combine_lang_model.cpp File Reference

#include "commandlineflags.h"
#include "commontraining.h"
#include "lang_model_helpers.h"
#include "tprintf.h"
#include "unicharset_training_utils.h"

Go to the source code of this file.

Functions
int	main (int argc, char **argv)

Function Documentation

main()

int main	(	int	argc,
		char **	argv
	)

Definition at line 41 of file combine_lang_model.cpp.

                                {
  // Sets properties on the input unicharset file, and writes:
  //   rootdir/lang/lang.charset_size=ddd.txt
  //   rootdir/lang/lang.traineddata
  //   rootdir/lang/lang.unicharset
  // If the 3 word lists are provided, the dawgs are also added
  // to the traineddata file.
  // The output unicharset and charset_size files are just for
  // human readability.
  tesseract::CheckSharedLibraryVersion();
  tesseract::ParseCommandLineFlags(argv[0], &argc, &argv, true);
 
  // If these reads fail, we get a warning message and an empty list of words.
  std::vector<std::string> words = split(tesseract::ReadFile(FLAGS_words.c_str()), '\n');
  std::vector<std::string> puncs = split(tesseract::ReadFile(FLAGS_puncs.c_str()), '\n');
  std::vector<std::string> numbers = split(tesseract::ReadFile(FLAGS_numbers.c_str()), '\n');
  // Load the input unicharset
  UNICHARSET unicharset;
  if (!unicharset.load_from_file(FLAGS_input_unicharset.c_str(), false)) {
    tprintf("Failed to load unicharset from %s\n", FLAGS_input_unicharset.c_str());
    return EXIT_FAILURE;
  }
  tprintf("Loaded unicharset of size %zu from file %s\n", unicharset.size(),
          FLAGS_input_unicharset.c_str());
 
  // Set unichar properties
  tprintf("Setting unichar properties\n");
  tesseract::SetupBasicProperties(/*report_errors*/ true,
                                  /*decompose (NFD)*/ false, &unicharset);
  tprintf("Setting script properties\n");
  tesseract::SetScriptProperties(FLAGS_script_dir.c_str(), &unicharset);
  // Combine everything into a traineddata file.
  return tesseract::CombineLangModel(unicharset, FLAGS_script_dir.c_str(),
                                     FLAGS_version_str.c_str(), FLAGS_output_dir.c_str(),
                                     FLAGS_lang.c_str(), FLAGS_pass_through_recoder, words, puncs,
                                     numbers, FLAGS_lang_is_rtl, /*reader*/ nullptr,
                                     /*writer*/ nullptr);
}