mirror of
https://github.com/explosion/spaCy.git
synced 2024-11-14 05:37:03 +03:00
139 lines
3.8 KiB
Python
139 lines
3.8 KiB
Python
from typing import Callable, Protocol, Iterator, Optional, Union, Tuple, Any, overload
|
|
from thinc.types import Floats1d, Ints2d, FloatsXd
|
|
from .doc import Doc
|
|
from .token import Token
|
|
from .underscore import Underscore
|
|
from ..lexeme import Lexeme
|
|
from ..vocab import Vocab
|
|
|
|
class SpanMethod(Protocol):
|
|
def __call__(self: Span, *args: Any, **kwargs: Any) -> Any: ... # type: ignore[misc]
|
|
|
|
class Span:
|
|
@classmethod
|
|
def set_extension(
|
|
cls,
|
|
name: str,
|
|
default: Optional[Any] = ...,
|
|
getter: Optional[Callable[[Span], Any]] = ...,
|
|
setter: Optional[Callable[[Span, Any], None]] = ...,
|
|
method: Optional[SpanMethod] = ...,
|
|
force: bool = ...,
|
|
) -> None: ...
|
|
@classmethod
|
|
def get_extension(
|
|
cls, name: str
|
|
) -> Tuple[
|
|
Optional[Any],
|
|
Optional[SpanMethod],
|
|
Optional[Callable[[Span], Any]],
|
|
Optional[Callable[[Span, Any], None]],
|
|
]: ...
|
|
@classmethod
|
|
def has_extension(cls, name: str) -> bool: ...
|
|
@classmethod
|
|
def remove_extension(
|
|
cls, name: str
|
|
) -> Tuple[
|
|
Optional[Any],
|
|
Optional[SpanMethod],
|
|
Optional[Callable[[Span], Any]],
|
|
Optional[Callable[[Span, Any], None]],
|
|
]: ...
|
|
def __init__(
|
|
self,
|
|
doc: Doc,
|
|
start: int,
|
|
end: int,
|
|
label: Union[str, int] = ...,
|
|
vector: Optional[Floats1d] = ...,
|
|
vector_norm: Optional[float] = ...,
|
|
kb_id: Union[str, int] = ...,
|
|
span_id: Union[str, int] = ...,
|
|
) -> None: ...
|
|
def __richcmp__(self, other: Span, op: int) -> bool: ...
|
|
def __hash__(self) -> int: ...
|
|
def __len__(self) -> int: ...
|
|
def __repr__(self) -> str: ...
|
|
@overload
|
|
def __getitem__(self, i: int) -> Token: ...
|
|
@overload
|
|
def __getitem__(self, i: slice) -> Span: ...
|
|
def __iter__(self) -> Iterator[Token]: ...
|
|
@property
|
|
def _(self) -> Underscore: ...
|
|
def as_doc(self, *, copy_user_data: bool = ...) -> Doc: ...
|
|
def get_lca_matrix(self) -> Ints2d: ...
|
|
def similarity(self, other: Union[Doc, Span, Token, Lexeme]) -> float: ...
|
|
@property
|
|
def doc(self) -> Doc: ...
|
|
@property
|
|
def vocab(self) -> Vocab: ...
|
|
@property
|
|
def sent(self) -> Span: ...
|
|
@property
|
|
def ents(self) -> Tuple[Span]: ...
|
|
@property
|
|
def has_vector(self) -> bool: ...
|
|
@property
|
|
def vector(self) -> Floats1d: ...
|
|
@property
|
|
def vector_norm(self) -> float: ...
|
|
@property
|
|
def tensor(self) -> FloatsXd: ...
|
|
@property
|
|
def text(self) -> str: ...
|
|
@property
|
|
def text_with_ws(self) -> str: ...
|
|
@property
|
|
def noun_chunks(self) -> Iterator[Span]: ...
|
|
@property
|
|
def root(self) -> Token: ...
|
|
def char_span(
|
|
self,
|
|
start_idx: int,
|
|
end_idx: int,
|
|
label: Union[int, str] = ...,
|
|
kb_id: Union[int, str] = ...,
|
|
vector: Optional[Floats1d] = ...,
|
|
id: Union[int, str] = ...,
|
|
alignment_mode: str = ...,
|
|
span_id: Union[int, str] = ...,
|
|
) -> Span: ...
|
|
@property
|
|
def conjuncts(self) -> Tuple[Token]: ...
|
|
@property
|
|
def lefts(self) -> Iterator[Token]: ...
|
|
@property
|
|
def rights(self) -> Iterator[Token]: ...
|
|
@property
|
|
def n_lefts(self) -> int: ...
|
|
@property
|
|
def n_rights(self) -> int: ...
|
|
@property
|
|
def subtree(self) -> Iterator[Token]: ...
|
|
start: int
|
|
end: int
|
|
start_char: int
|
|
end_char: int
|
|
@property
|
|
def label(self) -> int: ...
|
|
@property
|
|
def kb_id(self) -> int: ...
|
|
@property
|
|
def id(self) -> int: ...
|
|
@property
|
|
def ent_id(self) -> int: ...
|
|
@property
|
|
def orth_(self) -> str: ...
|
|
@property
|
|
def lemma_(self) -> str: ...
|
|
@property
|
|
def label_(self) -> str: ...
|
|
@property
|
|
def kb_id_(self) -> str: ...
|
|
@property
|
|
def id_(self) -> str: ...
|
|
@property
|
|
def ent_id_(self) -> str: ...
|