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

[expansion] using easier-to-access data structure for address dictionaries

Browse Source
This commit is contained in:
Al
2016-11-27 00:56:20 -08:00
parent 7de2aa21cd
commit eea11beb6a
4 changed files with 160 additions and 190 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
src/address_dictionary.h
View File

@@ -25,32 +25,31 @@
#define NULL_CANONICAL_INDEX -1
typedef union expansion_value {
uint32_t value;
struct {
uint32_t components:16;
uint32_t count:14;
uint32_t canonical:1;
uint32_t separable:1;
};
} expansion_value_t;
typedef struct address_expansion {
int32_t canonical_index;
char language[MAX_LANGUAGE_LEN];
uint32_t num_dictionaries;
uint16_t dictionary_ids[MAX_DICTIONARY_TYPES];
uint16_t address_components;
uint32_t address_components;
bool separable;
} address_expansion_t;
VECTOR_INIT(address_expansion_array, address_expansion_t)
KHASH_MAP_INIT_STR(str_expansions, address_expansion_array *)
typedef struct address_expansion_value {
uint32_t components;
address_expansion_array *expansions;
} address_expansion_value_t;
address_expansion_value_t *address_expansion_value_new(void);
address_expansion_value_t *address_expansion_value_new_with_expansion(address_expansion_t expansion);
void address_expansion_value_destroy(address_expansion_value_t *self);
VECTOR_INIT_FREE_DATA(address_expansion_value_array, address_expansion_value_t *, address_expansion_value_destroy)
typedef struct address_dictionary {
cstring_array *canonical;
khash_t(str_expansions) *expansions;
address_expansion_value_array *values;
trie_t *trie;
} address_dictionary_t;
@@ -66,7 +65,7 @@ bool search_address_dictionaries_tokens_with_phrases(char *str, token_array *tok
phrase_t search_address_dictionaries_prefix(char *str, size_t len, char *lang);
phrase_t search_address_dictionaries_suffix(char *str, size_t len, char *lang);
address_expansion_array *address_dictionary_get_expansions(char *key);
address_expansion_value_t *address_dictionary_get_expansions(uint32_t i);
bool address_expansion_in_dictionary(address_expansion_t expansion, uint16_t dictionary_id);
char *address_dictionary_get_canonical(uint32_t index);
int32_t address_dictionary_next_canonical_index(void);
@@ -81,7 +80,4 @@ bool address_dictionary_save(char *path);
bool address_dictionary_module_setup(char *filename);
void address_dictionary_module_teardown(void);
#endif