/**
* BLOB_CHOICE::BLOB_CHOICE
*
* Constructor to build a BLOB_CHOICE from another BLOB_CHOICE.
*/
BLOB_CHOICE::BLOB_CHOICE(const BLOB_CHOICE &other) {
unichar_id_ = other.unichar_id();
rating_ = other.rating();
certainty_ = other.certainty();
config_ = other.config();
script_id_ = other.script_id();
}
/**
* BLOB_CHOICE::BLOB_CHOICE
*
* Constructor to build a BLOB_CHOICE from another BLOB_CHOICE.
*/
BLOB_CHOICE::BLOB_CHOICE(const BLOB_CHOICE &other) {
unichar_id_ = other.unichar_id();
rating_ = other.rating();
certainty_ = other.certainty();
config_ = other.config();
config2_ = other.config2();
script_id_ = other.script_id();
language_model_state_ = NULL;
}
/**
* BLOB_CHOICE::BLOB_CHOICE
*
* Constructor to build a BLOB_CHOICE from another BLOB_CHOICE.
*/
BLOB_CHOICE::BLOB_CHOICE(const BLOB_CHOICE &other) {
unichar_id_ = other.unichar_id();
rating_ = other.rating();
certainty_ = other.certainty();
fontinfo_id_ = other.fontinfo_id();
fontinfo_id2_ = other.fontinfo_id2();
script_id_ = other.script_id();
language_model_state_ = NULL;
min_xheight_ = other.min_xheight_;
max_xheight_ = other.max_xheight_;
adapted_ = other.adapted_;
}
/**
* BLOB_CHOICE::BLOB_CHOICE
*
* Constructor to build a BLOB_CHOICE from another BLOB_CHOICE.
*/
BLOB_CHOICE::BLOB_CHOICE(const BLOB_CHOICE &other) {
unichar_id_ = other.unichar_id();
rating_ = other.rating();
certainty_ = other.certainty();
fontinfo_id_ = other.fontinfo_id();
fontinfo_id2_ = other.fontinfo_id2();
script_id_ = other.script_id();
matrix_cell_ = other.matrix_cell_;
min_xheight_ = other.min_xheight_;
max_xheight_ = other.max_xheight_;
yshift_ = other.yshift();
classifier_ = other.classifier_;
}
/**
* append_choices
*
* Checks to see whether or not the next choice is worth appending to
* the word being generated. If so then keeps going deeper into the word.
*
* This function assumes that Dict::go_deeper_fxn_ is set.
*/
void Dict::append_choices(
const char *debug,
const BLOB_CHOICE_LIST_VECTOR &char_choices,
const BLOB_CHOICE &blob_choice,
int char_choice_index,
const CHAR_FRAGMENT_INFO *prev_char_frag_info,
WERD_CHOICE *word,
float certainties[],
float *limit,
WERD_CHOICE *best_choice,
int *attempts_left,
void *more_args) {
int word_ending =
(char_choice_index == char_choices.length() - 1) ? true : false;
// Deal with fragments.
CHAR_FRAGMENT_INFO char_frag_info;
if (!fragment_state_okay(blob_choice.unichar_id(), blob_choice.rating(),
blob_choice.certainty(), prev_char_frag_info, debug,
word_ending, &char_frag_info)) {
return; // blob_choice must be an invalid fragment
}
// Search the next letter if this character is a fragment.
if (char_frag_info.unichar_id == INVALID_UNICHAR_ID) {
permute_choices(debug, char_choices, char_choice_index + 1,
&char_frag_info, word, certainties, limit,
best_choice, attempts_left, more_args);
return;
}
// Add the next unichar.
float old_rating = word->rating();
float old_certainty = word->certainty();
uint8_t old_permuter = word->permuter();
certainties[word->length()] = char_frag_info.certainty;
word->append_unichar_id_space_allocated(
char_frag_info.unichar_id, char_frag_info.num_fragments,
char_frag_info.rating, char_frag_info.certainty);
// Explore the next unichar.
(this->*go_deeper_fxn_)(debug, char_choices, char_choice_index,
&char_frag_info, word_ending, word, certainties,
limit, best_choice, attempts_left, more_args);
// Remove the unichar we added to explore other choices in it's place.
word->remove_last_unichar_id();
word->set_rating(old_rating);
word->set_certainty(old_certainty);
word->set_permuter(old_permuter);
}
// Creates a fake blob choice from the combination of the given fragments.
// unichar is the class to be made from the combination,
// expanded_fragment_lengths[choice_index] is the number of fragments to use.
// old_choices[choice_index] has the classifier output for each fragment.
// choice index initially indexes the last fragment and should be decremented
// expanded_fragment_lengths[choice_index] times to get the earlier fragments.
// Guarantees to return something non-null, or abort!
BLOB_CHOICE* Wordrec::rebuild_fragments(
const char* unichar,
const char* expanded_fragment_lengths,
int choice_index,
BLOB_CHOICE_LIST_VECTOR *old_choices) {
float rating = 0.0f;
float certainty = 0.0f;
inT16 min_xheight = -MAX_INT16;
inT16 max_xheight = MAX_INT16;
for (int fragment_pieces = expanded_fragment_lengths[choice_index] - 1;
fragment_pieces >= 0; --fragment_pieces, --choice_index) {
// Get a pointer to the classifier results from the old_choices.
BLOB_CHOICE_LIST *current_choices = old_choices->get(choice_index);
// Populate fragment with updated values and look for the
// fragment with the same values in current_choices.
// Update rating and certainty of the character being composed.
CHAR_FRAGMENT fragment;
fragment.set_all(unichar, fragment_pieces,
expanded_fragment_lengths[choice_index], false);
BLOB_CHOICE_IT choice_it(current_choices);
for (choice_it.mark_cycle_pt(); !choice_it.cycled_list();
choice_it.forward()) {
BLOB_CHOICE* choice = choice_it.data();
const CHAR_FRAGMENT *current_fragment =
getDict().getUnicharset().get_fragment(choice->unichar_id());
if (current_fragment && fragment.equals(current_fragment)) {
rating += choice->rating();
if (choice->certainty() < certainty) {
certainty = choice->certainty();
}
IntersectRange(choice->min_xheight(), choice->max_xheight(),
&min_xheight, &max_xheight);
break;
}
}
if (choice_it.cycled_list()) {
print_ratings_list("Failure", current_choices, unicharset);
tprintf("Failed to find fragment %s at index=%d\n",
fragment.to_string().string(), choice_index);
}
ASSERT_HOST(!choice_it.cycled_list()); // Be sure we found the fragment.
}
return new BLOB_CHOICE(getDict().getUnicharset().unichar_to_id(unichar),
rating, certainty, -1, -1, 0,
min_xheight, max_xheight, false);
}
LIST call_matcher( //call a matcher
TBLOB *ptblob, //previous
TBLOB *tessblob, //blob to match
TBLOB *ntblob, //next
void *, //unused parameter
TEXTROW * //always null anyway
) {
PBLOB *pblob; //converted blob
PBLOB *blob; //converted blob
PBLOB *nblob; //converted blob
LIST result; //tess output
BLOB_CHOICE *choice; //current choice
BLOB_CHOICE_LIST ratings; //matcher result
BLOB_CHOICE_IT it; //iterator
char choice_lengths[2] = {0, 0};
blob = make_ed_blob (tessblob);//convert blob
if (blob == NULL) {
// Since it is actually possible to get a NULL blob here, due to invalid
// segmentations, fake a really bad classification.
choice_lengths[0] = strlen(unicharset.id_to_unichar(1));
return append_choice(NULL, unicharset.id_to_unichar(1), choice_lengths,
static_cast<float>(MAX_NUM_INT_FEATURES),
static_cast<float>(kReallyBadCertainty), 0);
}
pblob = ptblob != NULL ? make_ed_blob (ptblob) : NULL;
nblob = ntblob != NULL ? make_ed_blob (ntblob) : NULL;
(*tess_matcher) (pblob, blob, nblob, tess_word, tess_denorm, ratings);
//match it
delete blob; //don't need that now
if (pblob != NULL)
delete pblob;
if (nblob != NULL)
delete nblob;
it.set_to_list (&ratings); //get list
result = NULL;
for (it.mark_cycle_pt (); !it.cycled_list (); it.forward ()) {
choice = it.data ();
choice_lengths[0] = strlen(choice->unichar ());
result = append_choice (result, choice->unichar (),
choice_lengths, choice->rating (),
choice->certainty (), choice->config ());
}
return result; //converted list
}
void Dict::ReplaceAmbig(int wrong_ngram_begin_index, int wrong_ngram_size,
UNICHAR_ID correct_ngram_id, WERD_CHOICE *werd_choice,
MATRIX *ratings) {
int num_blobs_to_replace = 0;
int begin_blob_index = 0;
int i;
// Rating and certainty for the new BLOB_CHOICE are derived from the
// replaced choices.
float new_rating = 0.0f;
float new_certainty = 0.0f;
BLOB_CHOICE* old_choice = nullptr;
for (i = 0; i < wrong_ngram_begin_index + wrong_ngram_size; ++i) {
if (i >= wrong_ngram_begin_index) {
int num_blobs = werd_choice->state(i);
int col = begin_blob_index + num_blobs_to_replace;
int row = col + num_blobs - 1;
BLOB_CHOICE_LIST* choices = ratings->get(col, row);
ASSERT_HOST(choices != nullptr);
old_choice = FindMatchingChoice(werd_choice->unichar_id(i), choices);
ASSERT_HOST(old_choice != nullptr);
new_rating += old_choice->rating();
new_certainty += old_choice->certainty();
num_blobs_to_replace += num_blobs;
} else {
begin_blob_index += werd_choice->state(i);
}
}
new_certainty /= wrong_ngram_size;
// If there is no entry in the ratings matrix, add it.
MATRIX_COORD coord(begin_blob_index,
begin_blob_index + num_blobs_to_replace - 1);
if (!coord.Valid(*ratings)) {
ratings->IncreaseBandSize(coord.row - coord.col + 1);
}
if (ratings->get(coord.col, coord.row) == nullptr)
ratings->put(coord.col, coord.row, new BLOB_CHOICE_LIST);
BLOB_CHOICE_LIST* new_choices = ratings->get(coord.col, coord.row);
BLOB_CHOICE* choice = FindMatchingChoice(correct_ngram_id, new_choices);
if (choice != nullptr) {
// Already there. Upgrade if new rating better.
if (new_rating < choice->rating())
choice->set_rating(new_rating);
if (new_certainty < choice->certainty())
choice->set_certainty(new_certainty);
// DO NOT SORT!! It will mess up the iterator in LanguageModel::UpdateState.
} else {
// Need a new choice with the correct_ngram_id.
choice = new BLOB_CHOICE(*old_choice);
choice->set_unichar_id(correct_ngram_id);
choice->set_rating(new_rating);
choice->set_certainty(new_certainty);
choice->set_classifier(BCC_AMBIG);
choice->set_matrix_cell(coord.col, coord.row);
BLOB_CHOICE_IT it (new_choices);
it.add_to_end(choice);
}
// Remove current unichar from werd_choice. On the last iteration
// set the correct replacement unichar instead of removing a unichar.
for (int replaced_count = 0; replaced_count < wrong_ngram_size;
++replaced_count) {
if (replaced_count + 1 == wrong_ngram_size) {
werd_choice->set_blob_choice(wrong_ngram_begin_index,
num_blobs_to_replace, choice);
} else {
werd_choice->remove_unichar_id(wrong_ngram_begin_index + 1);
}
}
if (stopper_debug_level >= 1) {
werd_choice->print("ReplaceAmbig() ");
tprintf("Modified blob_choices: ");
print_ratings_list("\n", new_choices, getUnicharset());
}
}
/// Recursive helper to find a match to the target_text (from text_index
/// position) in the choices (from choices_pos position).
/// @param choices is an array of GenericVectors, of length choices_length,
/// with each element representing a starting position in the word, and the
/// #GenericVector holding classification results for a sequence of consecutive
/// blobs, with index 0 being a single blob, index 1 being 2 blobs etc.
/// @param choices_pos
/// @param choices_length
/// @param target_text
/// @param text_index
/// @param rating
/// @param segmentation
/// @param best_rating
/// @param best_segmentation
void Tesseract::SearchForText(const GenericVector<BLOB_CHOICE_LIST*>* choices,
int choices_pos, int choices_length,
const GenericVector<UNICHAR_ID>& target_text,
int text_index,
float rating, GenericVector<int>* segmentation,
float* best_rating,
GenericVector<int>* best_segmentation) {
const UnicharAmbigsVector& table = getDict().getUnicharAmbigs().dang_ambigs();
for (int length = 1; length <= choices[choices_pos].size(); ++length) {
// Rating of matching choice or worst choice if no match.
float choice_rating = 0.0f;
// Find the corresponding best BLOB_CHOICE.
BLOB_CHOICE_IT choice_it(choices[choices_pos][length - 1]);
for (choice_it.mark_cycle_pt(); !choice_it.cycled_list();
choice_it.forward()) {
BLOB_CHOICE* choice = choice_it.data();
choice_rating = choice->rating();
UNICHAR_ID class_id = choice->unichar_id();
if (class_id == target_text[text_index]) {
break;
}
// Search ambigs table.
if (class_id < table.size() && table[class_id] != NULL) {
AmbigSpec_IT spec_it(table[class_id]);
for (spec_it.mark_cycle_pt(); !spec_it.cycled_list();
spec_it.forward()) {
const AmbigSpec *ambig_spec = spec_it.data();
// We'll only do 1-1.
if (ambig_spec->wrong_ngram[1] == INVALID_UNICHAR_ID &&
ambig_spec->correct_ngram_id == target_text[text_index])
break;
}
if (!spec_it.cycled_list())
break; // Found an ambig.
}
}
if (choice_it.cycled_list())
continue; // No match.
segmentation->push_back(length);
if (choices_pos + length == choices_length &&
text_index + 1 == target_text.size()) {
// This is a complete match. If the rating is good record a new best.
if (applybox_debug > 2) {
tprintf("Complete match, rating = %g, best=%g, seglength=%d, best=%d\n",
rating + choice_rating, *best_rating, segmentation->size(),
best_segmentation->size());
}
if (best_segmentation->empty() || rating + choice_rating < *best_rating) {
*best_segmentation = *segmentation;
*best_rating = rating + choice_rating;
}
} else if (choices_pos + length < choices_length &&
text_index + 1 < target_text.size()) {
if (applybox_debug > 3) {
tprintf("Match found for %d=%s:%s, at %d+%d, recursing...\n",
target_text[text_index],
unicharset.id_to_unichar(target_text[text_index]),
choice_it.data()->unichar_id() == target_text[text_index]
? "Match" : "Ambig",
choices_pos, length);
}
SearchForText(choices, choices_pos + length, choices_length, target_text,
text_index + 1, rating + choice_rating, segmentation,
best_rating, best_segmentation);
if (applybox_debug > 3) {
tprintf("End recursion for %d=%s\n", target_text[text_index],
unicharset.id_to_unichar(target_text[text_index]));
}
}
segmentation->truncate(segmentation->size() - 1);
}
}