mirror of
https://github.com/explosion/spaCy.git
synced 2024-11-13 13:17:06 +03:00
90 lines
1.8 KiB
Cython
90 lines
1.8 KiB
Cython
from libc.stdint cimport int8_t, uint8_t, uint16_t, uint32_t
|
|
|
|
from .typedefs cimport flags_t, attr_t, id_t, hash_t
|
|
from .parts_of_speech cimport univ_pos_t
|
|
|
|
|
|
cdef struct LexemeC:
|
|
const float* repvec
|
|
|
|
flags_t flags
|
|
|
|
attr_t id
|
|
attr_t length
|
|
|
|
attr_t orth
|
|
attr_t lower
|
|
attr_t norm
|
|
attr_t shape
|
|
attr_t prefix
|
|
attr_t suffix
|
|
|
|
attr_t cluster
|
|
|
|
float prob
|
|
float sentiment
|
|
float l2_norm
|
|
|
|
|
|
cdef struct Morphology:
|
|
uint8_t number
|
|
uint8_t tenspect # Tense/aspect/voice
|
|
uint8_t mood
|
|
uint8_t gender
|
|
uint8_t person
|
|
uint8_t case
|
|
uint8_t misc
|
|
|
|
|
|
cdef struct PosTag:
|
|
Morphology morph
|
|
int id
|
|
univ_pos_t pos
|
|
|
|
|
|
# Start and end will be offsets: i + ent.start will always take you to the
|
|
# "next" entity start. If inside an entity, ent.start will be negative ---
|
|
# the next entity is the start of the one the token is inside. If i _is_
|
|
# the start of an entity, then ent.start will be the beginning of the next one.
|
|
#
|
|
# The same/inverse is true for end. If ent.end has a negative value, we are either
|
|
# at the end of an entity, or outside one. If we're inside an entity, ent.end
|
|
# will have a positive value.
|
|
#
|
|
# This allows us to easily find the span of an entity we might be inside, while
|
|
# naturally sharing an API with iterating through all entities in the sentence
|
|
cdef struct Entity:
|
|
int32_t tag
|
|
uint16_t flags
|
|
int8_t start
|
|
int8_t end
|
|
|
|
|
|
cdef struct TokenC:
|
|
const LexemeC* lex
|
|
Morphology morph
|
|
Entity ent
|
|
univ_pos_t pos
|
|
int tag
|
|
int idx
|
|
int lemma
|
|
int sense
|
|
int head
|
|
int dep
|
|
bint sent_end
|
|
uint32_t l_kids
|
|
uint32_t r_kids
|
|
|
|
|
|
cdef struct Utf8Str:
|
|
id_t i
|
|
hash_t key
|
|
unsigned char* chars
|
|
int length
|
|
|
|
|
|
cdef struct UniStr:
|
|
Py_UNICODE* chars
|
|
size_t n
|
|
hash_t key
|