Add ability to build under android (without cube or scrollview).

android/AndroidManifest.xml

@@ -0,0 +1,4 @@
+<!--
+    This file is needed by the android_native_library rule to determine the
+    project directory for ndk-build.
+-->

android/Makefile.am

@@ -0,0 +1 @@
+EXTRA_DIST = AndroidManifest.xml jni/Android.mk jni/Application.mk

android/jni/Android.mk

@@ -0,0 +1,57 @@
+LOCAL_PATH := $(call my-dir)
+include $(CLEAR_VARS)
+
+LOCAL_MODULE := tesseract-$(APP_ABI)
+
+LOCAL_STATIC_LIBRARIES := \
+    mobile_base \
+    leptonica-$(APP_ABI)
+
+LOCAL_C_INCLUDES := $(APP_C_INCLUDES)
+
+LOCAL_C_INCLUDES += \
+  $(LOCAL_PATH)/../../api \
+  $(LOCAL_PATH)/../../ccmain\
+  $(LOCAL_PATH)/../../ccstruct\
+  $(LOCAL_PATH)/../../ccutil\
+  $(LOCAL_PATH)/../../classify\
+  $(LOCAL_PATH)/../../cutil\
+  $(LOCAL_PATH)/../../dict\
+  $(LOCAL_PATH)/../../image\
+  $(LOCAL_PATH)/../../textord\
+  $(LOCAL_PATH)/../../third_party\
+  $(LOCAL_PATH)/../../wordrec\
+  $(LOCAL_PATH)/../../opencl\
+  $(LOCAL_PATH)/../../viewer\
+  $(LOCAL_PATH)/../../../leptonica/include
+
+$(info local c includes=$(LOCAL_C_INCLUDES))
+$(info local path=$(LOCAL_PATH))
+LOCAL_SRC_FILES := $(wildcard $(LOCAL_PATH)/../../api/*.cpp $(LOCAL_PATH)/../../ccmain/*.cpp $(LOCAL_PATH)/../../ccstruct/*.cpp $(LOCAL_PATH)/../../ccutil/*.cpp $(LOCAL_PATH)/../../classify/*.cpp $(LOCAL_PATH)/../../cutil/*.cpp $(LOCAL_PATH)/../../dict/*.cpp $(LOCAL_PATH)/../../image/*.cpp $(LOCAL_PATH)/../../textord/*.cpp $(LOCAL_PATH)/../../viewer/*.cpp $(LOCAL_PATH)/../../wordrec/*.cpp)
+
+EXPLICIT_SRC_EXCLUDES := \
+  $(LOCAL_PATH)/../../ccmain/cubeclassifier.cpp \
+  $(LOCAL_PATH)/../../ccmain/cubeclassifier.h  \
+  $(LOCAL_PATH)/../../ccmain/cube_control.cpp \
+  $(LOCAL_PATH)/../../ccmain/cube_reco_context.cpp \
+  $(LOCAL_PATH)/../../ccmain/cube_reco_context.h \
+  $(LOCAL_PATH)/../../ccmain/tesseract_cube_combiner.cpp \
+  $(LOCAL_PATH)/../../ccmain/tesseract_cube_combiner.h \
+  $(LOCAL_PATH)/../../api/pdfrenderer.cpp \
+  $(LOCAL_PATH)/../../api/tesseractmain.cpp \
+
+LOCAL_SRC_FILES := $(filter-out $(EXPLICIT_SRC_EXCLUDES), $(LOCAL_SRC_FILES))
+
+LOCAL_SRC_FILES := $(LOCAL_SRC_FILES:$(LOCAL_PATH)/%=%)
+
+$(info local src files  = $(LOCAL_SRC_FILES))
+
+LOCAL_LDLIBS := -ldl -llog -ljnigraphics
+LOCAL_CFLAGS := -DANDROID_BUILD -DGRAPHICS_DISABLED
+
+include $(BUILD_SHARED_LIBRARY)
+
+$(call import-module,mobile/base)
+$(call import-module,mobile/base)
+$(call import-module,mobile/util/hash)
+$(call import-module,third_party/leptonica/android/jni)

android/jni/Application.mk

@@ -0,0 +1,13 @@
+# Include common.mk for building google3 native code.
+DEPOT_PATH := $(firstword $(subst /google3, ,$(abspath $(call my-dir))))
+ifneq ($(wildcard $(DEPOT_PATH)/google3/mobile/build/common.mk),)
+  include $(DEPOT_PATH)/google3/mobile/build/common.mk
+else
+  include $(DEPOT_PATH)/READONLY/google3/mobile/build/common.mk
+endif
+
+# Specify the hash namespace that we're using, based on the APP_STL we're using.
+APP_CFLAGS += -Werror -DHASH_NAMESPACE=__gnu_cxx -Wno-error=deprecated-register
+APP_PLATFORM := android-16
+APP_STL := gnustl_static
+NDK_TOOLCHAIN_VERSION := clang

api/baseapi.cpp

@@ -742,6 +742,7 @@ void TessBaseAPI::DumpPGM(const char* filename) {
   fclose(fp);
 }
 
+#ifndef ANDROID_BUILD
 /**
  * Placeholder for call to Cube and test that the input data is correct.
  * reskew is the direction of baselines in the skewed image in
@@ -786,6 +787,7 @@ int CubeAPITest(Boxa* boxa_blocks, Pixa* pixa_blocks,
   ASSERT_HOST(pr_word == word_count);
   return 0;
 }
+#endif
 
 /**
  * Runs page layout analysis in the mode set by SetPageSegMode.
@@ -1022,6 +1024,7 @@ bool TessBaseAPI::ProcessPagesMultipageTiff(const l_uint8 *data,
                                             int timeout_millisec,
                                             TessResultRenderer* renderer,
                                             int tessedit_page_number) {
+#ifndef ANDROID_BUILD
   Pix *pix = NULL;
 #ifdef USE_OPENCL
   OpenclDevice od;
@@ -1052,6 +1055,9 @@ bool TessBaseAPI::ProcessPagesMultipageTiff(const l_uint8 *data,
     if (tessedit_page_number >= 0) break;
   }
   return true;
+#else
+  return false;
+#endif
 }
 
 // Master ProcessPages calls ProcessPagesInternal and then does any post-
@@ -1087,6 +1093,7 @@ bool TessBaseAPI::ProcessPagesInternal(const char* filename,
                                        const char* retry_config,
                                        int timeout_millisec,
                                        TessResultRenderer* renderer) {
+#ifndef ANDROID_BUILD
   PERF_COUNT_START("ProcessPages")
   bool stdInput = !strcmp(filename, "stdin") || !strcmp(filename, "-");
   if (stdInput) {
@@ -1174,6 +1181,9 @@ bool TessBaseAPI::ProcessPagesInternal(const char* filename,
   }
   PERF_COUNT_END
   return true;
+#else
+  return false;
+#endif
 }
 
 bool TessBaseAPI::ProcessPage(Pix* pix, int page_index, const char* filename,
@@ -1207,8 +1217,10 @@ bool TessBaseAPI::ProcessPage(Pix* pix, int page_index, const char* filename,
     failed = Recognize(NULL) < 0;
   }
   if (tesseract_->tessedit_write_images) {
+#ifndef ANDROID_BUILD
     Pix* page_pix = GetThresholdedImage();
     pixWrite("tessinput.tif", page_pix, IFF_TIFF_G4);
+#endif
   }
   if (failed && retry_config != NULL && retry_config[0] != '\0') {
     // Save current config variables before switching modes.
@@ -2613,10 +2625,12 @@ int TessBaseAPI::NumDawgs() const {
   return tesseract_ == NULL ? 0 : tesseract_->getDict().NumDawgs();
 }
 
+#ifndef ANDROID_BUILD
 /** Return a pointer to underlying CubeRecoContext object if present. */
 CubeRecoContext *TessBaseAPI::GetCubeRecoContext() const {
   return (tesseract_ == NULL) ? NULL : tesseract_->GetCubeRecoContext();
 }
+#endif
 
 /** Escape a char string - remove <>&"' with HTML codes. */
 STRING HOcrEscape(const char* text) {

ccmain/control.cpp

@@ -387,12 +387,14 @@ bool Tesseract::recog_all_words(PAGE_RES* page_res,
     // ****************** Pass 5,6 *******************
     rejection_passes(page_res, monitor, target_word_box, word_config);
 
+#ifndef ANDROID_BUILD
     // ****************** Pass 7 *******************
     // Cube combiner.
     // If cube is loaded and its combiner is present, run it.
     if (tessedit_ocr_engine_mode == OEM_TESSERACT_CUBE_COMBINED) {
       run_cube_combiner(page_res);
     }
+#endif
 
     // ****************** Pass 8 *******************
     font_recognition_pass(page_res);
@@ -986,11 +988,13 @@ void Tesseract::classify_word_pass1(const WordData& word_data,
   BLOCK* block = word_data.block;
   prev_word_best_choice_ = word_data.prev_word != NULL
       ? word_data.prev_word->word->best_choice : NULL;
+#ifndef ANDROID_BUILD
   // If we only intend to run cube - run it and return.
   if (tessedit_ocr_engine_mode == OEM_CUBE_ONLY) {
     cube_word_pass1(block, row, *in_word);
     return;
   }
+#endif
   WERD_RES* word = *in_word;
   match_word_pass_n(1, word, row, block);
   if (!word->tess_failed && !word->word->flag(W_REP_CHAR)) {

ccmain/paramsd.h

@@ -24,7 +24,9 @@
 #define VARABLED_H
 
 #include "elst.h"
+#ifndef ANDROID_BUILD
 #include "scrollview.h"
+#endif
 #include "params.h"
 #include "tesseractclass.h"
 

ccmain/tessedit.cpp

@@ -194,7 +194,11 @@ bool Tesseract::init_tesseract_lang_data(
     if (tessdata_manager_debug_level) tprintf("Loaded ambigs\n");
   }
 
-  // Load Cube objects if necessary.
+  // The various OcrEngineMode settings (see publictypes.h) determine which
+  // engine-specific data files need to be loaded. Currently everything needs
+  // the base tesseract data, which supplies other useful information, but
+  // alternative engines, such as cube and LSTM are optional.
+#ifndef ANDROID_BUILD
   if (tessedit_ocr_engine_mode == OEM_CUBE_ONLY) {
     ASSERT_HOST(init_cube_objects(false, &tessdata_manager));
     if (tessdata_manager_debug_level)
@@ -204,7 +208,7 @@ bool Tesseract::init_tesseract_lang_data(
     if (tessdata_manager_debug_level)
       tprintf("Loaded Cube with combiner\n");
   }
-
+#endif
   // Init ParamsModel.
   // Load pass1 and pass2 weights (for now these two sets are the same, but in
   // the future separate sets of weights can be generated).
@@ -475,5 +479,4 @@ enum CMD_EVENTS
   RECOG_PSEUDO,
   ACTION_2_CMD_EVENT
 };
-
 }  // namespace tesseract

ccmain/tesseractclass.cpp

@@ -37,11 +37,15 @@
 #include "tesseractclass.h"
 
 #include "allheaders.h"
+#ifndef ANDROID_BUILD
 #include "cube_reco_context.h"
+#endif
 #include "edgblob.h"
 #include "equationdetect.h"
 #include "globals.h"
+#ifndef ANDROID_BUILD
 #include "tesseract_cube_combiner.h"
+#endif
 
 // Include automatically generated configuration file if running autoconf.
 #ifdef HAVE_CONFIG_H
@@ -547,8 +551,10 @@ Tesseract::Tesseract()
     reskew_(1.0f, 0.0f),
     most_recently_used_(this),
     font_table_size_(0),
+#ifndef ANDROID_BUILD
     cube_cntxt_(NULL),
     tess_cube_combiner_(NULL),
+#endif
     equ_detect_(NULL) {
 }
 
@@ -556,6 +562,7 @@ Tesseract::~Tesseract() {
   Clear();
   end_tesseract();
   sub_langs_.delete_data_pointers();
+#ifndef ANDROID_BUILD
   // Delete cube objects.
   if (cube_cntxt_ != NULL) {
     delete cube_cntxt_;
@@ -565,6 +572,7 @@ Tesseract::~Tesseract() {
     delete tess_cube_combiner_;
     tess_cube_combiner_ = NULL;
   }
+#endif
 }
 
 void Tesseract::Clear() {

ccmain/tesseractclass.h

@@ -97,12 +97,16 @@ class WERD_RES;
 namespace tesseract {
 
 class ColumnFinder;
+#ifndef ANDROID_BUILD
 class CubeLineObject;
 class CubeObject;
 class CubeRecoContext;
+#endif
 class EquationDetect;
 class Tesseract;
+#ifndef ANDROID_BUILD
 class TesseractCubeCombiner;
+#endif
 
 // A collection of various variables for statistics and debugging.
 struct TesseractStats {
@@ -382,6 +386,7 @@ class Tesseract : public Wordrec {
                              int *right_ok) const;
 
   //// cube_control.cpp ///////////////////////////////////////////////////
+#ifndef ANDROID_BUILD
   bool init_cube_objects(bool load_combiner,
                          TessdataManager *tessdata_manager);
   // Iterates through tesseract's results and calls cube on each word,
@@ -407,6 +412,7 @@ class Tesseract : public Wordrec {
                           Boxa** char_boxes, CharSamp*** char_samples);
   bool create_cube_box_word(Boxa *char_boxes, int num_chars,
                             TBOX word_box, BoxWord* box_word);
+#endif
   //// output.h //////////////////////////////////////////////////////////
 
   void output_pass(PAGE_RES_IT &page_res_it, const TBOX *target_word_box);
@@ -713,8 +719,8 @@ class Tesseract : public Wordrec {
   // Creates a fake best_choice entry in each WERD_RES with the correct text.
   void CorrectClassifyWords(PAGE_RES* page_res);
   // Call LearnWord to extract features for labelled blobs within each word.
-  // Features are written to the given filename.
-  void ApplyBoxTraining(const STRING& filename, PAGE_RES* page_res);
+  // Features are stored in an internal buffer.
+  void ApplyBoxTraining(const STRING& fontname, PAGE_RES* page_res);
 
   //// fixxht.cpp ///////////////////////////////////////////////////////
   // Returns the number of misfit blob tops in this word.
@@ -1089,7 +1095,9 @@ class Tesseract : public Wordrec {
                                   PAGE_RES_IT* pr_it,
                                   FILE *output_file);
 
+#ifndef ANDROID_BUILD
   inline CubeRecoContext *GetCubeRecoContext() { return cube_cntxt_; }
+#endif
 
  private:
   // The filename of a backup config file. If not null, then we currently
@@ -1129,9 +1137,11 @@ class Tesseract : public Wordrec {
   Tesseract* most_recently_used_;
   // The size of the font table, ie max possible font id + 1.
   int font_table_size_;
+#ifndef ANDROID_BUILD
   // Cube objects.
   CubeRecoContext* cube_cntxt_;
   TesseractCubeCombiner *tess_cube_combiner_;
+#endif
   // Equation detector. Note: this pointer is NOT owned by the class.
   EquationDetect* equ_detect_;
 };

ccstruct/imagedata.cpp

@@ -51,6 +51,7 @@ void WordFeature::ComputeSize(const GenericVector<WordFeature>& features,
 // Draws the features in the given window.
 void WordFeature::Draw(const GenericVector<WordFeature>& features,
                        ScrollView* window) {
+#ifndef GRAPHICS_DISABLED
   for (int f = 0; f < features.size(); ++f) {
     FCOORD pos(features[f].x_, features[f].y_);
     FCOORD dir;
@@ -61,6 +62,7 @@ void WordFeature::Draw(const GenericVector<WordFeature>& features,
     window->DrawTo(IntCastRounded(pos.x() + dir.x()),
                       IntCastRounded(pos.y() + dir.y()));
   }
+#endif
 }
 
 // Writes to the given file. Returns false in case of error.
@@ -244,6 +246,7 @@ int ImageData::MemoryUsed() const {
 
 // Draws the data in a new window.
 void ImageData::Display() const {
+#ifndef GRAPHICS_DISABLED
   const int kTextSize = 64;
   // Draw the image.
   Pix* pix = GetPix();
@@ -274,6 +277,7 @@ void ImageData::Display() const {
   win->Pen(ScrollView::GREEN);
   win->Update();
   window_wait(win);
+#endif
 }
 
 // Adds the supplied boxes and transcriptions that correspond to the correct

textord/drawedg.h

@@ -19,6 +19,7 @@
 
 #ifndef           DRAWEDG_H
 #define           DRAWEDG_H
+#ifndef ANDROID_BUILD
 
 #include          "scrollview.h"
 #include          "crakedge.h"
@@ -32,3 +33,4 @@ void draw_raw_edge(                   //draw the cracks
                    ScrollView::Color colour      //colour to draw in
                   );
 #endif
+#endif