show page 0 for multipage tiff;

Windows: use binary mode for fopen (issue 70);
autotools: fixed cutil/Makefile.am, improved tessdata/Makefile.am;

git-svn-id: https://tesseract-ocr.googlecode.com/svn/trunk@604 d0cd1f9f-072b-0410-8dd7-cf729c803f20

api/baseapi.cpp

@@ -440,7 +440,7 @@ Boxa* TessBaseAPI::GetComponentImages(PageIteratorLevel level,
 void TessBaseAPI::DumpPGM(const char* filename) {
   if (tesseract_ == NULL)
     return;
-  FILE *fp = fopen(filename, "w");
+  FILE *fp = fopen(filename, "wb");
   Pix* pix = tesseract_->pix_binary();
   int width = pixGetWidth(pix);
   int height = pixGetHeight(pix);
@@ -644,7 +644,7 @@ bool TessBaseAPI::ProcessPages(const char* filename,
   if (npages > 0) {
     for (; page < npages && (pix = pixReadTiff(filename, page)) != NULL;
          ++page) {
-      if (page > 0)
+      if (page >= 0)
         tprintf(_("Page %d\n"), page);
       char page_str[kMaxIntSize];
       snprintf(page_str, kMaxIntSize - 1, "%d", page);
@@ -665,7 +665,7 @@ bool TessBaseAPI::ProcessPages(const char* filename,
       pixDestroy(&pix);
     } else {
       // The file is not an image file, so try it as a list of filenames.
-      FILE* fimg = fopen(filename, "r");
+      FILE* fimg = fopen(filename, "rb");
       if (fimg == NULL) {
         tprintf(_("File %s cannot be opened!\n"), filename);
         return false;
@@ -742,7 +742,7 @@ bool TessBaseAPI::ProcessPage(Pix* pix, int page_index, const char* filename,
   }
   if (failed && retry_config != NULL && retry_config[0] != '\0') {
     // Save current config variables before switching modes.
-    FILE* fp = fopen(kOldVarsFile, "w");
+    FILE* fp = fopen(kOldVarsFile, "wb");
     PrintVariables(fp);
     fclose(fp);
     // Switch to alternate mode for retry.

api/tesseractmain.cpp

@@ -110,7 +110,7 @@ int main(int argc, char **argv) {
            tesseract::TessBaseAPI::Version());
 
   
-  FILE* fin = fopen(image, "r");
+  FILE* fin = fopen(image, "rb");
   if (fin == NULL) {
     printf("Cannot open input file: %s\n", image);
     exit(2);
@@ -134,7 +134,7 @@ int main(int argc, char **argv) {
   api.GetBoolVariable("tessedit_create_boxfile", &output_box);
   STRING outfile = output;
   outfile += output_hocr ? ".html" : output_box ? ".box" : ".txt";
-  FILE* fout = fopen(outfile.string(), "w");
+  FILE* fout = fopen(outfile.string(), "wb");
   if (fout == NULL) {
     tprintf(_("Cannot create output file %s\n"), outfile.string());
     exit(1);

ccmain/paramsd.cpp

@@ -325,7 +325,7 @@ void ParamsEditor::WriteParams(char *filename,
   FILE *fp;                      // input file
   char msg_str[255];
                                  // if file exists
-  if ((fp = fopen (filename, "r")) != NULL) {
+  if ((fp = fopen (filename, "rb")) != NULL) {
     fclose(fp);
     sprintf (msg_str, "Overwrite file " "%s" "? (Y/N)", filename);
     int a = sv_window_->ShowYesNoDialog(msg_str);
@@ -333,7 +333,7 @@ void ParamsEditor::WriteParams(char *filename,
   }
 
 
-  fp = fopen (filename, "w");  // can we write to it?
+  fp = fopen (filename, "wb");  // can we write to it?
   if (fp == NULL) {
     sv_window_->AddMessage("Cant write to file " "%s" "", filename);
     return;

ccmain/tessedit.cpp

@@ -69,13 +69,13 @@ void Tesseract::read_config_file(const char *filename, bool init_only) {
   path += "configs/";
   path += filename;
   FILE* fp;
-  if ((fp = fopen(path.string(), "r")) != NULL) {
+  if ((fp = fopen(path.string(), "rb")) != NULL) {
     fclose(fp);
   } else {
     path = datadir;
     path += "tessconfigs/";
     path += filename;
-    if ((fp = fopen(path.string(), "r")) != NULL) {
+    if ((fp = fopen(path.string(), "rb")) != NULL) {
       fclose(fp);
     } else {
       path = filename;
@@ -148,7 +148,7 @@ bool Tesseract::init_tesseract_lang_data(
   }
 
   if (((STRING &)tessedit_write_params_to_file).length() > 0) {
-    FILE *params_file = fopen(tessedit_write_params_to_file.string(), "w");
+    FILE *params_file = fopen(tessedit_write_params_to_file.string(), "wb");
     if (params_file != NULL) {
       ParamUtils::PrintParams(params_file, this->params());
       fclose(params_file);

ccmain/tesseract_cube_combiner.cpp

@@ -57,7 +57,7 @@ bool TesseractCubeCombiner::LoadCombinerNet() {
                           ".tesseract_cube.nn";
 
   // Return false if file does not exist
-  FILE *fp = fopen(net_file_name.c_str(), "r");
+  FILE *fp = fopen(net_file_name.c_str(), "rb");
   if (fp == NULL)
     return false;
   else

ccstruct/blread.cpp

@@ -49,7 +49,7 @@ bool read_unlv_file(                    //print list of sides
   BLOCK_IT block_it = blocks;    //block iterator
 
   name += UNLV_EXT;              //add extension
-  if ((pdfp = fopen (name.string (), "r")) == NULL) {
+  if ((pdfp = fopen (name.string (), "rb")) == NULL) {
     return false;                //didn't read one
   }
   else {

ccutil/boxread.cpp

@@ -35,7 +35,7 @@ FILE* OpenBoxFile(const STRING& fname) {
 
   filename += ".box";
   FILE* box_file = NULL;
-  if (!(box_file = fopen(filename.string(), "r"))) {
+  if (!(box_file = fopen(filename.string(), "rb"))) {
     CANTOPENFILE.error("read_next_box", TESSEXIT,
                        "Cant open box file %s",
                        filename.string());

ccutil/params.cpp

@@ -58,7 +58,7 @@ bool ParamUtils::ReadParamsFile(const char *file, bool init_only,
     nameoffset = 0;
   }
 
-  fp = fopen(file + nameoffset, "r");
+  fp = fopen(file + nameoffset, "rb");
   if (fp == NULL) {
     tprintf("read_params_file: Can't open %s\n", file + nameoffset);
     return true;

ccutil/tessdatamanager.cpp

@@ -128,7 +128,8 @@ bool TessdataManager::CombineDataFiles(
         kTessdataFileSuffixes[i], &type, &text_file));
     STRING filename = language_data_path_prefix;
     filename += kTessdataFileSuffixes[i];
-    file_ptr[i] =  fopen(filename.string(), text_file ? "r" : "rb");
+    // file_ptr[i] =  fopen(filename.string(), text_file ? "r" : "rb");
+    file_ptr[i] =  fopen(filename.string(), "rb");
     if (file_ptr[i] != NULL) {
       offset_table[type] = ftell(output_file);
       CopyFile(file_ptr[i], output_file, text_file, -1);
@@ -181,7 +182,8 @@ bool TessdataManager::OverwriteComponents(
   // Open the files with the new components.
   for (i = 0; i < num_new_components; ++i) {
     TessdataTypeFromFileName(component_filenames[i], &type, &text_file);
-    file_ptr[type] = fopen(component_filenames[i], text_file ? "r" : "rb");
+    // file_ptr[type] = fopen(component_filenames[i], text_file ? "r" : "rb");
+    file_ptr[type] = fopen(component_filenames[i], "rb");
   }
 
   // Write updated data to the output traineddata file.

ccutil/tprintf.cpp

@@ -64,7 +64,7 @@ const char *format, ...          //special message
   va_end(args);
 
   if (debugfp == NULL && strlen (debug_file.string ()) > 0)
-    debugfp = fopen (debug_file.string (), "w");
+    debugfp = fopen (debug_file.string (), "wb");
   else if (debugfp != NULL && strlen (debug_file.string ()) == 0) {
     fclose(debugfp);
     debugfp = NULL;

ccutil/unicharset.h

@@ -226,7 +226,7 @@ class UNICHARSET {
   // Opens the file indicated by filename and saves unicharset to that file.
   // Returns true if the operation is successful.
   bool save_to_file(const char * const filename) const {
-    FILE* file = fopen(filename, "w+");
+    FILE* file = fopen(filename, "w+b");
     if (file == NULL) return false;
     bool result = save_to_file(file);
     fclose(file);
@@ -241,7 +241,7 @@ class UNICHARSET {
   // from the given file. The previous data is lost.
   // Returns true if the operation is successful.
   bool load_from_file(const char* const filename, bool skip_fragments) {
-    FILE* file = fopen(filename, "r");
+    FILE* file = fopen(filename, "rb");
     if (file == NULL) return false;
     bool result = load_from_file(file, skip_fragments);
     fclose(file);

classify/blobclass.cpp

@@ -91,7 +91,7 @@ void LearnBlob(const FEATURE_DEFS_STRUCT &FeatureDefs, const STRING& filename,
   // the name of the file is the name of the image plus TRAIN_SUFFIX
   if (FeatureFile == NULL) {
     Filename += TRAIN_SUFFIX;
-    FeatureFile = Efopen(Filename.string(), "w");
+    FeatureFile = Efopen(Filename.string(), "wb");
     cprintf("TRAINING ... Font name = %s\n", CurrFontName.string());
   }
 

classify/picofeat.cpp

@@ -78,16 +78,16 @@ FEATURE_SET Classify::ExtractPicoFeatures(TBLOB *Blob) {
   iterate(RemainingOutlines) {
     Outline = (MFOUTLINE) first_node (RemainingOutlines);
     /*---------Debug--------------------------------------------------*
-    OFile = fopen ("f:/ims/debug/pfOutline.logCPP", "r");
+    OFile = fopen ("f:/ims/debug/pfOutline.logCPP", "rb");
     if (OFile == NULL)
     {
-      OFile = Efopen ("f:/ims/debug/pfOutline.logCPP", "w");
+      OFile = Efopen ("f:/ims/debug/pfOutline.logCPP", "wb");
       WriteOutline(OFile, Outline);
     }
     else
     {
       fclose (OFile);
-      OFile = Efopen ("f:/ims/debug/pfOutline.logCPP", "a");
+      OFile = Efopen ("f:/ims/debug/pfOutline.logCPP", "ab");
     }
     WriteOutline(OFile, Outline);
     fclose (OFile);
@@ -97,16 +97,16 @@ FEATURE_SET Classify::ExtractPicoFeatures(TBLOB *Blob) {
   if (classify_norm_method == baseline)
     NormalizePicoX(FeatureSet);
   /*---------Debug--------------------------------------------------*
-  File = fopen ("f:/ims/debug/pfFeatSet.logCPP", "r");
+  File = fopen ("f:/ims/debug/pfFeatSet.logCPP", "rb");
   if (File == NULL)
   {
-    File = Efopen ("f:/ims/debug/pfFeatSet.logCPP", "w");
+    File = Efopen ("f:/ims/debug/pfFeatSet.logCPP", "wb");
     WriteFeatureSet(File, FeatureSet);
   }
   else
   {
     fclose (File);
-    File = Efopen ("f:/ims/debug/pfFeatSet.logCPP", "a");
+    File = Efopen ("f:/ims/debug/pfFeatSet.logCPP", "ab");
   }
   WriteFeatureSet(File, FeatureSet);
   fclose (File);

cube/cached_file.cpp

@@ -52,7 +52,7 @@ bool CachedFile::Open() {
     return true;
   }
 
-  fp_ = fopen(file_name_.c_str(), "r");
+  fp_ = fopen(file_name_.c_str(), "rb");
   if (fp_ == NULL) {
     return false;
   }

cube/char_samp_set.cpp

@@ -94,7 +94,7 @@ CharSampSet * CharSampSet::FromCharDumpFile(string file_name) {
   FILE *fp;
   unsigned int val32;
   // open the file
-  fp = fopen(file_name.c_str(), "r");
+  fp = fopen(file_name.c_str(), "rb");
   if (fp == NULL) {
     return NULL;
   }
@@ -123,7 +123,7 @@ FILE *CharSampSet::CreateCharDumpFile(string file_name) {
   FILE *fp;
   unsigned int val32;
   // create the file
-  fp =  fopen(file_name.c_str(), "w");
+  fp =  fopen(file_name.c_str(), "wb");
   if (!fp) {
     return NULL;
   }

cube/conv_net_classifier.cpp

@@ -227,7 +227,7 @@ bool ConvNetCharClassifier::LoadFoldingSets(const string &data_file_path,
   fold_file_name += ".cube.fold";
 
   // folding sets are optional
-  FILE *fp = fopen(fold_file_name.c_str(), "r");
+  FILE *fp = fopen(fold_file_name.c_str(), "rb");
   if (fp == NULL) {
     return true;
   }
@@ -320,7 +320,7 @@ bool ConvNetCharClassifier::LoadNets(const string &data_file_path,
   char_net_file += ".cube.nn";
 
   // neural network is optional
-  FILE *fp = fopen(char_net_file.c_str(), "r");
+  FILE *fp = fopen(char_net_file.c_str(), "rb");
   if (fp == NULL) {
     return true;
   }

cube/cube_tuning_params.cpp

@@ -187,7 +187,7 @@ bool CubeTuningParams::Load(string tuning_params_file) {
 
 // Save the parameters to a file
 bool CubeTuningParams::Save(string file_name) {
-  FILE *params_file = fopen(file_name.c_str(), "w");
+  FILE *params_file = fopen(file_name.c_str(), "wb");
   if (params_file == NULL) {
     fprintf(stderr, "Cube ERROR (CubeTuningParams::Save): error opening file "
             "%s for write.\n", file_name.c_str());

cube/cube_utils.cpp

@@ -267,7 +267,7 @@ unsigned char *CubeUtils::GetImageData(Pix *pix, int left, int top,
 // read file contents to a string
 bool CubeUtils::ReadFileToString(const string &file_name, string *str) {
   str->clear();
-  FILE *fp = fopen(file_name.c_str(), "r");
+  FILE *fp = fopen(file_name.c_str(), "rb");
   if (fp == NULL) {
     return false;
   }

cube/hybrid_neural_net_classifier.cpp

@@ -223,7 +223,7 @@ bool HybridNeuralNetCharClassifier::LoadFoldingSets(
   fold_file_name += ".cube.fold";
 
   // folding sets are optional
-  FILE *fp = fopen(fold_file_name.c_str(), "r");
+  FILE *fp = fopen(fold_file_name.c_str(), "rb");
   if (fp == NULL) {
     return true;
   }
@@ -316,7 +316,7 @@ bool HybridNeuralNetCharClassifier::LoadNets(const string &data_file_path,
   hybrid_net_file += ".cube.hybrid";
 
   // neural network is optional
-  FILE *fp = fopen(hybrid_net_file.c_str(), "r");
+  FILE *fp = fopen(hybrid_net_file.c_str(), "rb");
   if (fp == NULL) {
     return true;
   }

cutil/Makefile.am

@@ -1,5 +1,5 @@
 SUBDIRS =
-AM_CPPFLAGS = -I$(top_srcdir)/ccutil
+AM_CPPFLAGS = -I$(top_srcdir)/ccutil -I$(top_srcdir)/viewer
 
 include_HEADERS = \
     bitvec.h callcpp.h const.h cutil.h cutil_class.h danerror.h efio.h \

cutil/cutil.cpp

@@ -92,7 +92,7 @@ FILE *open_file(const char *filename, const char *mode) {
 bool exists_file(const char *filename) {
   bool exists = false;
   FILE *f = NULL;
-  if ((f = fopen(filename, "r")) != NULL) {
+  if ((f = fopen(filename, "rb")) != NULL) {
     fclose(f);
     exists = true;
   }

tessdata/Makefile.am

@@ -1,65 +1,6 @@
 datadir = @datadir@/tessdata
-data_DATA = ara.cube.bigrams \
+data_DATA =  *.traineddata \
-	ara.cube.fold \
+    *.cube.*
-	ara.cube.lm \
-	ara.cube.nn \
-	ara.cube.params \
-	ara.cube.size \
-	ara.cube.word-freq \
-	ara.traineddata \
-	eng.cube.bigrams \
-	eng.cube.lm \
-	eng.cube.params \
-	eng.cube.word-freq \
-	eng.traineddata \
-	eng.cube.fold \
-	eng.cube.nn \
-	eng.cube.size \
-	eng.tesseract_cube.nn \
-	bul.traineddata \
-	cat.traineddata \
-	ces.traineddata \
-	chi_sim.traineddata \
-	chi_tra.traineddata \
-	dan.traineddata \
-	dan-frak.traineddata \
-	deu.traineddata \
-	ell.traineddata \
-	fin.traineddata \
-	fra.traineddata \
-	heb.traineddata \
-  hin.cube.bigrams \
-  hin.cube.lm \
-  hin.cube.params \
-  hin.cube.word-freq \
-  hin.tesseract_cube.nn \
-  hin.traineddata \
-	hrv.traineddata \
-	hun.traineddata \
-	ind.traineddata \
-	ita.traineddata \
-	jpn.traineddata \
-	kor.traineddata \
-	lav.traineddata \
-	lit.traineddata \
-	nld.traineddata \
-	nor.traineddata \
-	osd.traineddata \
-	pol.traineddata \
-	por.traineddata \
-	ron.traineddata \
-	rus.traineddata \
-	slk.traineddata \
-	slk-frak.traineddata \
-	slv.traineddata \
-	spa.traineddata \
-	srp.traineddata \
-	swe.traineddata \
-	tgl.traineddata \
-	tha.traineddata \
-	tur.traineddata \
-	ukr.traineddata \
-	vie.traineddata
 
 SUBDIRS = configs tessconfigs
 

textord/tablefind.cpp

@@ -2028,7 +2028,7 @@ void TableFinder::WriteToPix(const FCOORD& reskew) {
 
   BOXA* table_array = boxaCreate(num_boxes);
   // text file containing detected table bounding boxes
-  FILE* fptr = fopen("tess-table.txt", "w");
+  FILE* fptr = fopen("tess-table.txt", "wb");
   GridSearch<ColSegment, ColSegment_CLIST, ColSegment_C_IT>
       table_search(&table_grid_);
   table_search.StartFullSearch();

training/cntraining.cpp

@@ -155,7 +155,7 @@ int main (
   int num_fonts = 0;
   while ((PageName = GetNextFilename(argc, argv)) != NULL) {
     printf("Reading %s ...\n", PageName);
-    TrainingPage = Efopen(PageName, "r");
+    TrainingPage = Efopen(PageName, "rb");
     ReadTrainingSamples(FeatureDefs, PROGRAM_FEATURE_TYPE,
                         100, 1.0f / 64.0f, 0.0f, NULL, TrainingPage, &CharList);
     fclose(TrainingPage);
@@ -238,7 +238,7 @@ void WriteNormProtos (
   }
   strcat (Filename, "normproto");
   printf ("\nWriting %s ...", Filename);
-  File = Efopen (Filename, "w");
+  File = Efopen (Filename, "wb");
   fprintf(File,"%0d\n",Clusterer->SampleSize);
   WriteParamDesc(File,Clusterer->SampleSize,Clusterer->ParamDesc);
   iterate(LabeledProtoList)

training/commontraining.cpp

@@ -370,10 +370,10 @@ void WriteTrainingSamples (
 
     /* if file does not exist, create a new one with an appropriate
        header; otherwise append samples to the existing file */
-    File = fopen (Filename, "r");
+    File = fopen (Filename, "rb");
     if (File == NULL)
     {
-      File = Efopen (Filename, "w");
+      File = Efopen (Filename, "wb");
       WriteOldParamDesc(
           File,
           FeatureDefs.FeatureDesc[ShortNameToFeatureType(
@@ -382,7 +382,7 @@ void WriteTrainingSamples (
     else
     {
       fclose (File);
-      File = Efopen (Filename, "a");
+      File = Efopen (Filename, "ab");
     }
 
     // append samples onto the file

training/mftraining.cpp

@@ -225,7 +225,7 @@ int main (int argc, char **argv) {
 
   // Populate fontinfo_table with font properties.
   if (InputFontInfoFile != NULL) {
-    FILE* f = fopen(InputFontInfoFile, "r");
+    FILE* f = fopen(InputFontInfoFile, "rb");
     if (f == NULL) {
       fprintf(stderr, "Failed to load font_properties\n");
     } else {
@@ -292,7 +292,7 @@ int main (int argc, char **argv) {
       delete[] short_name;
     }
 
-    TrainingPage = Efopen (PageName, "r");
+    TrainingPage = Efopen (PageName, "rb");
     LIST char_list = NIL_LIST;
     ReadTrainingSamples(FeatureDefs, PROGRAM_FEATURE_TYPE,
                         0, 1.0f / 128.0f, 1.0f / 64.0f, &unicharset_training,
@@ -371,11 +371,7 @@ int main (int argc, char **argv) {
     strcat (Filename, "/");
   }
   strcat (Filename, "inttemp");
-#ifdef __UNIX__
-  OutFile = Efopen (Filename, "w");
-#else
   OutFile = Efopen (Filename, "wb");
-#endif
   classify->WriteIntTemplates(OutFile, IntTemplates, unicharset_training);
   fclose (OutFile);
   strcpy (Filename, "");
@@ -425,7 +421,7 @@ void WriteMicrofeat(
     strcat (Filename, "/");
   }
   strcat (Filename, "Microfeat");
-  File = Efopen (Filename, "w");
+  File = Efopen (Filename, "wb");
   printf ("\nWriting Merged %s ...", Filename);
   iterate(ClassList)
   {
@@ -516,7 +512,7 @@ void InitXHeights(const char *filename,
                    int xheights[]) {
   for (int i = 0; i < fontinfo_table.size(); ++i) xheights[i] = -1;
   if (filename == NULL) return;
-  FILE *f = fopen(filename, "r");
+  FILE *f = fopen(filename, "rb");
   if (f == NULL) {
     fprintf(stderr, "Failed to load font xheights from %s\n", filename);
     return;
@@ -542,7 +538,7 @@ void AddSpacingInfo(const char *filename,
                      const UNICHARSET &unicharset,
                      const int xheights[],
                      UnicityTable<FontInfo> *fontinfo_table) {
-  FILE* fontinfo_file = fopen(filename, "r");
+  FILE* fontinfo_file = fopen(filename, "rb");
   if (fontinfo_file == NULL) return;
   tprintf("Reading spacing from %s ...\n", filename);
   int scale = kBlnXHeight / xheights[fontinfo_id];

training/unicharset_extractor.cpp

@@ -134,7 +134,7 @@ int main(int argc, char** argv) {
   for (; tessoptind < argc; ++tessoptind) {
     printf("Extracting unicharset from %s\n", argv[tessoptind]);
 
-    FILE* box_file = fopen(argv[tessoptind], "r");
+    FILE* box_file = fopen(argv[tessoptind], "rb");
     if (box_file == NULL) {
       printf("Cannot open box file %s\n", argv[tessoptind]);
       return -1;

training/wordlist2dawg.cpp

@@ -91,7 +91,7 @@ int main(int argc, char** argv) {
   } else if (argc == 7) {
     // Place words of different lengths in separate Dawgs.
     char str[CHARS_PER_LINE];
-    FILE *word_file = fopen(wordlist_filename, "r");
+    FILE *word_file = fopen(wordlist_filename, "rb");
     if (word_file == NULL) {
       tprintf("Failed to open wordlist file %s\n", wordlist_filename);
       exit(1);

wordrec/drawfx.cpp

@@ -95,5 +95,5 @@ void create_fxdebug_win() {  //make gradients win
   //                      DEBUG_XPOS,DEBUG_YPOS,
   //                      DEBUG_XSIZE,DEBUG_YSIZE);
   //      else
-  //              fx_debug=fopen(fx_debugfile.string(),"w");
+  //              fx_debug=fopen(fx_debugfile.string(),"wb");
 }