spaCy/spacy/pt/__init__.py

21 lines
472 B
Python
Raw Normal View History

2016-12-17 14:25:44 +03:00
# encoding: utf8
from __future__ import unicode_literals, print_function
from os import path
from ..language import Language
2016-12-08 22:41:41 +03:00
from ..attrs import LANG
2016-12-18 18:54:19 +03:00
from .language_data import *
class Portuguese(Language):
lang = 'pt'
2016-12-08 22:41:41 +03:00
class Defaults(Language.Defaults):
lex_attr_getters = dict(Language.Defaults.lex_attr_getters)
lex_attr_getters[LANG] = lambda text: 'pt'
2016-12-08 22:41:41 +03:00
tokenizer_exceptions = TOKENIZER_EXCEPTIONS
stop_words = STOP_WORDS