Function MarkRowsWithModel

deps/tesseract/ccmain/paragraphs.cpp:807–844

Given a paragraph model mark rows[row_start, row_end) as said model start or body lines. Case 1: model->first_indent_ != model->body_indent_ Differentiating the paragraph start lines from the paragraph body lines in this case is easy, we just see how far each line is indented. Case 2: model->first_indent_ == model->body_indent_ Here, we find end-of-paragraph lines by looking for "short lines." W

//     We mark a line as short (end of paragraph) if the offside indent
//     is greater than eop_threshold.
void MarkRowsWithModel(GenericVector<RowScratchRegisters> *rows,
                       int row_start, int row_end,
                       const ParagraphModel *model,
                       bool ltr,
                       int eop_threshold) {
  if (!AcceptableRowArgs(0, 0, __func__, rows, row_start, row_end))
    return;
  for (int row = row_start; row < row_end; row++) {
    bool valid_first = ValidFirstLine(rows, row, model);
    bool valid_body = ValidBodyLine(rows, row, model);
    if (valid_first && !valid_body) {
      (*rows)[row].AddStartLine(model);
    } else if (valid_body && !valid_first) {
      (*rows)[row].AddBodyLine(model);
    } else if (valid_body && valid_first) {
      bool after_eop = (row == row_start);
      if (row > row_start) {
        if (eop_threshold > 0) {
          if (model->justification() == JUSTIFICATION_LEFT) {
            after_eop = (*rows)[row - 1].rindent_ > eop_threshold;
          } else {
            after_eop = (*rows)[row - 1].lindent_ > eop_threshold;
          }
        } else {
         after_eop = FirstWordWouldHaveFit((*rows)[row - 1], (*rows)[row],
                                           model->justification());
        }
      }
      if (after_eop) {
        (*rows)[row].AddStartLine(model);
      } else {
        (*rows)[row].AddBodyLine(model);
      }
    } else {
      // Do nothing. Stray row.
    }
  }
}

// GeometricClassifierState holds all of the information we'll use while
// trying to determine a paragraph model for the text lines in a block of