2014-12-19 19:54:03 +00:00
|
|
|
from libcpp.vector cimport vector
|
|
|
|
|
|
|
|
|
|
from preshed.maps cimport PreshMap
|
|
|
|
|
from cymem.cymem cimport Pool
|
|
|
|
|
from murmurhash.mrmr cimport hash64
|
|
|
|
|
|
2015-01-11 23:26:22 +00:00
|
|
|
from .structs cimport LexemeC, TokenC, UniStr
|
2014-12-19 19:54:03 +00:00
|
|
|
from .typedefs cimport utf8_t, id_t, hash_t
|
|
|
|
|
from .strings cimport StringStore
|
|
|
|
|
|
|
|
|
|
|
2015-01-11 23:26:22 +00:00
|
|
|
cdef LexemeC EMPTY_LEXEME
|
2014-12-24 06:42:00 +00:00
|
|
|
|
|
|
|
|
|
2014-12-19 19:54:03 +00:00
|
|
|
cdef union LexemesOrTokens:
|
2015-01-11 23:26:22 +00:00
|
|
|
const LexemeC* const* lexemes
|
2014-12-19 19:54:03 +00:00
|
|
|
TokenC* tokens
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
cdef struct _Cached:
|
|
|
|
|
LexemesOrTokens data
|
|
|
|
|
bint is_lex
|
|
|
|
|
int length
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
cdef class Vocab:
|
2015-01-13 13:03:48 +00:00
|
|
|
cpdef public lexeme_props_getter
|
2014-12-19 19:54:03 +00:00
|
|
|
cdef Pool mem
|
|
|
|
|
cpdef readonly StringStore strings
|
2015-01-13 13:03:48 +00:00
|
|
|
cdef vector[const LexemeC*] lexemes
|
2015-07-08 10:36:38 +00:00
|
|
|
cdef readonly object pos_tags
|
2014-12-19 19:54:03 +00:00
|
|
|
|
2015-01-11 23:26:22 +00:00
|
|
|
cdef const LexemeC* get(self, Pool mem, UniStr* s) except NULL
|
2015-01-13 13:03:48 +00:00
|
|
|
cdef int _add_lex_to_vocab(self, hash_t key, const LexemeC* lex) except -1
|
2015-04-19 08:31:31 +00:00
|
|
|
|
2014-12-19 19:54:03 +00:00
|
|
|
cdef PreshMap _map
|
2015-06-05 14:26:39 +00:00
|
|
|
cdef readonly int repvec_length
|
