Commit Graph

4800 Commits

Author SHA1 Message Date
Egor Pugin
5b65982fa4
Merge pull request #2111 from stweil/clean
Remove several unused methods
2018-12-08 20:07:35 +03:00
Stefan Weil
81ab302d52 FPRow: Remove three unused methods
This fixes warnings from the Intel compiler:

    src/textord/cjkpitch.cpp(319): warning #177:
      function "<unnamed>::FPRow::good_gaps" was declared but never referenced
    src/textord/cjkpitch.cpp(383): warning #177:
      function "<unnamed>::FPRow::is_bad" was declared but never referenced
    src/textord/cjkpitch.cpp(387): warning #177:
      function "<unnamed>::FPRow::is_unknown" was declared but never referenced

Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-12-08 16:43:52 +01:00
Stefan Weil
404f9cd147 SimpleStats: Remove unused method
This fixes a warning from the Intel compiler:

    src/textord/cjkpitch.cpp(79): warning #177:
      function "<unnamed>::SimpleStats::maximum" was declared
      but never referenced

Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-12-08 16:39:46 +01:00
Stefan Weil
a9121d28f3
Merge pull request #2107 from stweil/march
Add check whether compiler supports -march=native flag
2018-12-08 10:53:09 +01:00
Stefan Weil
2c044df959 Fix wrong x_fsize in hOCR output (regression)
The regression was caused by the latest commit
c9e85ab78f.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-12-08 10:39:31 +01:00
Stefan Weil
2ccc5810f3 Add check whether compiler supports -march=native flag
Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-12-05 20:13:28 +01:00
zdenop
ad40131385
Merge pull request #2104 from stweil/fix
Fix two runtime errors
2018-12-04 11:27:48 +01:00
Stefan Weil
c9e85ab78f Fix wrong font attributes in hOCR output
Instrumented code throws this runtime error during OCR:

    ../../src/api/baseapi.cpp:1616:5: runtime error: load of value 128,
      which is not a valid value for type 'bool'
    ../../src/api/baseapi.cpp:1627:5: runtime error: load of value 128,
      which is not a valid value for type 'bool'

If there is no font information (typical for Tesseract with a LSTM model),
the font attributes got random values resulting in wrong hOCR output.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-12-04 10:52:46 +01:00
Stefan Weil
0bdae8f8bf GENERIC_2D_ARRAY: Fix runtime error in assignment operator
Instrumented code throws this runtime error during OCR:

    ../../src/ccstruct/matrix.h:84:11: runtime error:
      null pointer passed as argument 2, which is declared to never be null

Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-12-04 10:48:46 +01:00
zdenop
bee874a1c7
Merge pull request #2098 from stweil/dp-perf
Add config variable for selection of dot product function
2018-12-01 07:50:56 +01:00
Stefan Weil
f0a4d04187 Add config variable for selection of dot product function
All also a C++ implementation with more aggressive compiler options
which is optimized for the CPU where the software was built.

It is now possible to select the function used for the dot product
with -c dotproduct=FUNCTION where FUNCTION can be one of those values:

* auto      selection based on detected hardware (default)
* generic   C++ code with default compiler options
* native    C++ code optimized for build host
* avx       optimized code for AVX
* sse       optimized code for SSE

Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-12-01 00:19:28 +01:00
zdenop
b527b37825
Merge pull request #2097 from stweil/namespace
SIMDDetect: Use tesseract namespace and format code
2018-12-01 00:02:18 +01:00
zdenop
d69dc27fa3
Merge pull request #2096 from stweil/clean
Clean code and fix crash with early use of tprintf
2018-12-01 00:02:08 +01:00
Stefan Weil
1910b1a72b SIMDDetect: Use tesseract namespace and format code
Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-30 23:36:39 +01:00
Stefan Weil
66d3275d0b IntSimdMatrixSSE: Remove unused include statement and simplify code
Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-30 23:14:11 +01:00
Stefan Weil
048eb34934 Add missing static attribute to local inline functions
Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-30 23:14:11 +01:00
Stefan Weil
b73370aac9 Remove unneeded test for nullptr
IntSimdMatrix::GetFastestMultiplier never returns a nullptr.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-30 23:14:11 +01:00
Stefan Weil
e2419b1968 Fix potential crash in tprintf
Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-30 23:14:11 +01:00
Stefan Weil
6b6d9de497 Fix potential crash in STRING class
Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-30 23:14:11 +01:00
zdenop
b6057f5755
Merge pull request #2095 from stweil/optimize
Use -ffast-math for calculation of dot product
2018-11-30 23:04:59 +01:00
Stefan Weil
59fb3370bb Use -ffast-math for calculation of dot product
This reduces the code size for intsimdmatrixavx2 from 2700 to 2668
and slightly improves the performance for fast models with AVX2.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-30 22:52:04 +01:00
Stefan Weil
fda3ba9009 IntSimdMatrixSSE: Fix comment
Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-30 22:13:32 +01:00
zdenop
07b140364f
Merge pull request #2093 from stweil/python
Updates for Python scripts
2018-11-30 08:10:20 +01:00
zdenop
53600c677e
Merge pull request #2092 from stweil/format
Format new ALTO code with clang-format
2018-11-30 08:08:52 +01:00
zdenop
53ee0595f0
Merge pull request #2091 from stweil/alto
Add configuration file for ALTO to installation
2018-11-30 08:08:04 +01:00
zdenop
f6493dd5e8
Merge pull request #2090 from stweil/inline
Optimize performance by using inline functions
2018-11-30 08:07:45 +01:00
Stefan Weil
c59c45fb3e Fix Amharic font list
This was reported for the Python code by LGTM.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-30 08:00:22 +01:00
Stefan Weil
57d0ae06c0 Use Python3 for LGTM
The Python scripts require Python3 and give errors with Python2.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-30 07:11:53 +01:00
Stefan Weil
b148644c1b Make Python script executable
Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-30 07:08:45 +01:00
Stefan Weil
ed48b2a8f5 Format new ALTO code with clang-format
Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-30 06:37:25 +01:00
Stefan Weil
e817d93e62 Add configuration file for ALTO to installation
Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-30 06:17:04 +01:00
Jake Sebright
d7cee03a94 Add support for ALTO output 2018-11-30 06:09:36 +01:00
Stefan Weil
3c047f0ac8 Optimize performance by using inline function DotProduct
This improves performace for the "best" models because it
avoids function calls.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-29 21:43:41 +01:00
Stefan Weil
e161501df6 Optimize performance by using inline MatrixDotVectorInternal
This improves performace for the "best" models because it
avoids function calls.

The compiler also knows the passed values for the parameters
add_bias_fwd and skip_bias_back.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-29 21:37:32 +01:00
Egor Pugin
685b136d89
Fix incorrect condition. 2018-11-29 19:02:54 +03:00
Stefan Weil
9e49539429
Merge pull request #2079 from ajschaeper/master
Add example in README.md
2018-11-26 09:44:24 +01:00
ajschaeper
0223abcb96 Add example in README.md 2018-11-26 09:30:22 +01:00
Egor Pugin
267b79982d
Merge pull request #2076 from jbarlow83/pythonize-training
RFC: Pythonize tesstrain.sh and friends
2018-11-25 13:31:48 +03:00
James R. Barlow
8aa25239ae Fix some of Codacy's complaints 2018-11-24 16:59:01 -08:00
James R. Barlow
9122e6249e Autoreformat code
This increases the deviation from the bash scripts so is done separately.
2018-11-24 00:50:29 -08:00
James R. Barlow
d9ae7ecc49 Pythonize tesstrain.sh -> tesstrain.py
This is a lightweight, semi-Pythonic conversion of tesstrain.sh that currently
supports only LSTM and not the Tesseract 3 training mode.

I attempted to keep source changes minimal so it would be easy to compare
bash to Python in code review and confirm equivalence.

Python 3.6+ is required.  Ubuntu 18.04 ships Python 3.6 and it is a mandatory
package (the package manager is also written in Python), so it is available
in the baseline Tesseract 4.0 system.

There are minor output and behavioral changes, and advantages.  Python's loggingis used.  Temporary files are only deleted on success, so they can be inspected
if training files.  Console output is more terse and the log file is more
verbose.  And there are progress bars!  (The python3-tqdm package is required.)
Where tesstrain.sh would sometimes fail without explanation and return an error
code of 1, it is much easier to find the point of failure in this version.
That was also the main motivation for this work.

Argument checking is also more comprehensive.
2018-11-24 00:45:35 -08:00
pndaza
fc8a3d5bbc combine condition with next 2018-11-24 09:21:05 +06:30
pndaza
5c85d8e03d add missed letters and symbols - 0x104a to 0x104f - 2018-11-24 09:14:31 +06:30
Egor Pugin
b08624acc0 Reapply: Add sw build system script (future cppan replacement). 2018-11-23 01:47:02 +03:00
Egor Pugin
19580b18bc Revert "Add sw build system script (future cppan replacement)."
This reverts commit b1e20043fd.
2018-11-23 01:44:58 +03:00
Egor Pugin
e98661c0e7
Merge pull request #2065 from thegrizzlylabs/fix-configure-android
fix(configure) Don't add rt on Android
2018-11-21 23:29:44 +03:00
Egor Pugin
b1e20043fd Add sw build system script (future cppan replacement). 2018-11-20 00:12:26 +03:00
zdenop
def7cdd641
Merge pull request #2063 from stweil/tprintf
Remove unused include statements for tprintf.h
2018-11-18 19:02:07 +01:00
Stefan Weil
9b783822a0 Remove unused include statements for tprintf.h
Format also a call of tprintf and add a missing explicit include statement.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
2018-11-18 17:25:01 +01:00
zdenop
670ce8e4cf
Merge pull request #2060 from stweil/overflow
Fix wrong results from function streamtofloat
2018-11-18 08:11:55 +01:00