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

[phrases] Prefix/suffix trie search using the new characters, fixing length of matched prefixes/suffixes and exiting early on falling off the the trie

Browse Source
This commit is contained in:
Al
2015-08-10 16:02:38 -04:00
parent 51572d6575
commit e511eede74
1 changed files with 11 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
src/trie_search.c
View File

@@ -449,12 +449,14 @@ phrase_t trie_search_suffixes_from_index(trie_t *self, char *word, size_t len, u
continue; continue;
} else if (in_tail && tail_remaining == 0 && i == char_len - 1) { } else if (in_tail && tail_remaining == 0 && i == char_len - 1) {
log_debug("tail match!\n"); log_debug("tail match!\n");
phrase_start = index; phrase_start = index + char_len;
phrase_len = len - index; phrase_len = len - index - char_len;
value = tail_value; value = tail_value;
index = 0;
break; break;
} else if (in_tail) { } else if (in_tail) {
log_debug("Done with tail\n"); log_debug("Done with tail\n");
index = 0;
break; break;
} }
@@ -463,6 +465,7 @@ phrase_t trie_search_suffixes_from_index(trie_t *self, char *word, size_t len, u
if (node.check != last_node_id) { if (node.check != last_node_id) {
log_debug("node.check = %d and last_node_id = %d\n", node.check, last_node_id); log_debug("node.check = %d and last_node_id = %d\n", node.check, last_node_id);
index = 0;
break; break;
} else if (node.base < 0) { } else if (node.base < 0) {
log_debug("Searching tail\n"); log_debug("Searching tail\n");
@@ -483,6 +486,7 @@ phrase_t trie_search_suffixes_from_index(trie_t *self, char *word, size_t len, u
phrase_start = index; phrase_start = index;
phrase_len = len - index; phrase_len = len - index;
value = tail_value; value = tail_value;
index = 0;
break; break;
} }
} }
@@ -508,10 +512,12 @@ phrase_t trie_search_suffixes_from_index(trie_t *self, char *word, size_t len, u
inline phrase_t trie_search_suffixes_from_index_get_suffix_char(trie_t *self, char *word, size_t len, uint32_t start_node_id) { inline phrase_t trie_search_suffixes_from_index_get_suffix_char(trie_t *self, char *word, size_t len, uint32_t start_node_id) {
if (word == NULL || len == 0) return NULL_PHRASE; if (word == NULL || len == 0) return NULL_PHRASE;
trie_node_t node = trie_get_node(self, start_node_id); trie_node_t node = trie_get_node(self, start_node_id);
uint32_t node_id = trie_get_transition_index(self, node, TRIE_SUFFIX_CHAR); unsigned char suffix_char = TRIE_SUFFIX_CHAR[0];
uint32_t node_id = trie_get_transition_index(self, node, suffix_char);
node = trie_get_node(self, node_id); node = trie_get_node(self, node_id);
if (node.check != start_node_id) { if (node.check != start_node_id) {
log_debug("node.check != start_node_id\n");
return NULL_PHRASE; return NULL_PHRASE;
} }
@@ -639,7 +645,8 @@ phrase_t trie_search_prefixes_from_index(trie_t *self, char *word, size_t len, u
inline phrase_t trie_search_prefixes_from_index_get_prefix_char(trie_t *self, char *word, size_t len, uint32_t start_node_id) { inline phrase_t trie_search_prefixes_from_index_get_prefix_char(trie_t *self, char *word, size_t len, uint32_t start_node_id) {
trie_node_t node = trie_get_node(self, start_node_id); trie_node_t node = trie_get_node(self, start_node_id);
uint32_t node_id = trie_get_transition_index(self, node, TRIE_PREFIX_CHAR); unsigned char prefix_char = TRIE_PREFIX_CHAR[0];
uint32_t node_id = trie_get_transition_index(self, node, prefix_char);
node = trie_get_node(self, node_id); node = trie_get_node(self, node_id);
if (node.check != start_node_id) { if (node.check != start_node_id) {