// Reads all boxes from the string. Otherwise, as ReadAllBoxes.
bool ReadMemBoxes(int target_page, bool skip_blanks, const char* box_data,
GenericVector<TBOX>* boxes,
GenericVector<STRING>* texts,
GenericVector<STRING>* box_texts,
GenericVector<int>* pages) {
STRING box_str(box_data);
GenericVector<STRING> lines;
box_str.split('\n', &lines);
if (lines.empty()) return false;
int num_boxes = 0;
for (int i = 0; i < lines.size(); ++i) {
int page = 0;
STRING utf8_str;
TBOX box;
if (!ParseBoxFileStr(lines[i].string(), &page, &utf8_str, &box)) {
continue;
}
if (skip_blanks && (utf8_str == " " || utf8_str == "\t")) continue;
if (target_page >= 0 && page != target_page) continue;
if (boxes != NULL) boxes->push_back(box);
if (texts != NULL) texts->push_back(utf8_str);
if (box_texts != NULL) {
STRING full_text;
MakeBoxFileStr(utf8_str.string(), box, target_page, &full_text);
box_texts->push_back(full_text);
}
if (pages != NULL) pages->push_back(page);
++num_boxes;
}
return num_boxes > 0;
}
// Returns a string debug representation of the given sample:
// font, unichar_str, bounding box, page.
STRING TrainingSampleSet::SampleToString(const TrainingSample &sample) const {
STRING boxfile_str;
MakeBoxFileStr(unicharset_.id_to_unichar(sample.class_id()),
sample.bounding_box(), sample.page_num(), &boxfile_str);
return STRING(fontinfo_table_.get(sample.font_id()).name) + " " + boxfile_str;
}
