// Sets all the properties for this unicharset given a src unicharset with
// everything set. The unicharsets don't have to be the same, and graphemes
// are correctly accounted for.
void UNICHARSET::PartialSetPropertiesFromOther(int start_index,
const UNICHARSET& src) {
for (int ch = start_index; ch < size_used; ++ch) {
const char* utf8 = id_to_unichar(ch);
UNICHAR_PROPERTIES properties;
if (src.GetStrProperties(utf8, &properties)) {
// Setup the script_id, other_case, and mirror properly.
const char* script = src.get_script_from_script_id(properties.script_id);
properties.script_id = add_script(script);
const char* other_case = src.id_to_unichar(properties.other_case);
if (contains_unichar(other_case)) {
properties.other_case = unichar_to_id(other_case);
} else {
properties.other_case = ch;
}
const char* mirror_str = src.id_to_unichar(properties.mirror);
if (contains_unichar(mirror_str)) {
properties.mirror = unichar_to_id(mirror_str);
} else {
properties.mirror = ch;
}
unichars[ch].properties.CopyFrom(properties);
set_normed_ids(ch);
} else {
tprintf("Failed to get properties for index %d = %s\n", ch, utf8);
}
}
}
/**
* print_ratings_info
*
* Send all the ratings out to the logfile.
*
* @param fp file to use
* @param ratings list of results
* @param current_unicharset unicharset that can be used
* for id-to-unichar conversion
*/
void print_ratings_info(FILE *fp,
BLOB_CHOICE_LIST *ratings,
const UNICHARSET ¤t_unicharset) {
inT32 index; // to list
inT32 best_index; // to list
FLOAT32 best_rat; // rating
FLOAT32 best_cert; // certainty
const char* first_char = NULL; // character
FLOAT32 first_rat; // rating
FLOAT32 first_cert; // certainty
const char* sec_char = NULL; // character
FLOAT32 sec_rat = 0.0f; // rating
FLOAT32 sec_cert = 0.0f; // certainty
BLOB_CHOICE_IT c_it = ratings; // iterator
index = ratings->length();
if (index > 0) {
first_char = current_unicharset.id_to_unichar(c_it.data()->unichar_id());
first_rat = c_it.data()->rating();
first_cert = -c_it.data()->certainty();
if (index > 1) {
sec_char = current_unicharset.id_to_unichar(
c_it.data_relative(1)->unichar_id());
sec_rat = c_it.data_relative(1)->rating();
sec_cert = -c_it.data_relative(1)->certainty();
} else {
sec_char = NULL;
sec_rat = -1;
sec_cert = -1;
}
} else {
first_char = NULL;
first_rat = -1;
first_cert = -1;
}
best_index = -1;
best_rat = -1;
best_cert = -1;
for (index = 0, c_it.mark_cycle_pt(); !c_it.cycled_list();
c_it.forward(), index++) {
if (strcmp(current_unicharset.id_to_unichar(c_it.data()->unichar_id()),
blob_answer) == 0) {
best_index = index;
best_rat = c_it.data()->rating();
best_cert = -c_it.data()->certainty();
}
}
if (first_char != NULL && (*first_char == '\0' || *first_char == ' '))
first_char = NULL;
if (sec_char != NULL && (*sec_char == '\0' || *sec_char == ' '))
sec_char = NULL;
fprintf(matcher_fp,
" " INT32FORMAT " " INT32FORMAT " %g %g %s %g %g %s %g %g\n",
ratings->length(), best_index, best_rat, best_cert,
first_char != NULL ? first_char : "~",
first_rat, first_cert, sec_char != NULL ? sec_char : "~",
sec_rat, sec_cert);
}
// Print the best guesses out of the match rating matrix.
void MATRIX::print(const UNICHARSET &unicharset) const {
tprintf("Ratings Matrix (top 3 choices)\n");
int dim = dimension();
int band_width = bandwidth();
int row, col;
for (col = 0; col < dim; ++col) {
for (row = col; row < dim && row < col + band_width; ++row) {
BLOB_CHOICE_LIST *rating = this->get(col, row);
if (rating == NOT_CLASSIFIED) continue;
BLOB_CHOICE_IT b_it(rating);
tprintf("col=%d row=%d ", col, row);
for (b_it.mark_cycle_pt(); !b_it.cycled_list(); b_it.forward()) {
tprintf("%s rat=%g cert=%g " ,
unicharset.id_to_unichar(b_it.data()->unichar_id()),
b_it.data()->rating(), b_it.data()->certainty());
}
tprintf("\n");
}
tprintf("\n");
}
tprintf("\n");
for (col = 0; col < dim; ++col) tprintf("\t%d", col);
tprintf("\n");
for (row = 0; row < dim; ++row) {
for (col = 0; col <= row; ++col) {
if (col == 0) tprintf("%d\t", row);
if (row >= col + band_width) {
tprintf(" \t");
continue;
}
BLOB_CHOICE_LIST *rating = this->get(col, row);
if (rating != NOT_CLASSIFIED) {
BLOB_CHOICE_IT b_it(rating);
int counter = 0;
for (b_it.mark_cycle_pt(); !b_it.cycled_list(); b_it.forward()) {
tprintf("%s ",
unicharset.id_to_unichar(b_it.data()->unichar_id()));
++counter;
if (counter == 3) break;
}
tprintf("\t");
} else {
tprintf(" \t");
}
}
tprintf("\n");
}
}
// Print the best guesses out of the match rating matrix.
void MATRIX::print(const UNICHARSET &unicharset) {
tprintf("Ratings Matrix (top choices)\n");
int row, col;
for (col = 0; col < this->dimension(); ++col) tprintf("\t%d", col);
tprintf("\n");
for (row = 0; row < this->dimension(); ++row) {
for (col = 0; col <= row; ++col) {
if (col == 0) tprintf("%d\t", row);
BLOB_CHOICE_LIST *rating = this->get(col, row);
if (rating != NOT_CLASSIFIED) {
BLOB_CHOICE_IT b_it(rating);
int counter = 0;
for (b_it.mark_cycle_pt(); !b_it.cycled_list(); b_it.forward()) {
tprintf("%s ", unicharset.id_to_unichar(b_it.data()->unichar_id()));
++counter;
if (counter == 3) break;
}
tprintf("\t");
} else {
tprintf(" \t");
}
}
tprintf("\n");
}
}
// For each id in src, if it does not occur in this, add it, as in
// SetPropertiesFromOther, otherwise expand the ranges, as in
// ExpandRangesFromOther.
void UNICHARSET::AppendOtherUnicharset(const UNICHARSET& src) {
int initial_used = size_used;
for (int ch = 0; ch < src.size_used; ++ch) {
const UNICHAR_PROPERTIES& src_props = src.unichars[ch].properties;
const char* utf8 = src.id_to_unichar(ch);
if (strcmp(utf8, " ") != 0 && src_props.AnyRangeEmpty()) {
// Only use fully valid entries.
tprintf("Bad properties for index %d, char %s: "
"%d,%d %d,%d %d,%d %d,%d %d,%d\n",
ch, utf8, src_props.min_bottom, src_props.max_bottom,
src_props.min_top, src_props.max_top,
src_props.min_width, src_props.max_width,
src_props.min_bearing, src_props.max_bearing,
src_props.min_advance, src_props.max_advance);
continue;
}
int id = size_used;
if (contains_unichar(utf8)) {
id = unichar_to_id(utf8);
// Just expand current ranges.
unichars[id].properties.ExpandRangesFrom(src_props);
} else {
unichar_insert(utf8);
unichars[id].properties.SetRangesEmpty();
}
}
// Set properties, including mirror and other_case, WITHOUT reordering
// the unicharset.
PartialSetPropertiesFromOther(initial_used, src);
}
/**
* print_ratings_info
*
* Send all the ratings out to the logfile.
*
* @param fp file to use
* @param ratings list of results
* @param current_unicharset unicharset that can be used
* for id-to-unichar conversion
*/
void print_ratings_info(FILE *fp,
BLOB_CHOICE_LIST *ratings,
const UNICHARSET ¤t_unicharset) {
inT32 index; // to list
const char* first_char = NULL; // character
FLOAT32 first_rat; // rating
FLOAT32 first_cert; // certainty
const char* sec_char = NULL; // character
FLOAT32 sec_rat = 0.0f; // rating
FLOAT32 sec_cert = 0.0f; // certainty
BLOB_CHOICE_IT c_it = ratings; // iterator
index = ratings->length();
if (index > 0) {
first_char = current_unicharset.id_to_unichar(c_it.data()->unichar_id());
first_rat = c_it.data()->rating();
first_cert = -c_it.data()->certainty();
if (index > 1) {
sec_char = current_unicharset.id_to_unichar(
c_it.data_relative(1)->unichar_id());
sec_rat = c_it.data_relative(1)->rating();
sec_cert = -c_it.data_relative(1)->certainty();
} else {
sec_char = NULL;
sec_rat = -1;
sec_cert = -1;
}
} else {
first_char = NULL;
first_rat = -1;
first_cert = -1;
}
if (first_char != NULL && (*first_char == '\0' || *first_char == ' '))
first_char = NULL;
if (sec_char != NULL && (*sec_char == '\0' || *sec_char == ' '))
sec_char = NULL;
tprintf(" " INT32FORMAT " %s %g %g %s %g %g\n",
ratings->length(),
first_char != NULL ? first_char : "~",
first_rat, first_cert, sec_char != NULL ? sec_char : "~",
sec_rat, sec_cert);
}
// Makes this a copy of src. Clears this completely first, so the automatic
// ids will not be present in this if not in src. Does NOT reorder the set!
void UNICHARSET::CopyFrom(const UNICHARSET& src) {
clear();
for (int ch = 0; ch < src.size_used; ++ch) {
const UNICHAR_PROPERTIES& src_props = src.unichars[ch].properties;
const char* utf8 = src.id_to_unichar(ch);
unichar_insert(utf8);
unichars[ch].properties.ExpandRangesFrom(src_props);
}
// Set properties, including mirror and other_case, WITHOUT reordering
// the unicharset.
PartialSetPropertiesFromOther(0, src);
}
/**
* string_and_lengths
*
* Populates the given word_str with unichars from unichar_ids and
* and word_lengths_str with the corresponding unichar lengths.
* Uses current_unicharset to make unichar id -> unichar conversions.
*/
void WERD_CHOICE::string_and_lengths(const UNICHARSET ¤t_unicharset,
STRING *word_str,
STRING *word_lengths_str) const {
*word_str = "";
if (word_lengths_str != NULL) *word_lengths_str = "";
for (int i = 0; i < length_; ++i) {
const char *ch = current_unicharset.id_to_unichar(unichar_ids_[i]);
*word_str += ch;
if (word_lengths_str != NULL) {
*word_lengths_str += strlen(ch);
}
}
}
// Helper prints the given set of blob choices.
static void PrintPath(int length, const BLOB_CHOICE** blob_choices,
const UNICHARSET& unicharset,
const char *label, FILE *output_file) {
float rating = 0.0f;
float certainty = 0.0f;
for (int i = 0; i < length; ++i) {
const BLOB_CHOICE* blob_choice = blob_choices[i];
fprintf(output_file, "%s",
unicharset.id_to_unichar(blob_choice->unichar_id()));
rating += blob_choice->rating();
if (certainty > blob_choice->certainty())
certainty = blob_choice->certainty();
}
fprintf(output_file, "\t%s\t%.4f\t%.4f\n",
label, rating, certainty);
}
// For each id in src, if it does not occur in this, add it, as in
// SetPropertiesFromOther, otherwise expand the ranges, as in
// ExpandRangesFromOther.
void UNICHARSET::AppendOtherUnicharset(const UNICHARSET& src) {
int initial_used = size_used;
for (int ch = 0; ch < src.size_used; ++ch) {
const UNICHAR_PROPERTIES& src_props = src.unichars[ch].properties;
const char* utf8 = src.id_to_unichar(ch);
int id = size_used;
if (contains_unichar(utf8)) {
id = unichar_to_id(utf8);
// Just expand current ranges.
unichars[id].properties.ExpandRangesFrom(src_props);
} else {
unichar_insert_backwards_compatible(utf8);
unichars[id].properties.SetRangesEmpty();
}
}
// Set properties, including mirror and other_case, WITHOUT reordering
// the unicharset.
PartialSetPropertiesFromOther(initial_used, src);
}
// Helper to set the properties for an input unicharset file, writes to the
// output file. If an appropriate script unicharset can be found in the
// script_dir directory, then the tops and bottoms are expanded using the
// script unicharset.
// If non-empty, xheight data for the fonts are written to the xheights_file.
void SetPropertiesForInputFile(const string& script_dir,
const string& input_unicharset_file,
const string& output_unicharset_file,
const string& output_xheights_file) {
UNICHARSET unicharset;
// Load the input unicharset
unicharset.load_from_file(input_unicharset_file.c_str());
tprintf("Loaded unicharset of size %d from file %s\n", unicharset.size(),
input_unicharset_file.c_str());
// Set unichar properties
tprintf("Setting unichar properties\n");
SetupBasicProperties(true, false, &unicharset);
string xheights_str;
for (int s = 0; s < unicharset.get_script_table_size(); ++s) {
// Load the unicharset for the script if available.
string filename = script_dir + "/" +
unicharset.get_script_from_script_id(s) + ".unicharset";
UNICHARSET script_set;
if (script_set.load_from_file(filename.c_str())) {
unicharset.SetPropertiesFromOther(script_set);
}
// Load the xheights for the script if available.
filename = script_dir + "/" + unicharset.get_script_from_script_id(s) +
".xheights";
string script_heights;
if (File::ReadFileToString(filename, &script_heights))
xheights_str += script_heights;
}
if (!output_xheights_file.empty())
File::WriteStringToFileOrDie(xheights_str, output_xheights_file);
for (int c = SPECIAL_UNICHAR_CODES_COUNT; c < unicharset.size(); ++c) {
if (unicharset.PropertiesIncomplete(c)) {
tprintf("Warning: properties incomplete for index %d = %s\n",
c, unicharset.id_to_unichar(c));
}
}
// Write the output unicharset
tprintf("Writing unicharset to file %s\n", output_unicharset_file.c_str());
unicharset.save_to_file(output_unicharset_file.c_str());
}
/* static */
ScriptPos WERD_CHOICE::ScriptPositionOf(bool print_debug,
const UNICHARSET& unicharset,
const TBOX& blob_box,
UNICHAR_ID unichar_id) {
ScriptPos retval = tesseract::SP_NORMAL;
int top = blob_box.top();
int bottom = blob_box.bottom();
int min_bottom, max_bottom, min_top, max_top;
unicharset.get_top_bottom(unichar_id,
&min_bottom, &max_bottom,
&min_top, &max_top);
int sub_thresh_top = min_top - kMinSubscriptOffset;
int sub_thresh_bot = kBlnBaselineOffset - kMinSubscriptOffset;
int sup_thresh_bot = max_bottom + kMinSuperscriptOffset;
if (bottom <= kMaxDropCapBottom) {
retval = tesseract::SP_DROPCAP;
} else if (top < sub_thresh_top && bottom < sub_thresh_bot) {
retval = tesseract::SP_SUBSCRIPT;
} else if (bottom > sup_thresh_bot) {
retval = tesseract::SP_SUPERSCRIPT;
}
if (print_debug) {
const char *pos = ScriptPosToString(retval);
tprintf("%s Character %s[bot:%d top: %d] "
"bot_range[%d,%d] top_range[%d, %d] "
"sub_thresh[bot:%d top:%d] sup_thresh_bot %d\n",
pos, unicharset.id_to_unichar(unichar_id),
bottom, top,
min_bottom, max_bottom, min_top, max_top,
sub_thresh_bot, sub_thresh_top,
sup_thresh_bot);
}
return retval;
}