mirror of https://github.com/explosion/spaCy.git
48 lines
1.1 KiB
Python
48 lines
1.1 KiB
Python
from typing import Optional, Callable
|
|
from thinc.api import Model
|
|
from .tokenizer_exceptions import TOKENIZER_EXCEPTIONS
|
|
from .punctuation import TOKENIZER_PREFIXES, TOKENIZER_SUFFIXES, TOKENIZER_INFIXES
|
|
from .stop_words import STOP_WORDS
|
|
from ...language import Language
|
|
from ...pipeline import Lemmatizer
|
|
|
|
|
|
class BengaliDefaults(Language.Defaults):
|
|
tokenizer_exceptions = TOKENIZER_EXCEPTIONS
|
|
prefixes = TOKENIZER_PREFIXES
|
|
suffixes = TOKENIZER_SUFFIXES
|
|
infixes = TOKENIZER_INFIXES
|
|
stop_words = STOP_WORDS
|
|
|
|
|
|
class Bengali(Language):
|
|
lang = "bn"
|
|
Defaults = BengaliDefaults
|
|
|
|
|
|
@Bengali.factory(
|
|
"lemmatizer",
|
|
assigns=["token.lemma"],
|
|
default_config={
|
|
"model": None,
|
|
"mode": "rule",
|
|
"overwrite": False,
|
|
"scorer": {"@scorers": "spacy.lemmatizer_scorer.v1"},
|
|
},
|
|
default_score_weights={"lemma_acc": 1.0},
|
|
)
|
|
def make_lemmatizer(
|
|
nlp: Language,
|
|
model: Optional[Model],
|
|
name: str,
|
|
mode: str,
|
|
overwrite: bool,
|
|
scorer: Optional[Callable],
|
|
):
|
|
return Lemmatizer(
|
|
nlp.vocab, model, name, mode=mode, overwrite=overwrite, scorer=scorer
|
|
)
|
|
|
|
|
|
__all__ = ["Bengali"]
|