Method fix_rep_char

deps/tesseract/ccmain/control.cpp:1629–1666

* fix_rep_char() * The word is a repeated char. (Leader.) Find the repeated char character. * Create the appropriate single-word or multi-word sequence according to * the size of spaces in between blobs, and correct the classifications * where some of the characters disagree with the majority. */

 * where some of the characters disagree with the majority.
 */
void Tesseract::fix_rep_char(PAGE_RES_IT* page_res_it) {
  WERD_RES *word_res = page_res_it->word();
  const WERD_CHOICE &word = *(word_res->best_choice);

  // Find the frequency of each unique character in the word.
  SortHelper<UNICHAR_ID> rep_ch(word.length());
  for (int i = 0; i < word.length(); ++i) {
    rep_ch.Add(word.unichar_id(i), 1);
  }

  // Find the most frequent result.
  UNICHAR_ID maxch_id = INVALID_UNICHAR_ID; // most common char
  int max_count = rep_ch.MaxCount(&maxch_id);
  // Find the best exemplar of a classifier result for maxch_id.
  BLOB_CHOICE* best_choice = FindBestMatchingChoice(maxch_id, word_res);
  if (best_choice == NULL) {
    tprintf("Failed to find a choice for %s, occurring %d times\n",
            word_res->uch_set->debug_str(maxch_id).string(), max_count);
    return;
  }
  word_res->done = TRUE;

  // Measure the mean space.
  int gap_count = 0;
  WERD* werd = word_res->word;
  C_BLOB_IT blob_it(werd->cblob_list());
  C_BLOB* prev_blob = blob_it.data();
  for (blob_it.forward(); !blob_it.at_first(); blob_it.forward()) {
    C_BLOB* blob = blob_it.data();
    int gap = blob->bounding_box().left();
    gap -= prev_blob->bounding_box().right();
    ++gap_count;
    prev_blob = blob;
  }
  // Just correct existing classification.
  CorrectRepcharChoices(best_choice, word_res);
  word_res->reject_map.initialise(word.length());
}

ACCEPTABLE_WERD_TYPE Tesseract::acceptable_word_string(
    const UNICHARSET& char_set, const char *s, const char *lengths) {