void Wordrec::ProcessSegSearchPainPoint(
float pain_point_priority,
const MATRIX_COORD &pain_point, const char* pain_point_type,
GenericVector<SegSearchPending>* pending, WERD_RES *word_res,
LMPainPoints *pain_points, BlamerBundle *blamer_bundle) {
if (segsearch_debug_level > 0) {
tprintf("Classifying pain point %s priority=%.4f, col=%d, row=%d\n",
pain_point_type, pain_point_priority,
pain_point.col, pain_point.row);
}
ASSERT_HOST(pain_points != NULL);
MATRIX *ratings = word_res->ratings;
// Classify blob [pain_point.col pain_point.row]
if (!pain_point.Valid(*ratings)) {
ratings->IncreaseBandSize(pain_point.row + 1 - pain_point.col);
}
ASSERT_HOST(pain_point.Valid(*ratings));
BLOB_CHOICE_LIST *classified = classify_piece(word_res->seam_array,
pain_point.col, pain_point.row,
pain_point_type,
word_res->chopped_word,
blamer_bundle);
BLOB_CHOICE_LIST *lst = ratings->get(pain_point.col, pain_point.row);
if (lst == NULL) {
ratings->put(pain_point.col, pain_point.row, classified);
} else {
// We can not delete old BLOB_CHOICEs, since they might contain
// ViterbiStateEntries that are parents of other "active" entries.
// Thus if the matrix cell already contains classifications we add
// the new ones to the beginning of the list.
BLOB_CHOICE_IT it(lst);
it.add_list_before(classified);
delete classified; // safe to delete, since empty after add_list_before()
classified = NULL;
}
if (segsearch_debug_level > 0) {
print_ratings_list("Updated ratings matrix with a new entry:",
ratings->get(pain_point.col, pain_point.row),
getDict().getUnicharset());
ratings->print(getDict().getUnicharset());
}
// Insert initial "pain points" to join the newly classified blob
// with its left and right neighbors.
if (classified != NULL && !classified->empty()) {
if (pain_point.col > 0) {
pain_points->GeneratePainPoint(
pain_point.col - 1, pain_point.row, LM_PPTYPE_SHAPE, 0.0,
true, segsearch_max_char_wh_ratio, word_res);
}
if (pain_point.row + 1 < ratings->dimension()) {
pain_points->GeneratePainPoint(
pain_point.col, pain_point.row + 1, LM_PPTYPE_SHAPE, 0.0,
true, segsearch_max_char_wh_ratio, word_res);
}
}
(*pending)[pain_point.col].SetBlobClassified(pain_point.row);
}
// Helper recursively prints all paths through the ratings matrix, starting
// at column col.
static void PrintMatrixPaths(int col, int dim,
const MATRIX& ratings,
int length, const BLOB_CHOICE** blob_choices,
const UNICHARSET& unicharset,
const char *label, FILE *output_file) {
for (int row = col; row < dim && row - col < ratings.bandwidth(); ++row) {
if (ratings.get(col, row) != NOT_CLASSIFIED) {
BLOB_CHOICE_IT bc_it(ratings.get(col, row));
for (bc_it.mark_cycle_pt(); !bc_it.cycled_list(); bc_it.forward()) {
blob_choices[length] = bc_it.data();
if (row + 1 < dim) {
PrintMatrixPaths(row + 1, dim, ratings, length + 1, blob_choices,
unicharset, label, output_file);
} else {
PrintPath(length + 1, blob_choices, unicharset, label, output_file);
}
}
}
}
}
void LMPainPoints::GenerateInitial(WERD_RES *word_res) {
MATRIX *ratings = word_res->ratings;
AssociateStats associate_stats;
for (int col = 0; col < ratings->dimension(); ++col) {
int row_end = MIN(ratings->dimension(), col + ratings->bandwidth() + 1);
for (int row = col + 1; row < row_end; ++row) {
MATRIX_COORD coord(col, row);
if (coord.Valid(*ratings) &&
ratings->get(col, row) != NOT_CLASSIFIED) continue;
// Add an initial pain point if needed.
if (ratings->Classified(col, row - 1, dict_->WildcardID()) ||
(col + 1 < ratings->dimension() &&
ratings->Classified(col + 1, row, dict_->WildcardID()))) {
GeneratePainPoint(col, row, LM_PPTYPE_SHAPE, 0.0,
true, max_char_wh_ratio_, word_res);
}
}
}
}
void Wordrec::SegSearch(CHUNKS_RECORD *chunks_record,
WERD_CHOICE *best_choice,
BLOB_CHOICE_LIST_VECTOR *best_char_choices,
WERD_CHOICE *raw_choice,
STATE *output_best_state) {
int row, col = 0;
if (segsearch_debug_level > 0) {
tprintf("Starting SegSearch on ratings matrix:\n");
chunks_record->ratings->print(getDict().getUnicharset());
}
// Start with a fresh best_choice since rating adjustments
// used by the chopper and the new segmentation search are not compatible.
best_choice->set_rating(WERD_CHOICE::kBadRating);
// Clear best choice accumulator (that is used for adaption), so that
// choices adjusted by chopper do not interfere with the results from the
// segmentation search.
getDict().ClearBestChoiceAccum();
MATRIX *ratings = chunks_record->ratings;
// Priority queue containing pain points generated by the language model
// The priority is set by the language model components, adjustments like
// seam cost and width priority are factored into the priority.
HEAP *pain_points = MakeHeap(segsearch_max_pain_points);
// best_path_by_column records the lowest cost path found so far for each
// column of the chunks_record->ratings matrix over all the rows.
BestPathByColumn *best_path_by_column =
new BestPathByColumn[ratings->dimension()];
for (col = 0; col < ratings->dimension(); ++col) {
best_path_by_column[col].avg_cost = WERD_CHOICE::kBadRating;
best_path_by_column[col].best_vse = NULL;
}
language_model_->InitForWord(prev_word_best_choice_, &denorm_,
assume_fixed_pitch_char_segment,
best_choice->certainty(),
segsearch_max_char_wh_ratio,
pain_points, chunks_record);
MATRIX_COORD *pain_point;
float pain_point_priority;
BestChoiceBundle best_choice_bundle(
output_best_state, best_choice, raw_choice, best_char_choices);
// pending[i] stores a list of the parent/child pair of BLOB_CHOICE_LISTs,
// where i is the column of the child. Initially all the classified entries
// in the ratings matrix from column 0 (with parent NULL) are inserted into
// pending[0]. As the language model state is updated, new child/parent
// pairs are inserted into the lists. Next, the entries in pending[1] are
// considered, and so on. It is important that during the update the
// children are considered in the non-decreasing order of their column, since
// this guarantess that all the parents would be up to date before an update
// of a child is done.
SEG_SEARCH_PENDING_LIST *pending =
new SEG_SEARCH_PENDING_LIST[ratings->dimension()];
// Search for the ratings matrix for the initial best path.
for (row = 0; row < ratings->dimension(); ++row) {
if (ratings->get(0, row) != NOT_CLASSIFIED) {
pending[0].add_sorted(
SEG_SEARCH_PENDING::compare, true,
new SEG_SEARCH_PENDING(row, NULL, LanguageModel::kAllChangedFlag));
}
}
UpdateSegSearchNodes(0, &pending, &best_path_by_column, chunks_record,
pain_points, &best_choice_bundle);
// Keep trying to find a better path by fixing the "pain points".
int num_futile_classifications = 0;
while (!(language_model_->AcceptableChoiceFound() ||
num_futile_classifications >=
segsearch_max_futile_classifications)) {
// Get the next valid "pain point".
int pop;
while (true) {
pop = HeapPop(pain_points, &pain_point_priority, &pain_point);
if (pop == EMPTY) break;
if (pain_point->Valid(*ratings) &&
ratings->get(pain_point->col, pain_point->row) == NOT_CLASSIFIED) {
break;
} else {
delete pain_point;
}
}
if (pop == EMPTY) {
if (segsearch_debug_level > 0) tprintf("Pain points queue is empty\n");
break;
}
if (segsearch_debug_level > 0) {
tprintf("Classifying pain point priority=%.4f, col=%d, row=%d\n",
pain_point_priority, pain_point->col, pain_point->row);
}
BLOB_CHOICE_LIST *classified = classify_piece(
chunks_record->chunks, chunks_record->splits,
pain_point->col, pain_point->row);
ratings->put(pain_point->col, pain_point->row, classified);
if (segsearch_debug_level > 0) {
print_ratings_list("Updated ratings matrix with a new entry:",
ratings->get(pain_point->col, pain_point->row),
getDict().getUnicharset());
chunks_record->ratings->print(getDict().getUnicharset());
}
// Insert initial "pain points" to join the newly classified blob
// with its left and right neighbors.
if (!classified->empty()) {
float worst_piece_cert;
bool fragmented;
if (pain_point->col > 0) {
language_model_->GetWorstPieceCertainty(
pain_point->col-1, pain_point->row, chunks_record->ratings,
&worst_piece_cert, &fragmented);
language_model_->GeneratePainPoint(
pain_point->col-1, pain_point->row, false,
LanguageModel::kInitialPainPointPriorityAdjustment,
worst_piece_cert, fragmented, best_choice->certainty(),
segsearch_max_char_wh_ratio, NULL, NULL,
chunks_record, pain_points);
}
if (pain_point->row+1 < ratings->dimension()) {
language_model_->GetWorstPieceCertainty(
pain_point->col, pain_point->row+1, chunks_record->ratings,
&worst_piece_cert, &fragmented);
language_model_->GeneratePainPoint(
pain_point->col, pain_point->row+1, true,
LanguageModel::kInitialPainPointPriorityAdjustment,
worst_piece_cert, fragmented, best_choice->certainty(),
segsearch_max_char_wh_ratio, NULL, NULL,
chunks_record, pain_points);
}
}
// Record a pending entry with the pain_point and each of its parents.
int parent_row = pain_point->col - 1;
if (parent_row < 0) { // this node has no parents
pending[pain_point->col].add_sorted(
SEG_SEARCH_PENDING::compare, true,
new SEG_SEARCH_PENDING(pain_point->row, NULL,
LanguageModel::kAllChangedFlag));
} else {
for (int parent_col = 0; parent_col < pain_point->col; ++parent_col) {
if (ratings->get(parent_col, parent_row) != NOT_CLASSIFIED) {
pending[pain_point->col].add_sorted(
SEG_SEARCH_PENDING::compare, true,
new SEG_SEARCH_PENDING(pain_point->row,
ratings->get(parent_col, parent_row),
LanguageModel::kAllChangedFlag));
}
}
}
UpdateSegSearchNodes(pain_point->col, &pending, &best_path_by_column,
chunks_record, pain_points, &best_choice_bundle);
if (!best_choice_bundle.updated) ++num_futile_classifications;
if (segsearch_debug_level > 0) {
tprintf("num_futile_classifications %d\n", num_futile_classifications);
}
// Clean up
best_choice_bundle.updated = false;
delete pain_point; // done using this pain point
}
if (segsearch_debug_level > 0) {
tprintf("Done with SegSearch (AcceptableChoiceFound: %d\n",
language_model_->AcceptableChoiceFound());
}
// Clean up.
FreeHeapData(pain_points, MATRIX_COORD::Delete);
delete[] best_path_by_column;
delete[] pending;
for (row = 0; row < ratings->dimension(); ++row) {
for (col = 0; col <= row; ++col) {
BLOB_CHOICE_LIST *rating = ratings->get(col, row);
if (rating != NOT_CLASSIFIED) language_model_->DeleteState(rating);
}
}
}
void Wordrec::UpdateSegSearchNodes(
int starting_col,
SEG_SEARCH_PENDING_LIST *pending[],
BestPathByColumn *best_path_by_column[],
CHUNKS_RECORD *chunks_record,
HEAP *pain_points,
BestChoiceBundle *best_choice_bundle) {
MATRIX *ratings = chunks_record->ratings;
for (int col = starting_col; col < ratings->dimension(); ++col) {
if (segsearch_debug_level > 0) {
tprintf("\n\nUpdateSegSearchNodes: evaluate children in col=%d\n", col);
}
// Iterate over the pending list for this column.
SEG_SEARCH_PENDING_LIST *pending_list = &((*pending)[col]);
SEG_SEARCH_PENDING_IT pending_it(pending_list);
GenericVector<int> non_empty_rows;
while (!pending_it.empty()) {
// Update language model state of this child+parent pair.
SEG_SEARCH_PENDING *p = pending_it.extract();
if (non_empty_rows.length() == 0 ||
non_empty_rows[non_empty_rows.length()-1] != p->child_row) {
non_empty_rows.push_back(p->child_row);
}
BLOB_CHOICE_LIST *current_node = ratings->get(col, p->child_row);
LanguageModelFlagsType new_changed =
language_model_->UpdateState(p->changed, col, p->child_row,
current_node, p->parent, pain_points,
best_path_by_column,
chunks_record, best_choice_bundle);
if (new_changed) {
// Since the language model state of this entry changed, add all the
// pairs with it as a parent and each of its children to pending, so
// that the children are updated as well.
int child_col = p->child_row + 1;
for (int child_row = child_col;
child_row < ratings->dimension(); ++child_row) {
if (ratings->get(child_col, child_row) != NOT_CLASSIFIED) {
SEG_SEARCH_PENDING *new_pending =
new SEG_SEARCH_PENDING(child_row, current_node, 0);
SEG_SEARCH_PENDING *actual_new_pending =
reinterpret_cast<SEG_SEARCH_PENDING *>(
(*pending)[child_col].add_sorted_and_find(
SEG_SEARCH_PENDING::compare, true, new_pending));
if (new_pending != actual_new_pending) delete new_pending;
actual_new_pending->changed |= new_changed;
if (segsearch_debug_level > 0) {
tprintf("Added child(col=%d row=%d) parent(col=%d row=%d)"
" changed=0x%x to pending\n", child_col,
actual_new_pending->child_row,
col, p->child_row, actual_new_pending->changed);
}
}
}
} // end if new_changed
delete p; // clean up
pending_it.forward();
} // end while !pending_it.empty()
language_model_->GeneratePainPointsFromColumn(
col, non_empty_rows, best_choice_bundle->best_choice->certainty(),
pain_points, best_path_by_column, chunks_record);
} // end for col
if (best_choice_bundle->updated) {
language_model_->GeneratePainPointsFromBestChoice(
pain_points, chunks_record, best_choice_bundle);
}
language_model_->CleanUp();
}
void Wordrec::UpdateSegSearchNodes(
float rating_cert_scale,
int starting_col,
GenericVector<SegSearchPending>* pending,
WERD_RES *word_res,
LMPainPoints *pain_points,
BestChoiceBundle *best_choice_bundle,
BlamerBundle *blamer_bundle) {
MATRIX *ratings = word_res->ratings;
ASSERT_HOST(ratings->dimension() == pending->size());
ASSERT_HOST(ratings->dimension() == best_choice_bundle->beam.size());
for (int col = starting_col; col < ratings->dimension(); ++col) {
if (!(*pending)[col].WorkToDo()) continue;
int first_row = col;
int last_row = MIN(ratings->dimension() - 1,
col + ratings->bandwidth() - 1);
if ((*pending)[col].SingleRow() >= 0) {
first_row = last_row = (*pending)[col].SingleRow();
}
if (segsearch_debug_level > 0) {
tprintf("\n\nUpdateSegSearchNodes: col=%d, rows=[%d,%d], alljust=%d\n",
col, first_row, last_row,
(*pending)[col].IsRowJustClassified(INT32_MAX));
}
// Iterate over the pending list for this column.
for (int row = first_row; row <= last_row; ++row) {
// Update language model state of this child+parent pair.
BLOB_CHOICE_LIST *current_node = ratings->get(col, row);
LanguageModelState *parent_node =
col == 0 ? NULL : best_choice_bundle->beam[col - 1];
if (current_node != NULL &&
language_model_->UpdateState((*pending)[col].IsRowJustClassified(row),
col, row, current_node, parent_node,
pain_points, word_res,
best_choice_bundle, blamer_bundle) &&
row + 1 < ratings->dimension()) {
// Since the language model state of this entry changed, process all
// the child column.
(*pending)[row + 1].RevisitWholeColumn();
if (segsearch_debug_level > 0) {
tprintf("Added child col=%d to pending\n", row + 1);
}
} // end if UpdateState.
} // end for row.
} // end for col.
if (best_choice_bundle->best_vse != NULL) {
ASSERT_HOST(word_res->StatesAllValid());
if (best_choice_bundle->best_vse->updated) {
pain_points->GenerateFromPath(rating_cert_scale,
best_choice_bundle->best_vse, word_res);
if (!best_choice_bundle->fixpt.empty()) {
pain_points->GenerateFromAmbigs(best_choice_bundle->fixpt,
best_choice_bundle->best_vse, word_res);
}
}
}
// The segsearch is completed. Reset all updated flags on all VSEs and reset
// all pendings.
for (int col = 0; col < pending->size(); ++col) {
(*pending)[col].Clear();
ViterbiStateEntry_IT
vse_it(&best_choice_bundle->beam[col]->viterbi_state_entries);
for (vse_it.mark_cycle_pt(); !vse_it.cycled_list(); vse_it.forward()) {
vse_it.data()->updated = false;
}
}
}