bool Dawg::match_words(WERD_CHOICE *word, inT32 index,
NODE_REF node, UNICHAR_ID wildcard) const {
EDGE_REF edge;
inT32 word_end;
if (wildcard != INVALID_UNICHAR_ID && word->unichar_id(index) == wildcard) {
bool any_matched = false;
NodeChildVector vec;
this->unichar_ids_of(node, &vec);
for (int i = 0; i < vec.size(); ++i) {
word->set_unichar_id(vec[i].unichar_id, index);
if (match_words(word, index, node, wildcard))
any_matched = true;
}
word->set_unichar_id(wildcard, index);
return any_matched;
} else {
word_end = index == word->length() - 1;
edge = edge_char_of(node, word->unichar_id(index), word_end);
if (edge != NO_EDGE) { // normal edge in DAWG
node = next_node(edge);
if (word_end) {
if (debug_level_ > 1) word->print("match_words() found: ");
return true;
} else if (node != 0) {
return match_words(word, index+1, node, wildcard);
}
}
}
return false;
}
bool Dawg::prefix_in_dawg(const WERD_CHOICE &word,
bool requires_complete) const {
if (word.length() == 0) return !requires_complete;
NODE_REF node = 0;
int end_index = word.length() - 1;
for (int i = 0; i < end_index; i++) {
EDGE_REF edge = edge_char_of(node, word.unichar_id(i), false);
if (edge == NO_EDGE) {
return false;
}
if ((node = next_node(edge)) == 0) {
// This only happens if all words following this edge terminate --
// there are no larger words. See Trie::add_word_to_dawg()
return false;
}
}
// Now check the last character.
return edge_char_of(node, word.unichar_id(end_index), requires_complete) !=
NO_EDGE;
}