tommy/libpostal
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[fix] tokenized trie search was skipping tokens in some cases

Browse Source
This commit is contained in:
Al
2015-08-02 14:36:21 -06:00
parent 3178eda501
commit b27030e39f
1 changed files with 6 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
src/trie_search.c
View File

@@ -319,8 +319,8 @@ phrase_array *trie_search_tokens_from_index(trie_t *self, char *str, token_array
i = last_match_index; i = last_match_index;
last_match_index = -1; last_match_index = -1;
phrase_start = 0; phrase_start = 0;
last_node_id = start_node_id; node_id = last_node_id = start_node_id;
last_node = trie_get_node(self, last_node_id); node = last_node = trie_get_node(self, start_node_id);
continue; continue;
} else if (last_state == SEARCH_STATE_PARTIAL_MATCH) { } else if (last_state == SEARCH_STATE_PARTIAL_MATCH) {
log_debug("last_state == SEARCH_STATE_PARTIAL_MATCH\n"); log_debug("last_state == SEARCH_STATE_PARTIAL_MATCH\n");
@@ -331,8 +331,8 @@ phrase_array *trie_search_tokens_from_index(trie_t *self, char *str, token_array
// this token was not a phrase // this token was not a phrase
log_debug("Plain token=%.*s\n", (int)token.len, str + token.offset); log_debug("Plain token=%.*s\n", (int)token.len, str + token.offset);
} }
last_node_id = start_node_id; node_id = last_node_id = start_node_id;
last_node = trie_get_node(self, last_node_id); node = last_node = trie_get_node(self, start_node_id);
} else { } else {
state = SEARCH_STATE_PARTIAL_MATCH; state = SEARCH_STATE_PARTIAL_MATCH;
@@ -378,8 +378,8 @@ phrase_array *trie_search_tokens_from_index(trie_t *self, char *str, token_array
log_debug("node->match no continuation\n"); log_debug("node->match no continuation\n");
phrase_array_push(phrases, (phrase_t){phrase_start, last_match_index - phrase_start + 1, data}); phrase_array_push(phrases, (phrase_t){phrase_start, last_match_index - phrase_start + 1, data});
last_match_index = -1; last_match_index = -1;
last_node_id = start_node_id; node_id = last_node_id = start_node_id;
last_node = trie_get_node(self, last_node_id); node = last_node = trie_get_node(self, start_node_id);
state = SEARCH_STATE_BEGIN; state = SEARCH_STATE_BEGIN;
} else { } else {
log_debug("Has continuation, node_id=%d\n", continuation_id); log_debug("Has continuation, node_id=%d\n", continuation_id);