hand_poly = NULL;
+  left_it.set_to_list (&leftside);
+  right_it.set_to_list (&rightside);
+                                 //make default box
+  left_it.add_to_end (new ICOORDELT (xmin, ymin));
+  left_it.add_to_end (new ICOORDELT (xmin, ymax));
+  right_it.add_to_end (new ICOORDELT (xmax, ymin));
+  right_it.add_to_end (new ICOORDELT (xmax, ymax));
+}
+
+
+/**********************************************************************
+ * BLOCK::set_sides
+ *
+ * Sets left and right vertex lists
+ **********************************************************************/
+
+//void                                                          BLOCK::set_sides(                       //set vertex lists
+//ICOORDELT_LIST             *left,                        //left vertices
+//ICOORDELT_LIST             *right                        //right vertices
+//)
+//{
+//      ICOORDELT_IT                            left_it= &leftside;                             //boundaries
+//      ICOORDELT_IT                            right_it= &rightside;
+
+//      leftside.clear();
+//      left_it.move_to_first();
+//      left_it.add_list_before(left);
+//      rightside.clear();
+//      right_it.move_to_first();
+//      right_it.add_list_before(right);
+//}
+
+/**********************************************************************
+ * BLOCK::contains
+ *
+ * Return TRUE if the given point is within the block.
+ **********************************************************************/
+
+//BOOL8                                                         BLOCK::contains(                                        //test containment
+//ICOORD                                                        pt                                                                              //point to test
+//)
+//{
+//      BLOCK_RECT_IT                           it=this;                                                                //rectangle iterator
+//      ICOORD                                          bleft,tright;                                           //corners of rectangle
+
+//      for (it.start_block();!it.cycled_rects();it.forward())
+//      {
+//              it.bounding_box(bleft,tright);                                                  //get rectangle
+//              if (pt.x()>=bleft.x() && pt.x()<=tright.x()                     //inside rect
+//              && pt.y()>=bleft.y() && pt.y()<=tright.y())
+//                      return TRUE;                                                                                            //is inside
+//      }
+//      return FALSE;                                                                                                           //not inside
+//}
+
+/**********************************************************************
+ * BLOCK::move
+ *
+ * Reposition block
+ **********************************************************************/
+
+//void                                                     BLOCK::move(                                         // reposition block
+//const ICOORD                                  vec                                                                     // by vector
+//)
+//{
+//   ROW_IT                                             row_it( &rows );
+//      ICOORDELT_IT                            it( &leftside );
+
+//      for( row_it.mark_cycle_pt(); !row_it.cycled_list(); row_it.forward() )
+//         row_it.data()->move( vec );
+
+//      for( it.mark_cycle_pt(); !it.cycled_list(); it.forward() )
+//         *(it.data()) += vec;
+
+//   it.set_to_list( &rightside );
+
+//   for( it.mark_cycle_pt(); !it.cycled_list(); it.forward() )
+//         *(it.data()) += vec;
+
+//   box.move( vec );
+//}
+
+/**********************************************************************
+ * decreasing_top_order
+ *
+ * Sort Comparator: Return <0 if row1 top < row2 top
+ **********************************************************************/
+
+int decreasing_top_order(  //
+                         const void *row1,
+                         const void *row2) {
+  return (*(ROW **) row2)->bounding_box ().top () -
+    (*(ROW **) row1)->bounding_box ().top ();
+}
+
+
+/**********************************************************************
+ * BLOCK::sort_rows
+ *
+ * Order rows so that they are in order of decreasing Y coordinate
+ **********************************************************************/
+
+void BLOCK::sort_rows() {  // order on "top"
+  ROW_IT row_it(&rows); 
+
+  row_it.sort (decreasing_top_order);
+}
+
+
+/**********************************************************************
+ * BLOCK::compress
+ *
+ * Delete space between the rows. (And maybe one day, compress the rows)
+ * Fill space of block from top down, left aligning rows.
+ **********************************************************************/
+
+void BLOCK::compress() {  // squash it up
+  #define           ROW_SPACING 5
+
+  ROW_IT row_it(&rows); 
+  ROW *row;
+  ICOORD row_spacing (0, ROW_SPACING);
+
+  ICOORDELT_IT icoordelt_it;
+
+  sort_rows(); 
+
+  box = BOX (box.topleft (), box.topleft ());
+  box.move_bottom_edge (ROW_SPACING);
+  for (row_it.mark_cycle_pt (); !row_it.cycled_list (); row_it.forward ()) {
+    row = row_it.data ();
+    row->move (box.botleft () - row_spacing -
+      row->bounding_box ().topleft ());
+    box += row->bounding_box ();
+  }
+
+  leftside.clear ();
+  icoordelt_it.set_to_list (&leftside);
+  icoordelt_it.add_to_end (new ICOORDELT (box.left (), box.bottom ()));
+  icoordelt_it.add_to_end (new ICOORDELT (box.left (), box.top ()));
+  rightside.clear ();
+  icoordelt_it.set_to_list (&rightside);
+  icoordelt_it.add_to_end (new ICOORDELT (box.right (), box.bottom ()));
+  icoordelt_it.add_to_end (new ICOORDELT (box.right (), box.top ()));
+}
+
+
+/**********************************************************************
+ * BLOCK::check_pitch
+ *
+ * Check whether the block is fixed or prop, set the flag, and set
+ * the pitch if it is fixed.
+ **********************************************************************/
+
+void BLOCK::check_pitch() {  // check prop
+  //      tprintf("Missing FFT fixed pitch stuff!\n");
+  pitch = -1;
+}
+
+
+/**********************************************************************
+ * BLOCK::compress
+ *
+ * Compress and move in a single operation.
+ **********************************************************************/
+
+void BLOCK::compress(                  // squash it up
+                     const ICOORD vec  // and move
+                    ) {
+  box.move (vec);
+  compress(); 
+}
+
+
+/**********************************************************************
+ * BLOCK::print
+ *
+ * Print the info on a block
+ **********************************************************************/
+
+void BLOCK::print(            //print list of sides
+                  FILE *,     //file to print on
+                  BOOL8 dump  //print full detail
+                 ) {
+  ICOORDELT_IT it = &leftside;   //iterator
+
+  box.print ();
+  tprintf ("Proportional= %s\n", proportional ? "TRUE" : "FALSE");
+  tprintf ("Kerning= %d\n", kerning);
+  tprintf ("Spacing= %d\n", spacing);
+  tprintf ("Fixed_pitch=%d\n", pitch);
+  tprintf ("Filename= %s\n", filename.string ());
+
+  if (dump) {
+    tprintf ("Left side coords are:\n");
+    for (it.mark_cycle_pt (); !it.cycled_list (); it.forward ())
+      tprintf ("(%d,%d) ", it.data ()->x (), it.data ()->y ());
+    tprintf ("\n");
+    tprintf ("Right side coords are:\n");
+    it.set_to_list (&rightside);
+    for (it.mark_cycle_pt (); !it.cycled_list (); it.forward ())
+      tprintf ("(%d,%d) ", it.data ()->x (), it.data ()->y ());
+    tprintf ("\n");
+  }
+}
+
+
+/**********************************************************************
+ * BLOCK::plot
+ *
+ * Plot the outline of a block in the given colour.
+ **********************************************************************/
+
+//void                                                          BLOCK::plot(                                            //draw outline
+//WINDOW                                                        window,                                                         //window to draw in
+//INT32                                                         serial,                                                         //serial number
+//COLOUR                                                        colour                                                          //colour to draw in
+//)
+//{
+//      ICOORD                                          startpt;                                                                //start of outline
+//      ICOORD                                          endpt;                                                          //end of outline
+//      ICOORD                                          prevpt;                                                         //previous point
+//      ICOORDELT_IT                            it= &leftside;                                          //iterator
+//      char                                                    number[32];                                                     //block id
+
+//      line_color_index(window,colour);                                                                //set the colour
+//      text_color_index(window,colour);
+//      character_height(window,(float)BLOCK_LABEL_HEIGHT);
+//      text_font_index(window,6);
+
+//      if (!leftside.empty())
+//      {
+//              startpt= *(it.data());                                                                          //bottom left corner
+////            fprintf(stderr,"Block %d bottom left is (%d,%d)\n",
+////                    serial,startpt.x(),startpt.y());
+//              sprintf(number,"%d",serial);
+//              text2d(window,startpt.x(),startpt.y(),number,0,FALSE);
+
+//              move2d(window,startpt.x(),startpt.y());
+//              do
+//              {
+//                      prevpt= *(it.data());                                                                   //previous point
+//                      it.forward();                                                                                           //move to next point
+//                      draw2d(window,prevpt.x(),it.data()->y());                       //draw round corner
+//                      draw2d(window,it.data()->x(),it.data()->y());
+//              }
+//              while (!it.at_last());                                                                          //until end of list
+//              endpt= *(it.data());                                                                                    //end point
+
+//              move2d(window,startpt.x(),startpt.y());                         //other side of boundary
+//              it.set_to_list(&rightside);
+//              prevpt=startpt;
+//              for (it.mark_cycle_pt();!it.cycled_list();it.forward())
+//              {
+//                      draw2d(window,prevpt.x(),it.data()->y());                       //draw round corner
+//                      draw2d(window,it.data()->x(),it.data()->y());
+//                      prevpt= *(it.data());                                                                   //previous point
+//              }
+//              draw2d(window,endpt.x(),endpt.y());                                             //close boundary
+//              if (hand_block!=NULL)
+//                      hand_block->plot(window,colour,serial);
+//      }
+//}
+
+/**********************************************************************
+ * BLOCK::show
+ *
+ * Show the image corresponding to a block as its set of rectangles.
+ **********************************************************************/
+
+//void                                                          BLOCK::show(                                            //show image block
+//IMAGE                                                         *image,                                                         //image to show
+//WINDOW                                                        window                                                          //window to show in
+//)
+//{
+//      BLOCK_RECT_IT                           it=this;                                                                //rectangle iterator
+//      ICOORD                                          bleft,tright;                                           //corners of rectangle
+
+//      for (it.start_block();!it.cycled_rects();it.forward())
+//      {
+//              it.bounding_box(bleft,tright);                                                  //get rectangle
+////            fprintf(stderr,"Drawing a block with a bottom left of (%d,%d)\n",
+////                    bleft.x(),bleft.y());
+//              show_sub_image(image,bleft.x(),bleft.y(),
+//                      tright.x()-bleft.x(),tright.y()-bleft.y(),
+//                      window,bleft.x(),bleft.y());                                                    //show it
+//      }
+//}
+
+/**********************************************************************
+ * BLOCK::operator=
+ *
+ * Assignment - duplicate the block structure, but with an EMPTY row list.
+ **********************************************************************/
+
+BLOCK & BLOCK::operator= (       //assignment
+const BLOCK & source             //from this
+) {
+  this->ELIST_LINK::operator= (source);
+  this->PDBLK::operator= (source);
+  proportional = source.proportional;
+  kerning = source.kerning;
+  spacing = source.spacing;
+  filename = source.filename;    //STRINGs assign ok
+  if (!rows.empty ())
+    rows.clear ();
+  //   if ( !leftside.empty() )
+  //              leftside.clear();
+  // if ( !rightside.empty() )
+  //              rightside.clear();
+  //   leftside.deep_copy( &source.leftside );
+  //   rightside.deep_copy( &source.rightside );
+  //      box=source.box;
+  return *this;
+}

ccstruct/ocrblock.h

@@ -0,0 +1,228 @@
+/**********************************************************************
+ * File:        ocrblock.h  (Formerly block.h)
+ * Description: Page block class definition.
+ * Author:		Ray Smith
+ * Created:		Thu Mar 14 17:32:01 GMT 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           OCRBLOCK_H
+#define           OCRBLOCK_H
+
+#include          "img.h"
+#include          "ocrrow.h"
+#include          "pageblk.h"
+#include          "pdblock.h"
+
+class BLOCK;                     //forward decl
+
+ELISTIZEH_S (BLOCK)
+class BLOCK:public ELIST_LINK, public PDBLK
+//page block
+{
+  friend class BLOCK_RECT_IT;    //block iterator
+
+                                 //block label
+  friend void scan_hpd_blocks(const char *name,
+                              PAGE_BLOCK_LIST *page_blocks,  //head of full pag
+                              INT32 &block_no,               //no of blocks
+                              BLOCK_IT *block_it);
+  friend BOOL8 read_vec_file(              //read uscan output
+                             STRING name,  //basename of file
+                             INT32 xsize,  //page size //output list
+                             INT32 ysize,
+                             BLOCK_LIST *blocks);
+  friend BOOL8 read_pd_file(              //read uscan output
+                            STRING name,  //basename of file
+                            INT32 xsize,  //page size //output list
+                            INT32 ysize,
+                            BLOCK_LIST *blocks);
+
+  public:
+    BLOCK() {  //empty constructor
+      hand_block = NULL;
+      hand_poly = NULL;
+    }
+    BLOCK(                   //simple constructor
+          const char *name,  //filename
+          BOOL8 prop,        //proportional
+          INT16 kern,        //kerning
+          INT16 space,       //spacing
+          INT16 xmin,        //bottom left
+          INT16 ymin,
+          INT16 xmax,        //top right
+          INT16 ymax);
+
+    //      void                                            set_sides(                                                 //set vertex lists
+    //              ICOORDELT_LIST       *left,                        //list of left vertices
+    //              ICOORDELT_LIST       *right);                      //list of right vertices
+
+    ~BLOCK () {                  //destructor
+    }
+
+    void set_stats(                   //set space size etc.
+                   BOOL8 prop,        //proportional
+                   INT16 kern,        //inter char size
+                   INT16 space,       //inter word size
+                   INT16 ch_pitch) {  //pitch if fixed
+      proportional = prop;
+      kerning = (INT8) kern;
+      spacing = space;
+      pitch = ch_pitch;
+    }
+    void set_xheight(  //set char size
+                     INT32 height) {
+      xheight = height;
+    }
+    void set_font_class(  //set font class
+                        INT16 font) {
+      font_class = font;
+    }
+    //      TEXT_REGION*                            text_region()
+    //      {
+    //              return hand_block;
+    //      }
+    //      POLY_BLOCK*                                     poly_block()
+    //      {
+    //              return hand_poly;
+    //      }
+    BOOL8 prop() const {  //return proportional
+      return proportional;
+    }
+    INT32 fixed_pitch() const {  //return pitch
+      return pitch;
+    }
+    INT16 kern() const {  //return kerning
+      return kerning;
+    }
+    INT16 font() const {  //return font class
+      return font_class;
+    }
+    INT16 space() const {  //return spacing
+      return spacing;
+    }
+    const char *name() const {  //return filename
+      return filename.string ();
+    }
+    INT32 x_height() const {  //return xheight
+      return xheight;
+    }
+    ROW_LIST *row_list() {  //get rows
+      return &rows;
+    }
+    C_BLOB_LIST *blob_list() {  //get blobs
+      return &c_blobs;
+    }
+    C_BLOB_LIST *reject_blobs() { 
+      return &rej_blobs;
+    }
+    //      void                                                    bounding_box(                                           //get box
+    //              ICOORD&                                 bottom_left,                                            //bottom left
+    //              ICOORD&                                 top_right) const                                        //topright
+    //      {
+    //              bottom_left=box.botleft();
+    //              top_right=box.topright();
+    //      }
+    //      const BOX&                                      bounding_box() const                            //get real box
+    //      {
+    //              return box;
+    //      }
+
+    //      BOOL8                                                   contains(                                                       //is pt inside block
+    //              ICOORD                                  pt);
+
+    //      void                                                    move(                                                                   // reposition block
+    //         const ICOORD                 vec);                                                                   // by vector
+
+    void sort_rows();  //decreasing y order
+
+    void compress();  //shrink white space
+
+    void check_pitch();  //check proportional
+
+    void compress(                    //shrink white space
+                  const ICOORD vec);  //and move by vector
+
+    void print(              //print summary/table
+               FILE *fp,     //file to print on
+               BOOL8 dump);  //dump whole table
+
+    //      void                                                    plot(                                                                   //draw histogram
+    //              WINDOW                                  window,                                                         //window to draw in
+    //              INT32                                           serial,                                                         //serial number
+    //              COLOUR                                  colour);                                                                //colour to draw in
+
+    //      void                                                    show(                                                                   //show image
+    //              IMAGE                                           *image,                                                         //image to show
+    //              WINDOW                                  window);                                                                //window to show in
+
+    void prep_serialise() {  //set ptrs to counts
+      filename.prep_serialise ();
+      rows.prep_serialise ();
+      c_blobs.prep_serialise ();
+      rej_blobs.prep_serialise ();
+      leftside.prep_serialise ();
+      rightside.prep_serialise ();
+    }
+
+    void dump(  //write external bits
+              FILE *f) {
+      filename.dump (f);
+      rows.dump (f);
+      c_blobs.dump (f);
+      rej_blobs.dump (f);
+      leftside.dump (f);
+      rightside.dump (f);
+      if (hand_block != NULL)
+        hand_block->serialise (f);
+    }
+
+    void de_dump(  //read external bits
+                 FILE *f) {
+      filename.de_dump (f);
+      rows.de_dump (f);
+      c_blobs.de_dump (f);
+      rej_blobs.de_dump (f);
+      leftside.de_dump (f);
+      rightside.de_dump (f);
+      if (hand_block != NULL)
+        hand_block = TEXT_REGION::de_serialise (f);
+    }
+
+                                 //assignment
+    make_serialise (BLOCK) BLOCK & operator= (
+      const BLOCK & source);     //from this
+
+  private:
+    BOOL8 proportional;          //proportional
+    INT8 kerning;                //inter blob gap
+    INT16 spacing;               //inter word gap
+    INT16 pitch;                 //pitch of non-props
+    INT16 font_class;            //correct font class
+    INT32 xheight;               //height of chars
+    STRING filename;             //name of block
+    //      TEXT_REGION*                            hand_block;                                                     //if it exists
+    //      POLY_BLOCK*                                     hand_poly;                                                      //wierd as well
+    ROW_LIST rows;               //rows in block
+    C_BLOB_LIST c_blobs;         //before textord
+    C_BLOB_LIST rej_blobs;       //duff stuff
+    //      ICOORDELT_LIST                          leftside;                                                       //left side vertices
+    //      ICOORDELT_LIST                          rightside;                                                      //right side vertices
+    //      BOX                                                     box;                                                                    //bounding box
+};
+
+int decreasing_top_order(  //
+                         const void *row1,
+                         const void *row2);
+#endif

ccstruct/ocrrow.cpp

@@ -0,0 +1,216 @@
+/**********************************************************************
+ * File:        ocrrow.cpp  (Formerly row.c)
+ * Description: Code for the ROW class.
+ * Author:					Ray Smith
+ * Created:					Tue Oct 08 15:58:04 BST 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#include "mfcpch.h"
+#include          "ocrrow.h"
+#include          "blobbox.h"
+
+ELISTIZE_S (ROW)
+/**********************************************************************
+ * ROW::ROW
+ *
+ * Constructor to build a ROW. Only the stats stuff are given here.
+ * The words are added directly.
+ **********************************************************************/
+ROW::ROW (                       //constructor
+INT32 spline_size,               //no of segments
+INT32 * xstarts,                 //segment boundaries
+double *coeffs,                  //coefficients
+float x_height,                  //line height
+float ascenders,                 //ascender size
+float descenders,                //descender drop
+INT16 kern,                      //char gap
+INT16 space                      //word gap
+):
+baseline(spline_size, xstarts, coeffs) { 
+  kerning = kern;                //just store stuff
+  spacing = space;
+  xheight = x_height;
+  ascrise = ascenders;
+  descdrop = descenders;
+}
+
+
+/**********************************************************************
+ * ROW::ROW
+ *
+ * Constructor to build a ROW. Only the stats stuff are given here.
+ * The words are added directly.
+ **********************************************************************/
+
+ROW::ROW(                 //constructor
+         TO_ROW *to_row,  //source row
+         INT16 kern,      //char gap
+         INT16 space      //word gap
+        ) {
+  kerning = kern;                //just store stuff
+  spacing = space;
+  xheight = to_row->xheight;
+  ascrise = to_row->ascrise;
+  descdrop = to_row->descdrop;
+  baseline = to_row->baseline;
+}
+
+
+/**********************************************************************
+ * ROW::recalc_bounding_box
+ *
+ * Set the bounding box correctly
+ **********************************************************************/
+
+void ROW::recalc_bounding_box() {  //recalculate BB
+  WERD *word;                    //current word
+  WERD_IT it = &words;           //words of ROW
+  INT16 left;                    //of word
+  INT16 prev_left;               //old left
+
+  if (!it.empty ()) {
+    word = it.data ();
+    prev_left = word->bounding_box ().left ();
+    it.forward ();
+    while (!it.at_first ()) {
+      word = it.data ();
+      left = word->bounding_box ().left ();
+      if (left < prev_left) {
+        it.move_to_first ();
+                                 //words in BB order
+        it.sort (word_comparator);
+        break;
+      }
+      prev_left = left;
+      it.forward ();
+    }
+  }
+  for (it.mark_cycle_pt (); !it.cycled_list (); it.forward ()) {
+    word = it.data ();
+    if (it.at_first ())
+      word->set_flag (W_BOL, TRUE);
+    else
+                                 //not start of line
+      word->set_flag (W_BOL, FALSE);
+    if (it.at_last ())
+      word->set_flag (W_EOL, TRUE);
+    else
+                                 //not end of line
+      word->set_flag (W_EOL, FALSE);
+                                 //extend BB as reqd
+    bound_box += word->bounding_box ();
+  }
+}
+
+
+/**********************************************************************
+ * ROW::move
+ *
+ * Reposition row by vector
+ **********************************************************************/
+
+void ROW::move(                  // reposition row
+               const ICOORD vec  // by vector
+              ) {
+  WERD_IT it(&words);  // word iterator
+
+  for (it.mark_cycle_pt (); !it.cycled_list (); it.forward ())
+    it.data ()->move (vec);
+
+  bound_box.move (vec);
+  baseline.move (vec);
+}
+
+
+/**********************************************************************
+ * ROW::print
+ *
+ * Display members
+ **********************************************************************/
+
+void ROW::print(          //print
+                FILE *fp  //file to print on
+               ) {
+  tprintf ("Kerning= %d\n", kerning);
+  tprintf ("Spacing= %d\n", spacing);
+  bound_box.print ();
+  tprintf ("Xheight= %f\n", xheight);
+  tprintf ("Ascrise= %f\n", ascrise);
+  tprintf ("Descdrop= %f\n", descdrop);
+}
+
+
+/**********************************************************************
+ * ROW::plot
+ *
+ * Draw the ROW in the given colour.
+ **********************************************************************/
+
+#ifndef GRAPHICS_DISABLED
+void ROW::plot(                //draw it
+               WINDOW window,  //window to draw in
+               COLOUR colour   //colour to draw in
+              ) {
+  WERD *word;                    //current word
+  WERD_IT it = &words;           //words of ROW
+
+  for (it.mark_cycle_pt (); !it.cycled_list (); it.forward ()) {
+    word = it.data ();
+    word->plot (window, colour); //all in one colour
+  }
+}
+#endif
+
+/**********************************************************************
+ * ROW::plot
+ *
+ * Draw the ROW in rainbow colours.
+ **********************************************************************/
+
+#ifndef GRAPHICS_DISABLED
+void ROW::plot(               //draw it
+               WINDOW window  //window to draw in
+              ) {
+  WERD *word;                    //current word
+  WERD_IT it = &words;           //words of ROW
+
+  for (it.mark_cycle_pt (); !it.cycled_list (); it.forward ()) {
+    word = it.data ();
+    word->plot (window);         //in rainbow colours
+  }
+}
+#endif
+
+/**********************************************************************
+ * ROW::operator=
+ *
+ * Assign rows by duplicating the row structure but NOT the WERDLIST
+ **********************************************************************/
+
+ROW & ROW::operator= (           //assignment
+const ROW & source               //from this
+) {
+  this->ELIST_LINK::operator= (source);
+  kerning = source.kerning;
+  spacing = source.spacing;
+  xheight = source.xheight;
+  ascrise = source.ascrise;
+  descdrop = source.descdrop;
+  if (!words.empty ())
+    words.clear ();
+  baseline = source.baseline;    //QSPLINES must do =
+  bound_box = source.bound_box;
+  return *this;
+}

ccstruct/ocrrow.h

@@ -0,0 +1,133 @@
+/**********************************************************************
+ * File:        ocrrow.h  (Formerly row.h)
+ * Description: Code for the ROW class.
+ * Author:		Ray Smith
+ * Created:		Tue Oct 08 15:58:04 BST 1991
+ *
+ * (C) Copyright 1991, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+
+#ifndef           OCRROW_H
+#define           OCRROW_H
+
+#include          <stdio.h>
+#include          "quspline.h"
+#include          "werd.h"
+
+class TO_ROW;
+
+class ROW:public ELIST_LINK
+{
+  friend void tweak_row_baseline(ROW *); 
+  public:
+    ROW() { 
+    }                            //empty constructor
+    ROW(                    //constructor
+        INT32 spline_size,  //no of segments
+        INT32 *xstarts,     //segment boundaries
+        double *coeffs,     //coefficients //ascender size
+        float x_height,
+        float ascenders,
+        float descenders,   //descender size
+        INT16 kern,         //char gap
+        INT16 space);       //word gap
+    ROW(               //constructor
+        TO_ROW *row,   //textord row
+        INT16 kern,    //char gap
+        INT16 space);  //word gap
+
+    WERD_LIST *word_list() {  //get words
+      return &words;
+    }
+
+    float base_line(                     //compute baseline
+                    float xpos) const {  //at the position
+                                 //get spline value
+      return (float) baseline.y (xpos);
+    }
+    float x_height() const {  //return x height
+      return xheight;
+    }
+    INT32 kern() const {  //return kerning
+      return kerning;
+    }
+    INT32 space() const {  //return spacing
+      return spacing;
+    }
+    float ascenders() const {  //return size
+      return ascrise;
+    }
+    float descenders() const {  //return size
+      return descdrop;
+    }
+    BOX bounding_box() const {  //return bounding box
+      return bound_box;
+    }
+
+    void recalc_bounding_box();  //recalculate BB
+
+    void move(                    // reposition row
+              const ICOORD vec);  // by vector
+
+    void print(            //print
+               FILE *fp);  //file to print on
+
+    void plot(                 //draw one
+              WINDOW window,   //window to draw in
+              COLOUR colour);  //uniform colour
+    void plot(                 //draw one
+              WINDOW window);  //in rainbow colours
+
+#ifndef GRAPHICS_DISABLED
+    void plot_baseline(                  //draw the baseline
+                       WINDOW window,    //window to draw in
+                       COLOUR colour) {  //colour to draw
+                                 //draw it
+      baseline.plot (window, colour);
+    }
+#endif
+
+    void prep_serialise() {  //set ptrs to counts
+      words.prep_serialise ();
+      baseline.prep_serialise ();
+    }
+
+    void dump(  //write external bits
+              FILE *f) {
+      words.dump (f);
+      baseline.dump (f);
+    }
+
+    void de_dump(  //read external bits
+                 FILE *f) {
+      words.de_dump (f);
+      baseline.de_dump (f);
+    }
+
+                                 //assignment
+    make_serialise (ROW) ROW & operator= (
+      const ROW & source);       //from this
+
+  private:
+    INT32 kerning;               //inter char gap
+    INT32 spacing;               //inter word gap
+    BOX bound_box;               //bounding box
+    float xheight;               //height of line
+    float ascrise;               //size of ascenders
+    float descdrop;              //-size of descenders
+    WERD_LIST words;             //words
+    QSPLINE baseline;            //baseline spline
+};
+
+ELISTIZEH_S (ROW)
+#endif

ccstruct/pageblk.cpp

@@ -0,0 +1,879 @@
+#include "mfcpch.h"
+#include          "pageblk.h"
+#include    <stdio.h>
+#include          <ctype.h>
+#include          <math.h>
+#ifdef __UNIX__
+#include          <unistd.h>
+#else
+#include          <io.h>
+#endif
+
+#include          "hpddef.h"     //must be last (handpd.dll)
+
+#define G_START 0
+#define I_START 1
+#define R_START 3
+#define S_START 5
+
+extern char blabel[NUM_BLOCK_ATTR][4][MAXLENGTH];
+extern char backlabel[NUM_BACKGROUNDS][MAXLENGTH];
+
+ELISTIZE_S (PAGE_BLOCK)
+void PAGE_BLOCK::pb_delete() { 
+  switch (pb_type) {
+    case PB_TEXT:
+      delete ((TEXT_BLOCK *) this);
+      break;
+    case PB_GRAPHICS:
+      delete ((GRAPHICS_BLOCK *) this);
+      break;
+    case PB_IMAGE:
+      delete ((IMAGE_BLOCK *) this);
+      break;
+    case PB_RULES:
+      delete ((RULE_BLOCK *) this);
+      break;
+    case PB_SCRIBBLE:
+      delete ((SCRIBBLE_BLOCK *) this);
+      break;
+    case PB_WEIRD:
+      delete ((WEIRD_BLOCK *) this);
+      break;
+    default:
+      break;
+  }
+}
+
+
+#define QUOTE_IT( parm ) #parm
+
+void PAGE_BLOCK::serialise(FILE *f) { 
+
+  if (fwrite (&pb_type, sizeof (PB_TYPE), 1, f) != 1)
+    WRITEFAILED.error (QUOTE_IT (PAGE_BLOCK::serialise), ABORT, NULL);
+  switch (pb_type) {
+    case PB_TEXT:
+      ((TEXT_BLOCK *) this)->serialise (f);
+      break;
+    case PB_GRAPHICS:
+      ((GRAPHICS_BLOCK *) this)->serialise (f);
+      break;
+    case PB_RULES:
+      ((RULE_BLOCK *) this)->serialise (f);
+      break;
+    case PB_IMAGE:
+      ((IMAGE_BLOCK *) this)->serialise (f);
+      break;
+    case PB_SCRIBBLE:
+      ((SCRIBBLE_BLOCK *) this)->serialise (f);
+      break;
+    case PB_WEIRD:
+      ((WEIRD_BLOCK *) this)->serialise (f);
+      break;
+    default:
+      break;
+  }
+}
+
+
+PAGE_BLOCK *PAGE_BLOCK::de_serialise(FILE *f) { 
+  PB_TYPE type;
+  TEXT_BLOCK *tblock;
+  GRAPHICS_BLOCK *gblock;
+  RULE_BLOCK *rblock;
+  IMAGE_BLOCK *iblock;
+  SCRIBBLE_BLOCK *sblock;
+  WEIRD_BLOCK *wblock;
+
+  if (fread ((void *) &type, sizeof (PB_TYPE), 1, f) != 1)
+    WRITEFAILED.error (QUOTE_IT (PAGE_BLOCK::serialise), ABORT, NULL);
+  switch (type) {
+    case PB_TEXT:
+      tblock = (TEXT_BLOCK *) alloc_struct (sizeof (TEXT_BLOCK));
+      return tblock->de_serialise (f);
+    case PB_GRAPHICS:
+      gblock = (GRAPHICS_BLOCK *) alloc_struct (sizeof (GRAPHICS_BLOCK));
+      return gblock->de_serialise (f);
+    case PB_RULES:
+      rblock = (RULE_BLOCK *) alloc_struct (sizeof (RULE_BLOCK));
+      return rblock->de_serialise (f);
+    case PB_IMAGE:
+      iblock = (IMAGE_BLOCK *) alloc_struct (sizeof (IMAGE_BLOCK));
+      return iblock->de_serialise (f);
+    case PB_SCRIBBLE:
+      sblock = (SCRIBBLE_BLOCK *) alloc_struct (sizeof (SCRIBBLE_BLOCK));
+      return sblock->de_serialise (f);
+    case PB_WEIRD:
+      wblock = (WEIRD_BLOCK *) alloc_struct (sizeof (SCRIBBLE_BLOCK));
+      return wblock->de_serialise (f);
+    default:
+      return NULL;
+  }
+}
+
+
+/**********************************************************************
+ * PAGE_BLOCK::serialise_asc()  Convert to ascii file.
+ *
+ **********************************************************************/
+
+void PAGE_BLOCK::serialise_asc(         //convert to ascii
+                               FILE *f  //file to use
+                              ) {
+  serialise_INT32(f, pb_type); 
+  switch (pb_type) {
+    case PB_TEXT:
+      ((TEXT_BLOCK *) this)->serialise_asc (f);
+      break;
+    case PB_GRAPHICS:
+      ((GRAPHICS_BLOCK *) this)->serialise_asc (f);
+      break;
+    case PB_RULES:
+      ((RULE_BLOCK *) this)->serialise_asc (f);
+      break;
+    case PB_IMAGE:
+      ((IMAGE_BLOCK *) this)->serialise_asc (f);
+      break;
+    case PB_SCRIBBLE:
+      ((SCRIBBLE_BLOCK *) this)->serialise_asc (f);
+      break;
+    case PB_WEIRD:
+      ((WEIRD_BLOCK *) this)->serialise_asc (f);
+      break;
+    default:
+      break;
+  }
+}
+
+
+/**********************************************************************
+ * PAGE_BLOCK::internal_serialise_asc()  Convert to ascii file.
+ *
+ **********************************************************************/
+
+void PAGE_BLOCK::internal_serialise_asc(         //convert to ascii
+                                        FILE *f  //file to use
+                                       ) {
+  ((POLY_BLOCK *) this)->serialise_asc (f);
+  serialise_INT32(f, pb_type); 
+  children.serialise_asc (f);
+}
+
+
+/**********************************************************************
+ * PAGE_BLOCK::de_serialise_asc()  Convert from ascii file.
+ *
+ **********************************************************************/
+
+void PAGE_BLOCK::de_serialise_asc(         //convert from ascii
+                                  FILE *f  //file to use
+                                 ) {
+  PAGE_BLOCK *page_block;        //new block for list
+  INT32 len;                     /*length to retrive */
+  PAGE_BLOCK_IT it;
+
+  ((POLY_BLOCK *) this)->de_serialise_asc (f);
+  pb_type = (PB_TYPE) de_serialise_INT32 (f);
+  //      children.de_serialise_asc(f);
+  len = de_serialise_INT32 (f);
+  it.set_to_list (&children);
+  for (; len > 0; len--) {
+    page_block = new_de_serialise_asc (f);
+    it.add_to_end (page_block);  /*put on the list */
+  }
+}
+
+
+/**********************************************************************
+ * PAGE_BLOCK::new_de_serialise_asc()  Convert from ascii file.
+ *
+ **********************************************************************/
+
+PAGE_BLOCK *PAGE_BLOCK::new_de_serialise_asc(         //convert from ascii
+                                             FILE *f  //file to use
+                                            ) {
+  PB_TYPE type;
+  TEXT_BLOCK *tblock;
+  GRAPHICS_BLOCK *gblock;
+  RULE_BLOCK *rblock;
+  IMAGE_BLOCK *iblock;
+  SCRIBBLE_BLOCK *sblock;
+  WEIRD_BLOCK *wblock;
+
+  type = (PB_TYPE) de_serialise_INT32 (f);
+  switch (type) {
+    case PB_TEXT:
+      tblock = new TEXT_BLOCK;
+      tblock->de_serialise_asc (f);
+      return tblock;
+    case PB_GRAPHICS:
+      gblock = new GRAPHICS_BLOCK;
+      gblock->de_serialise_asc (f);
+      return gblock;
+    case PB_RULES:
+      rblock = new RULE_BLOCK;
+      rblock->de_serialise_asc (f);
+      return rblock;
+    case PB_IMAGE:
+      iblock = new IMAGE_BLOCK;
+      iblock->de_serialise_asc (f);
+      return iblock;
+    case PB_SCRIBBLE:
+      sblock = new SCRIBBLE_BLOCK;
+      sblock->de_serialise_asc (f);
+      return sblock;
+    case PB_WEIRD:
+      wblock = new WEIRD_BLOCK;
+      wblock->de_serialise_asc (f);
+      return wblock;
+    default:
+      return NULL;
+  }
+}
+
+
+void PAGE_BLOCK::show_attrs(DEBUG_WIN *f) { 
+  PAGE_BLOCK_IT it;
+
+  switch (pb_type) {
+    case PB_TEXT:
+      ((TEXT_BLOCK *) this)->show_attrs (f);
+      break;
+    case PB_GRAPHICS:
+      ((GRAPHICS_BLOCK *) this)->show_attrs (f);
+      break;
+    case PB_RULES:
+      ((RULE_BLOCK *) this)->show_attrs (f);
+      break;
+    case PB_IMAGE:
+      ((IMAGE_BLOCK *) this)->show_attrs (f);
+      break;
+    case PB_SCRIBBLE:
+      ((SCRIBBLE_BLOCK *) this)->show_attrs (f);
+      break;
+    case PB_WEIRD:
+      ((WEIRD_BLOCK *) this)->show_attrs (f);
+      break;
+    default:
+      break;
+  }
+
+  if (!children.empty ()) {
+    f->dprintf ("containing subblocks\n");
+    it.set_to_list (&children);
+    for (it.mark_cycle_pt (); !it.cycled_list (); it.forward ())
+      it.data ()->show_attrs (f);
+    f->dprintf ("end of subblocks\n");
+  }
+}
+
+
+PAGE_BLOCK::PAGE_BLOCK (ICOORDELT_LIST * points, PB_TYPE type, PAGE_BLOCK_LIST * child):POLY_BLOCK (points,
+POLY_PAGE) {
+  PAGE_BLOCK_IT
+    c = &children;
+
+  pb_type = type;
+  children.clear ();
+  c.move_to_first ();
+  c.add_list_before (child);
+}
+
+
+PAGE_BLOCK::PAGE_BLOCK (ICOORDELT_LIST * points, PB_TYPE type):POLY_BLOCK (points,
+POLY_PAGE) {
+  pb_type = type;
+  children.clear ();
+}
+
+
+void PAGE_BLOCK::add_a_child(PAGE_BLOCK *newchild) { 
+  PAGE_BLOCK_IT c = &children;
+
+  c.move_to_first ();
+  c.add_to_end (newchild);
+}
+
+
+/**********************************************************************
+ * PAGE_BLOCK::rotate
+ *
+ * Rotate the PAGE_BLOCK and its children
+ **********************************************************************/
+
+void PAGE_BLOCK::rotate(  //cos,sin
+                        FCOORD rotation) {
+                                 //sub block iterator
+  PAGE_BLOCK_IT child_it = &children;
+  PAGE_BLOCK *child;             //child block
+
+  for (child_it.mark_cycle_pt (); !child_it.cycled_list ();
+  child_it.forward ()) {
+    child = child_it.data ();
+    child->rotate (rotation);
+  }
+  if (pb_type == PB_TEXT)
+    ((TEXT_BLOCK *) this)->rotate (rotation);
+  else
+    POLY_BLOCK::rotate(rotation); 
+}
+
+
+/**********************************************************************
+ * PAGE_BLOCK::move
+ *
+ * Move the PAGE_BLOCK and its children
+ **********************************************************************/
+
+void PAGE_BLOCK::move(ICOORD shift  //amount to move
+                     ) {
+                                 //sub block iterator
+  PAGE_BLOCK_IT child_it = &children;
+  PAGE_BLOCK *child;             //child block
+
+  for (child_it.mark_cycle_pt (); !child_it.cycled_list ();
+  child_it.forward ()) {
+    child = child_it.data ();
+    child->move (shift);
+  }
+  if (pb_type == PB_TEXT)
+    ((TEXT_BLOCK *) this)->move (shift);
+  else
+    POLY_BLOCK::move(shift); 
+}
+
+#ifndef GRAPHICS_DISABLED
+void PAGE_BLOCK::basic_plot(WINDOW window, COLOUR colour) { 
+  PAGE_BLOCK_IT c = &children;
+
+  POLY_BLOCK::plot (window, colour, 0);
+
+  if (!c.empty ())
+    for (c.mark_cycle_pt (); !c.cycled_list (); c.forward ())
+      c.data ()->plot (window, colour);
+}
+
+
+void PAGE_BLOCK::plot(WINDOW window, COLOUR colour) { 
+  TEXT_BLOCK *tblock;
+  WEIRD_BLOCK *wblock;
+
+  switch (pb_type) {
+    case PB_TEXT:
+      basic_plot(window, colour); 
+      tblock = (TEXT_BLOCK *) this;
+      tblock->plot (window, colour, REGION_COLOUR, SUBREGION_COLOUR);
+      break;
+    case PB_WEIRD:
+      wblock = (WEIRD_BLOCK *) this;
+      wblock->plot (window, colour);
+      break;
+    default:
+      basic_plot(window, colour); 
+      break;
+  }
+}
+#endif
+
+void show_all_in(PAGE_BLOCK *pblock, POLY_BLOCK *show_area, DEBUG_WIN *f) { 
+  PAGE_BLOCK_IT c;
+  INT16 i, pnum;
+
+  c.set_to_list (pblock->child ());
+  pnum = pblock->child ()->length ();
+  for (i = 0; i < pnum; i++, c.forward ()) {
+    if (show_area->contains (c.data ()))
+      c.data ()->show_attrs (f);
+    else if (show_area->overlap (c.data ()))
+      show_all_in (c.data (), show_area, f);
+  }
+}
+
+
+void delete_all_in(PAGE_BLOCK *pblock, POLY_BLOCK *delete_area) { 
+  PAGE_BLOCK_IT c;
+  INT16 i, pnum;
+
+  c.set_to_list (pblock->child ());
+  pnum = pblock->child ()->length ();
+  for (i = 0; i < pnum; i++, c.forward ()) {
+    if (delete_area->contains (c.data ()))
+      c.extract ()->pb_delete ();
+    else if (delete_area->overlap (c.data ()))
+      delete_all_in (c.data (), delete_area);
+  }
+}
+
+
+PAGE_BLOCK *smallest_containing(PAGE_BLOCK *pblock, POLY_BLOCK *other) { 
+  PAGE_BLOCK_IT c;
+
+  c.set_to_list (pblock->child ());
+  if (c.empty ())
+    return (pblock);
+
+  for (c.mark_cycle_pt (); !c.cycled_list (); c.forward ())
+    if (c.data ()->contains (other))
+      return (smallest_containing (c.data (), other));
+
+  return (pblock);
+}
+
+
+TEXT_BLOCK::TEXT_BLOCK (ICOORDELT_LIST * points, BOOL8 backg[NUM_BACKGROUNDS]):PAGE_BLOCK (points,
+PB_TEXT) {
+  int
+    i;
+
+  for (i = 0; i < NUM_BACKGROUNDS; i++)
+    background.set_bit (i, backg[i]);
+
+  text_regions.clear ();
+}
+
+
+void
+TEXT_BLOCK::set_attrs (BOOL8 backg[NUM_BACKGROUNDS]) {
+  int i;
+
+  for (i = 0; i < NUM_BACKGROUNDS; i++)
+    background.set_bit (i, backg[i]);
+}
+
+
+void TEXT_BLOCK::add_a_region(TEXT_REGION *newchild) { 
+  TEXT_REGION_IT c;
+
+  c.set_to_list (&text_regions);
+
+  c.move_to_first ();
+  c.add_to_end (newchild);
+}
+
+
+/**********************************************************************
+ * TEXT_BLOCK::rotate
+ *
+ * Rotate the TEXT_BLOCK and its children
+ **********************************************************************/
+
+void TEXT_BLOCK::rotate(  //cos,sin
+                        FCOORD rotation) {
+                                 //sub block iterator
+  TEXT_REGION_IT child_it = &text_regions;
+  TEXT_REGION *child;            //child block
+
+  for (child_it.mark_cycle_pt (); !child_it.cycled_list ();
+  child_it.forward ()) {
+    child = child_it.data ();
+    child->rotate (rotation);
+  }
+  POLY_BLOCK::rotate(rotation); 
+}
+
+
+/**********************************************************************
+ * TEXT_BLOCK::move
+ *
+ * Move the TEXT_BLOCK and its children
+ **********************************************************************/
+
+void TEXT_BLOCK::move(ICOORD shift  //amount to move
+                     ) {
+                                 //sub block iterator
+  TEXT_REGION_IT child_it = &text_regions;
+  TEXT_REGION *child;            //child block
+
+  for (child_it.mark_cycle_pt (); !child_it.cycled_list ();
+  child_it.forward ()) {
+    child = child_it.data ();
+    child->move (shift);
+  }
+  POLY_BLOCK::move(shift); 
+}
+
+
+/**********************************************************************
+ * TEXT_BLOCK::serialise_asc()  Convert to ascii file.
+ *
+ **********************************************************************/
+
+void TEXT_BLOCK::serialise_asc(         //convert to ascii
+                               FILE *f  //file to use
+                              ) {
+  ((PAGE_BLOCK *) this)->internal_serialise_asc (f);
+  serialise_INT32 (f, background.val);
+  text_regions.serialise_asc (f);
+}
+
+
+/**********************************************************************
+ * TEXT_BLOCK::de_serialise_asc()  Convert from ascii file.
+ *
+ **********************************************************************/
+
+void TEXT_BLOCK::de_serialise_asc(         //convert from ascii
+                                  FILE *f  //file to use
+                                 ) {
+  ((PAGE_BLOCK *) this)->de_serialise_asc (f);
+  background.val = de_serialise_INT32 (f);
+  text_regions.de_serialise_asc (f);
+}
+
+
+#ifndef GRAPHICS_DISABLED
+void TEXT_BLOCK::plot(WINDOW window,
+                      COLOUR colour,
+                      COLOUR region_colour,
+                      COLOUR subregion_colour) {
+  TEXT_REGION_IT t = &text_regions, tc;
+
+  PAGE_BLOCK::basic_plot(window, colour); 
+
+  if (!t.empty ())
+  for (t.mark_cycle_pt (); !t.cycled_list (); t.forward ()) {
+    t.data ()->plot (window, region_colour, t.data ()->id_no ());
+    tc.set_to_list (t.data ()->regions ());
+    if (!tc.empty ())
+      for (tc.mark_cycle_pt (); !tc.cycled_list (); tc.forward ())
+        tc.data ()->plot (window, subregion_colour, -1);
+  }
+}
+#endif
+
+
+void TEXT_BLOCK::show_attrs(DEBUG_WIN *f) { 
+  TEXT_REGION_IT it;
+
+  f->dprintf ("TEXT BLOCK\n");
+  print_background(f, background); 
+  if (!text_regions.empty ()) {
+    f->dprintf ("containing text regions:\n");
+    it.set_to_list (&text_regions);
+    for (it.mark_cycle_pt (); !it.cycled_list (); it.forward ())
+      it.data ()->show_attrs (f);
+    f->dprintf ("end of regions\n");
+  }
+}
+
+
+DLLSYM void show_all_tr_in(TEXT_BLOCK *tblock,
+                           POLY_BLOCK *show_area,
+                           DEBUG_WIN *f) {
+  TEXT_REGION_IT t, tc;
+  INT16 i, tnum, j, ttnum;
+
+  t.set_to_list (tblock->regions ());
+  tnum = tblock->regions ()->length ();
+  for (i = 0; i < tnum; i++, t.forward ()) {
+    if (show_area->contains (t.data ()))
+      t.data ()->show_attrs (f);
+    else if (show_area->overlap (t.data ())) {
+      tc.set_to_list (t.data ()->regions ());
+      ttnum = t.data ()->regions ()->length ();
+      for (j = 0; j < ttnum; j++, tc.forward ())
+        if (show_area->contains (tc.data ()))
+          tc.data ()->show_attrs (f);
+    }
+  }
+}
+
+
+void delete_all_tr_in(TEXT_BLOCK *tblock, POLY_BLOCK *delete_area) { 
+  TEXT_REGION_IT t, tc;
+  INT16 i, tnum, j, ttnum;
+
+  t.set_to_list (tblock->regions ());
+  tnum = tblock->regions ()->length ();
+  for (i = 0; i < tnum; i++, t.forward ()) {
+    if (delete_area->contains (t.data ()))
+      delete (t.extract ());
+    else if (delete_area->overlap (t.data ())) {
+      tc.set_to_list (t.data ()->regions ());
+      ttnum = t.data ()->regions ()->length ();
+      for (j = 0; j < ttnum; j++, tc.forward ())
+        if (delete_area->contains (tc.data ()))
+          delete (tc.extract ());
+    }
+  }
+}
+
+
+RULE_BLOCK::RULE_BLOCK (ICOORDELT_LIST * points, INT8 sing, INT8 colo):PAGE_BLOCK (points,
+PB_RULES) {
+  multiplicity = sing;
+  colour = colo;
+}
+
+
+void RULE_BLOCK::set_attrs(INT8 sing, INT8 colo) { 
+  multiplicity = sing;
+  colour = colo;
+}
+
+
+void RULE_BLOCK::show_attrs(DEBUG_WIN *f) { 
+  f->dprintf ("RULE BLOCK with attributes %s, %s\n",
+    blabel[R_START][multiplicity], blabel[R_START + 1][colour]);
+}
+
+
+/**********************************************************************
+ * RULE_BLOCK::serialise_asc()  Convert to ascii file.
+ *
+ **********************************************************************/
+
+void RULE_BLOCK::serialise_asc(         //convert to ascii
+                               FILE *f  //file to use
+                              ) {
+  ((PAGE_BLOCK *) this)->internal_serialise_asc (f);
+  serialise_INT32(f, multiplicity); 
+  serialise_INT32(f, colour); 
+}
+
+
+/**********************************************************************
+ * RULE_BLOCK::de_serialise_asc()  Convert from ascii file.
+ *
+ **********************************************************************/
+
+void RULE_BLOCK::de_serialise_asc(         //convert from ascii
+                                  FILE *f  //file to use
+                                 ) {
+  ((PAGE_BLOCK *) this)->de_serialise_asc (f);
+  multiplicity = de_serialise_INT32 (f);
+  colour = de_serialise_INT32 (f);
+}
+
+
+GRAPHICS_BLOCK::GRAPHICS_BLOCK (ICOORDELT_LIST * points, BOOL8 backg[NUM_BACKGROUNDS], INT8 foreg):PAGE_BLOCK (points,
+PB_GRAPHICS) {
+  int
+    i;
+
+  for (i = 0; i < NUM_BACKGROUNDS; i++)
+    background.set_bit (i, backg[i]);
+
+  foreground = foreg;
+}
+
+
+void
+GRAPHICS_BLOCK::set_attrs (BOOL8 backg[NUM_BACKGROUNDS], INT8 foreg) {
+  int i;
+
+  for (i = 0; i < NUM_BACKGROUNDS; i++)
+    background.set_bit (i, backg[i]);
+
+  foreground = foreg;
+}
+
+
+void GRAPHICS_BLOCK::show_attrs(DEBUG_WIN *f) { 
+  f->dprintf ("GRAPHICS BLOCK with attribute %s\n",
+    blabel[G_START][foreground]);
+  print_background(f, background); 
+}
+
+
+/**********************************************************************
+ * GRAPHICS_BLOCK::serialise_asc()  Convert to ascii file.
+ *
+ **********************************************************************/
+
+void GRAPHICS_BLOCK::serialise_asc(         //convert to ascii
+                                   FILE *f  //file to use
+                                  ) {
+  ((PAGE_BLOCK *) this)->internal_serialise_asc (f);
+  serialise_INT32 (f, background.val);
+  serialise_INT32(f, foreground); 
+}
+
+
+/**********************************************************************
+ * GRAPHICS_BLOCK::de_serialise_asc()  Convert from ascii file.
+ *
+ **********************************************************************/
+
+void GRAPHICS_BLOCK::de_serialise_asc(         //convert from ascii
+                                      FILE *f  //file to use
+                                     ) {
+  ((PAGE_BLOCK *) this)->de_serialise_asc (f);
+  background.val = de_serialise_INT32 (f);
+  foreground = de_serialise_INT32 (f);
+}
+
+
+IMAGE_BLOCK::IMAGE_BLOCK (ICOORDELT_LIST * points, INT8 colo, INT8 qual):PAGE_BLOCK (points,
+PB_IMAGE) {
+  colour = colo;
+  quality = qual;
+}
+
+
+void IMAGE_BLOCK::set_attrs(INT8 colo, INT8 qual) { 
+  colour = colo;
+  quality = qual;
+}
+
+
+void IMAGE_BLOCK::show_attrs(DEBUG_WIN *f) { 
+  f->dprintf ("IMAGE BLOCK with attributes %s, %s\n", blabel[I_START][colour],
+    blabel[I_START + 1][quality]);
+}
+
+
+/**********************************************************************
+ * IMAGE_BLOCK::serialise_asc()  Convert to ascii file.
+ *
+ **********************************************************************/
+
+void IMAGE_BLOCK::serialise_asc(         //convert to ascii
+                                FILE *f  //file to use
+                               ) {
+  ((PAGE_BLOCK *) this)->internal_serialise_asc (f);
+  serialise_INT32(f, colour); 
+  serialise_INT32(f, quality); 
+}
+
+
+/**********************************************************************
+ * IMAGE_BLOCK::de_serialise_asc()  Convert from ascii file.
+ *
+ **********************************************************************/
+
+void IMAGE_BLOCK::de_serialise_asc(         //convert from ascii
+                                   FILE *f  //file to use
+                                  ) {
+  ((PAGE_BLOCK *) this)->de_serialise_asc (f);
+  colour = de_serialise_INT32 (f);
+  quality = de_serialise_INT32 (f);
+}
+
+
+SCRIBBLE_BLOCK::SCRIBBLE_BLOCK (ICOORDELT_LIST * points, BOOL8 backg[NUM_BACKGROUNDS], INT8 foreg):PAGE_BLOCK (points,
+PB_SCRIBBLE) {
+  int
+    i;
+
+  for (i = 0; i < NUM_BACKGROUNDS; i++)
+    background.set_bit (i, backg[i]);
+
+  foreground = foreg;
+}
+
+
+void
+SCRIBBLE_BLOCK::set_attrs (BOOL8 backg[NUM_BACKGROUNDS], INT8 foreg) {
+  int i;
+
+  for (i = 0; i < NUM_BACKGROUNDS; i++)
+    background.set_bit (i, backg[i]);
+
+  foreground = foreg;
+}
+
+
+void SCRIBBLE_BLOCK::show_attrs(DEBUG_WIN *f) { 
+  f->dprintf ("SCRIBBLE BLOCK with attributes %s\n",
+    blabel[S_START][foreground]);
+  print_background(f, background); 
+}
+
+
+/**********************************************************************
+ * SCRIBBLE_BLOCK::serialise_asc()  Convert to ascii file.
+ *
+ **********************************************************************/
+
+void SCRIBBLE_BLOCK::serialise_asc(         //convert to ascii
+                                   FILE *f  //file to use
+                                  ) {
+  ((PAGE_BLOCK *) this)->internal_serialise_asc (f);
+  serialise_INT32 (f, background.val);
+  serialise_INT32(f, foreground); 
+}
+
+
+/**********************************************************************
+ * SCRIBBLE_BLOCK::de_serialise_asc()  Convert from ascii file.
+ *
+ **********************************************************************/
+
+void SCRIBBLE_BLOCK::de_serialise_asc(         //convert from ascii
+                                      FILE *f  //file to use
+                                     ) {
+  ((PAGE_BLOCK *) this)->de_serialise_asc (f);
+  background.val = de_serialise_INT32 (f);
+  foreground = de_serialise_INT32 (f);
+}
+
+
+WEIRD_BLOCK::WEIRD_BLOCK (ICOORDELT_LIST * points, INT32 id_no):PAGE_BLOCK (points,
+PB_WEIRD) {
+  id_number = id_no;
+}
+
+
+#ifndef GRAPHICS_DISABLED
+void WEIRD_BLOCK::plot(WINDOW window, COLOUR colour) { 
+  PAGE_BLOCK_IT c = this->child ();
+
+  POLY_BLOCK::plot(window, colour, id_number); 
+
+  if (!c.empty ())
+    for (c.mark_cycle_pt (); !c.cycled_list (); c.forward ())
+      c.data ()->plot (window, colour);
+}
+#endif
+
+
+void WEIRD_BLOCK::set_id(INT32 id_no) { 
+  id_number = id_no;
+}
+
+
+void WEIRD_BLOCK::show_attrs(DEBUG_WIN *f) { 
+  f->dprintf ("WEIRD BLOCK with id number %d\n", id_number);
+}
+
+
+/**********************************************************************
+ * WEIRD_BLOCK::serialise_asc()  Convert to ascii file.
+ *
+ **********************************************************************/
+
+void WEIRD_BLOCK::serialise_asc(         //convert to ascii
+                                FILE *f  //file to use
+                               ) {
+  ((PAGE_BLOCK *) this)->internal_serialise_asc (f);
+  serialise_INT32(f, id_number); 
+}
+
+
+/**********************************************************************
+ * WEIRD_BLOCK::de_serialise_asc()  Convert from ascii file.
+ *
+ **********************************************************************/
+
+void WEIRD_BLOCK::de_serialise_asc(         //convert from ascii
+                                   FILE *f  //file to use
+                                  ) {
+  ((PAGE_BLOCK *) this)->de_serialise_asc (f);
+  id_number = de_serialise_INT32 (f);
+}
+
+
+void print_background(DEBUG_WIN *f, BITS16 background) { 
+  int i;
+
+  f->dprintf ("Background is \n");
+  for (i = 0; i < NUM_BACKGROUNDS; i++) {
+    if (background.bit (i))
+      f->dprintf ("%s, ", backlabel[i]);
+  }
+
+  f->dprintf ("\n");
+
+}

ccstruct/pageblk.h

@@ -0,0 +1,318 @@
+#ifndef           PAGEBLK_C
+#define           PAGEBLK_C
+
+#include "elst.h"
+#include "txtregn.h"
+#include "bits16.h"
+
+#include          "hpddef.h"     //must be last (handpd.dll)
+
+enum PB_TYPE
+{
+  PB_TEXT,
+  PB_RULES,
+  PB_GRAPHICS,
+  PB_IMAGE,
+  PB_SCRIBBLE,
+  PB_WEIRD
+};
+
+class DLLSYM PAGE_BLOCK;         //forward decl
+class DLLSYM TEXT_BLOCK;         //forward decl
+class DLLSYM GRAPHICS_BLOCK;     //forward decl
+class DLLSYM RULE_BLOCK;         //forward decl
+class DLLSYM IMAGE_BLOCK;        //forward decl
+class DLLSYM SCRIBBLE_BLOCK;     //forward decl
+class DLLSYM WEIRD_BLOCK;        //forward decl
+
+ELISTIZEH_S (PAGE_BLOCK)
+class DLLSYM PAGE_BLOCK:public ELIST_LINK, public POLY_BLOCK
+//page block
+{
+  public:
+    PAGE_BLOCK() { 
+    }                            //empty constructor
+    PAGE_BLOCK(  //simple constructor
+               ICOORDELT_LIST *points,
+               PB_TYPE type,
+               PAGE_BLOCK_LIST *child);
+
+    PAGE_BLOCK(  //simple constructor
+               ICOORDELT_LIST *points,
+               PB_TYPE type);
+
+    ~PAGE_BLOCK () {             //destructor
+    }
+
+    void add_a_child(PAGE_BLOCK *newchild); 
+
+    PB_TYPE type() {  //get type
+      return pb_type;
+    }
+
+    PAGE_BLOCK_LIST *child() {  //get children
+      return &children;
+    }
+
+    void rotate(  //rotate it
+                FCOORD rotation);
+    void move(                //move it
+              ICOORD shift);  //vector
+
+    void basic_plot(WINDOW window, COLOUR colour); 
+
+    void plot(WINDOW window, COLOUR colour); 
+
+    void show_attrs(DEBUG_WIN *debug); 
+
+    NEWDELETE2 (PAGE_BLOCK) void pb_delete ();
+
+    void serialise(FILE *f); 
+
+    static PAGE_BLOCK *de_serialise(FILE *f); 
+
+    void prep_serialise() {  //set ptrs to counts
+      POLY_BLOCK::prep_serialise(); 
+      children.prep_serialise ();
+    }
+
+    void dump(  //write external bits
+              FILE *f) {
+      POLY_BLOCK::dump(f); 
+      children.dump (f);
+    }
+
+    void de_dump(  //read external bits
+                 FILE *f) {
+      POLY_BLOCK::de_dump(f); 
+      children.de_dump (f);
+    }
+
+    //note that due to the awful switched nature of the PAGE_BLOCK class,
+    //a PAGE_BLOCK_LIST cannot be de-serialised by the normal mechanism, since
+    //each element cannot be de-serialised in place.
+    //To fix this it is important to use read_poly_blocks or the code therein.
+    void serialise_asc(  //serialise to ascii
+                       FILE *f);
+    void internal_serialise_asc(  //serialise to ascii
+                                FILE *f);
+    void de_serialise_asc(  //serialise from ascii
+                          FILE *f);
+                                 //make one from ascii
+    static PAGE_BLOCK *new_de_serialise_asc(FILE *f); 
+
+  private:
+    PB_TYPE pb_type;
+    PAGE_BLOCK_LIST children;
+};
+
+DLLSYM void show_all_in(PAGE_BLOCK *pblock,
+                        POLY_BLOCK *show_area,
+                        DEBUG_WIN *f);
+
+DLLSYM void delete_all_in(PAGE_BLOCK *pblock, POLY_BLOCK *delete_area); 
+
+DLLSYM PAGE_BLOCK *smallest_containing(PAGE_BLOCK *pblock, POLY_BLOCK *other); 
+
+class DLLSYM TEXT_BLOCK:public PAGE_BLOCK
+                                 //text block
+{
+  public:
+    TEXT_BLOCK() { 
+    }                            //empty constructor
+    TEXT_BLOCK(ICOORDELT_LIST *points); 
+
+    TEXT_BLOCK (ICOORDELT_LIST * points, BOOL8 backg[NUM_BACKGROUNDS]);
+
+                                 //get children
+    TEXT_REGION_LIST *regions() { 
+      return &text_regions;
+    }
+
+    INT32 nregions() { 
+      return text_regions.length ();
+    }
+
+    void add_a_region(TEXT_REGION *newchild); 
+
+    void rotate(  //rotate it
+                FCOORD rotation);
+    void move(                //move it
+              ICOORD shift);  //vector
+
+    void plot(WINDOW window,
+              COLOUR colour,
+              COLOUR region_colour,
+              COLOUR subregion_colour);
+
+    void set_attrs (BOOL8 backg[NUM_BACKGROUNDS]);
+
+    void show_attrs(DEBUG_WIN *debug); 
+
+    void prep_serialise() {  //set ptrs to counts
+      PAGE_BLOCK::prep_serialise(); 
+      text_regions.prep_serialise ();
+    }
+
+    void dump(  //write external bits
+              FILE *f) {
+      PAGE_BLOCK::dump(f); 
+      text_regions.dump (f);
+    }
+
+    void de_dump(  //read external bits
+                 FILE *f) {
+      PAGE_BLOCK::de_dump(f); 
+      text_regions.de_dump (f);
+    }
+
+                                 //serialise to ascii
+    make_serialise (TEXT_BLOCK) void serialise_asc (
+      FILE * f);
+    void de_serialise_asc(  //serialise from ascii
+                          FILE *f);
+
+  private:
+    BITS16 background;
+
+    TEXT_REGION_LIST text_regions;
+};
+
+DLLSYM void delete_all_tr_in(TEXT_BLOCK *tblock, POLY_BLOCK *delete_area); 
+
+DLLSYM void show_all_tr_in(TEXT_BLOCK *tblock,
+                           POLY_BLOCK *show_area,
+                           DEBUG_WIN *f);
+
+class DLLSYM RULE_BLOCK:public PAGE_BLOCK
+                                 //rule block
+{
+  public:
+    RULE_BLOCK() { 
+    }                            //empty constructor
+    RULE_BLOCK(ICOORDELT_LIST *points, INT8 sing, INT8 colo); 
+
+    void set_attrs(INT8 sing, INT8 colo); 
+
+    void show_attrs(DEBUG_WIN *debug); 
+
+                                 //serialise to ascii
+    make_serialise (RULE_BLOCK) void serialise_asc (
+      FILE * f);
+    void de_serialise_asc(  //serialise from ascii
+                          FILE *f);
+
+  private:
+    INT8 multiplicity;
+    INT8 colour;
+
+};
+
+class DLLSYM GRAPHICS_BLOCK:public PAGE_BLOCK
+                                 //graphics block
+{
+  public:
+    GRAPHICS_BLOCK() { 
+    }                            //empty constructor
+    GRAPHICS_BLOCK (ICOORDELT_LIST * points,
+      BOOL8 backg[NUM_BACKGROUNDS], INT8 foreg);
+
+    void set_attrs (BOOL8 backg[NUM_BACKGROUNDS], INT8 foreg);
+
+    void show_attrs(DEBUG_WIN *debug); 
+
+                                 //serialise to ascii
+    make_serialise (GRAPHICS_BLOCK) void serialise_asc (
+      FILE * f);
+    void de_serialise_asc(  //serialise from ascii
+                          FILE *f);
+
+  private:
+    BITS16 background;
+    INT8 foreground;
+
+};
+
+class DLLSYM IMAGE_BLOCK:public PAGE_BLOCK
+                                 //image block
+{
+  public:
+    IMAGE_BLOCK() { 
+    }                            //empty constructor
+    IMAGE_BLOCK(ICOORDELT_LIST *points, INT8 colo, INT8 qual); 
+
+    void set_attrs(INT8 colo, INT8 qual); 
+
+    void show_attrs(DEBUG_WIN *debug); 
+
+                                 //serialise to ascii
+    make_serialise (IMAGE_BLOCK) void serialise_asc (
+      FILE * f);
+    void de_serialise_asc(  //serialise from ascii
+                          FILE *f);
+
+  private:
+    INT8 colour;
+    INT8 quality;
+
+};
+
+class DLLSYM SCRIBBLE_BLOCK:public PAGE_BLOCK
+                                 //scribble block
+{
+  public:
+    SCRIBBLE_BLOCK() { 
+    }                            //empty constructor
+    SCRIBBLE_BLOCK (ICOORDELT_LIST * points,
+      BOOL8 backg[NUM_BACKGROUNDS], INT8 foreg);
+
+    void set_attrs (BOOL8 backg[NUM_BACKGROUNDS], INT8 foreg);
+
+    void show_attrs(DEBUG_WIN *debug); 
+
+                                 //serialise to ascii
+    make_serialise (SCRIBBLE_BLOCK) void serialise_asc (
+      FILE * f);
+    void de_serialise_asc(  //serialise from ascii
+                          FILE *f);
+
+  private:
+    BITS16 background;
+    INT8 foreground;
+};
+
+class DLLSYM WEIRD_BLOCK:public PAGE_BLOCK
+                                 //weird block
+{
+  public:
+    WEIRD_BLOCK() { 
+    }                            //empty constructor
+    WEIRD_BLOCK(ICOORDELT_LIST *points, INT32 id_no); 
+
+    void set_id(INT32 id_no); 
+
+    void show_attrs(DEBUG_WIN *debug); 
+
+    void set_id_no(INT32 new_id) { 
+      id_number = new_id;
+    }
+
+    void plot(WINDOW window, COLOUR colour); 
+
+    INT32 id_no() { 
+      return id_number;
+    }
+
+                                 //serialise to ascii
+    make_serialise (WEIRD_BLOCK) void serialise_asc (
+      FILE * f);
+    void de_serialise_asc(  //serialise from ascii
+                          FILE *f);
+
+  private:
+    INT32 id_number;             //unique id
+
+};
+
+void print_background(DEBUG_WIN *f, BITS16 background); 
+#endif

ccstruct/pageres.cpp

@@ -0,0 +1,325 @@
+/**********************************************************************
+ * File:        pageres.cpp  (Formerly page_res.c)
+ * Description: Results classes used by control.c
+ * Author:		Phil Cheatle
+ * Created:     Tue Sep 22 08:42:49 BST 1992
+ *
+ * (C) Copyright 1992, Hewlett-Packard Ltd.
+ ** Licensed under the Apache License, Version 2.0 (the "License");
+ ** you may not use this file except in compliance with the License.
+ ** You may obtain a copy of the License at
+ ** http://www.apache.org/licenses/LICENSE-2.0
+ ** Unless required by applicable law or agreed to in writing, software
+ ** distributed under the License is distributed on an "AS IS" BASIS,
+ ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ ** See the License for the specific language governing permissions and
+ ** limitations under the License.
+ *
+ **********************************************************************/
+#include "mfcpch.h"
+#include          <stdlib.h>
+#ifdef __UNIX__
+#include          <assert.h>
+#endif
+#include          "pageres.h"
+#include          "notdll.h"
+
+ELISTIZE (BLOCK_RES)
+CLISTIZE (BLOCK_RES) ELISTIZE (ROW_RES) ELISTIZE (WERD_RES)
+/*************************************************************************
+ * PAGE_RES::PAGE_RES
+ *
+ * Constructor for page results
+ *************************************************************************/
+PAGE_RES::PAGE_RES(                            //recursive construct
+                   BLOCK_LIST *the_block_list  //real page
+                  ) {
+  BLOCK_IT block_it(the_block_list); 
+  BLOCK_RES_IT block_res_it(&block_res_list); 
+
+  char_count = 0;
+  rej_count = 0;
+  rejected = FALSE;
+
+  for (block_it.mark_cycle_pt ();
+  !block_it.cycled_list (); block_it.forward ()) {
+    block_res_it.add_to_end (new BLOCK_RES (block_it.data ()));
+  }
+}
+
+
+/*************************************************************************
+ * BLOCK_RES::BLOCK_RES
+ *
+ * Constructor for BLOCK results
+ *************************************************************************/
+
+BLOCK_RES::BLOCK_RES(                  //recursive construct
+                     BLOCK *the_block  //real BLOCK
+                    ) {
+  ROW_IT row_it (the_block->row_list ());
+  ROW_RES_IT row_res_it(&row_res_list); 
+
+  char_count = 0;
+  rej_count = 0;
+  font_class = -1;               //not assigned
+  x_height = -1.0;
+  font_assigned = FALSE;
+  bold = FALSE;
+  italic = FALSE;
+  row_count = 0;
+
+  block = the_block;
+
+  for (row_it.mark_cycle_pt (); !row_it.cycled_list (); row_it.forward ()) {
+    row_res_it.add_to_end (new ROW_RES (row_it.data ()));
+  }
+}
+
+
+/*************************************************************************
+ * ROW_RES::ROW_RES
+ *
+ * Constructor for ROW results
+ *************************************************************************/
+
+ROW_RES::ROW_RES(              //recursive construct
+                 ROW *the_row  //real ROW
+                ) {
+  WERD_IT word_it (the_row->word_list ());
+  WERD_RES_IT word_res_it(&word_res_list); 
+  WERD_RES *combo = NULL;        //current combination of fuzzies
+  WERD_RES *word_res;            //current word
+  WERD *copy_word;
+
+  char_count = 0;
+  rej_count = 0;
+  whole_word_rej_count = 0;
+  font_class = -1;
+  font_class_score = -1.0;
+  bold = FALSE;
+  italic = FALSE;
+
+  row = the_row;
+
+  for (word_it.mark_cycle_pt (); !word_it.cycled_list (); word_it.forward ()) {
+    word_res = new WERD_RES (word_it.data ());
+
+    if (word_res->word->flag (W_FUZZY_NON)) {
+      ASSERT_HOST (combo != NULL);
+      word_res->part_of_combo = TRUE;
+      combo->copy_on (word_res);
+    }
+    if (word_it.data_relative (1)->flag (W_FUZZY_NON)) {
+      if (combo == NULL) {
+        copy_word = new WERD;
+                                 //deep copy
+        *copy_word = *(word_it.data ());
+        combo = new WERD_RES (copy_word);
+        combo->combination = TRUE;
+        word_res_it.add_to_end (combo);
+      }
+      word_res->part_of_combo = TRUE;
+    }
+    else
+      combo = NULL;
+    word_res_it.add_to_end (word_res);
+  }
+}
+
+
+WERD_RES & WERD_RES::operator= ( //assign word_res
+const WERD_RES & source          //from this
+) {
+  this->ELIST_LINK::operator= (source);
+  if (source.combination) {
+    word = new WERD;
+    *word = *(source.word);      //deep copy
+  }
+  else
+    word = source.word;          //pt to same word
+
+  if (source.outword != NULL) {
+    outword = new WERD;
+    *outword = *(source.outword);//deep copy
+  }
+  else
+    outword = NULL;
+
+  denorm = source.denorm;
+  if (source.best_choice != NULL) {
+    best_choice = new WERD_CHOICE;
+    *best_choice = *(source.best_choice);
+    raw_choice = new WERD_CHOICE;
+    *raw_choice = *(source.raw_choice);
+  }
+  else {
+    best_choice = NULL;
+    raw_choice = NULL;
+  }
+  if (source.ep_choice != NULL) {
+    ep_choice = new WERD_CHOICE;
+    *ep_choice = *(source.ep_choice);
+  }
+  else
+    ep_choice = NULL;
+  reject_map = source.reject_map;
+  tess_failed = source.tess_failed;
+  tess_accepted = source.tess_accepted;
+  tess_would_adapt = source.tess_would_adapt;
+  done = source.done;
+  unlv_crunch_mode = source.unlv_crunch_mode;
+  italic = source.italic;
+  bold = source.bold;
+  font1 = source.font1;
+  font1_count = source.font1_count;
+  font2 = source.font2;
+  font2_count = source.font2_count;
+  x_height = source.x_height;
+  caps_height = source.caps_height;
+  guessed_x_ht = source.guessed_x_ht;
+  guessed_caps_ht = source.guessed_caps_ht;
+  combination = source.combination;
+  part_of_combo = source.part_of_combo;
+  reject_spaces = source.reject_spaces;
+  return *this;
+}
+
+
+WERD_RES::~WERD_RES () {
+  if (combination)
+    delete word;
+  if (outword != NULL)
+    delete outword;
+  if (best_choice != NULL) {
+    delete best_choice;
+    delete raw_choice;
+  }
+  if (ep_choice != NULL) {
+    delete ep_choice;
+  }
+}
+
+
+/*************************************************************************
+ * PAGE_RES_IT::restart_page
+ *
+ * Set things up at the start of the page
+ *************************************************************************/
+
+WERD_RES *PAGE_RES_IT::restart_page() { 
+  block_res_it.set_to_list (&page_res->block_res_list);
+  block_res_it.mark_cycle_pt ();
+  block_res = NULL;
+  row_res = NULL;
+  word_res = NULL;
+  next_block_res = NULL;
+  next_row_res = NULL;
+  next_word_res = NULL;
+  internal_forward(TRUE); 
+  return internal_forward (FALSE);
+}
+
+
+/*************************************************************************
+ * PAGE_RES_IT::internal_forward
+ *
+ * Find the next word on the page. Empty blocks and rows are skipped.
+ * The iterator maintains pointers to block, row and word for the previous,
+ * current and next words.  These are correct, regardless of block/row
+ * boundaries. NULL values denote start and end of the page.
+ *************************************************************************/
+
+WERD_RES *PAGE_RES_IT::internal_forward(BOOL8 new_block) { 
+  BOOL8 found_next_word = FALSE;
+  BOOL8 new_row = FALSE;
+
+  prev_block_res = block_res;
+  prev_row_res = row_res;
+  prev_word_res = word_res;
+  block_res = next_block_res;
+  row_res = next_row_res;
+  word_res = next_word_res;
+
+  while (!found_next_word && !block_res_it.cycled_list ()) {
+    if (new_block) {
+      new_block = FALSE;
+      row_res_it.set_to_list (&block_res_it.data ()->row_res_list);
+      row_res_it.mark_cycle_pt ();
+      new_row = TRUE;
+    }
+    while (!found_next_word && !row_res_it.cycled_list ()) {
+      if (new_row) {
+        new_row = FALSE;
+        word_res_it.set_to_list (&row_res_it.data ()->word_res_list);
+        word_res_it.mark_cycle_pt ();
+      }
+      while (!found_next_word && !word_res_it.cycled_list ()) {
+        next_block_res = block_res_it.data ();
+        next_row_res = row_res_it.data ();
+        next_word_res = word_res_it.data ();
+        found_next_word = TRUE;
+        do {
+          word_res_it.forward ();
+        }
+        while (word_res_it.data ()->part_of_combo);
+      }
+      if (!found_next_word) {    //end of row reached
+        row_res_it.forward ();
+        new_row = TRUE;
+      }
+    }
+    if (!found_next_word) {      //end of block reached
+      block_res_it.forward ();
+      new_block = TRUE;
+    }
+  }
+  if (!found_next_word) {        //end of page reached
+    next_block_res = NULL;
+    next_row_res = NULL;
+    next_word_res = NULL;
+  }
+  return word_res;
+}
+
+
+/*************************************************************************
+ * PAGE_RES_IT::forward_block
+ *
+ * Move to the first word of the next block
+ * Can be followed by subsequent calls to forward() BUT at the first word in
+ * the block, the prev block, row and word are all NULL.
+ *************************************************************************/
+
+WERD_RES *PAGE_RES_IT::forward_block() { 
+  if (block_res == next_block_res) {
+    block_res_it.forward ();;
+    block_res = NULL;
+    row_res = NULL;
+    word_res = NULL;
+    next_block_res = NULL;
+    next_row_res = NULL;
+    next_word_res = NULL;
+    internal_forward(TRUE); 
+  }
+  return internal_forward (FALSE);
+}
+
+
+void PAGE_RES_IT::rej_stat_word() { 
+  INT16 chars_in_word;
+  INT16 rejects_in_word = 0;
+
+  chars_in_word = word_res->reject_map.length ();
+  page_res->char_count += chars_in_word;
+  block_res->char_count += chars_in_word;
+  row_res->char_count += chars_in_word;
+
+  rejects_in_word = word_res->reject_map.reject_count ();
+
+  page_res->rej_count += rejects_in_word;
+  block_res->rej_count += rejects_in_word;
+  row_res->rej_count += rejects_in_word;
+  if (chars_in_word == rejects_in_word)
+    row_res->whole_word_rej_count += rejects_in_word;
+}