spaCy/spacy/lang/pl/__init__.py

36 lines
1009 B
Python
Raw Normal View History

2017-05-12 10:25:37 +03:00
# coding: utf8
from __future__ import unicode_literals
2017-06-27 00:38:28 +03:00
from .tokenizer_exceptions import TOKENIZER_EXCEPTIONS
2019-02-08 16:14:49 +03:00
from .punctuation import TOKENIZER_INFIXES
from .tag_map import TAG_MAP
2017-05-12 10:25:37 +03:00
from .stop_words import STOP_WORDS
from .lex_attrs import LEX_ATTRS
2017-05-12 10:25:37 +03:00
from ..tokenizer_exceptions import BASE_EXCEPTIONS
from ..norm_exceptions import BASE_NORMS
2017-05-12 10:25:37 +03:00
from ...language import Language
from ...attrs import LANG, NORM
from ...util import update_exc, add_lookups
2017-05-12 10:25:37 +03:00
class PolishDefaults(Language.Defaults):
lex_attr_getters = dict(Language.Defaults.lex_attr_getters)
lex_attr_getters.update(LEX_ATTRS)
lex_attr_getters[LANG] = lambda text: "pl"
lex_attr_getters[NORM] = add_lookups(
Language.Defaults.lex_attr_getters[NORM], BASE_NORMS
)
2017-06-27 00:42:27 +03:00
tokenizer_exceptions = update_exc(BASE_EXCEPTIONS, TOKENIZER_EXCEPTIONS)
stop_words = STOP_WORDS
tag_map = TAG_MAP
2019-02-08 16:14:49 +03:00
infixes = TOKENIZER_INFIXES
2017-05-12 10:25:37 +03:00
class Polish(Language):
lang = "pl"
Defaults = PolishDefaults
2017-05-12 10:25:37 +03:00
__all__ = ["Polish"]