2015-09-06 22:56:47 +03:00
|
|
|
from __future__ import unicode_literals, print_function
|
|
|
|
|
|
|
|
from os import path
|
|
|
|
|
|
|
|
from ..language import Language
|
2016-09-24 21:26:17 +03:00
|
|
|
from ..vocab import Vocab
|
|
|
|
from ..attrs import LANG
|
2015-09-06 22:56:47 +03:00
|
|
|
|
|
|
|
|
|
|
|
class German(Language):
|
2016-03-03 17:21:00 +03:00
|
|
|
lang = 'de'
|
2016-09-24 21:26:17 +03:00
|
|
|
|
|
|
|
class Defaults(Language.Defaults):
|
|
|
|
def Vocab(self, vectors=None, lex_attr_getters=None):
|
|
|
|
if lex_attr_getters is None:
|
|
|
|
lex_attr_getters = dict(self.lex_attr_getters)
|
|
|
|
if vectors is None:
|
|
|
|
vectors = self.Vectors()
|
|
|
|
# set a dummy lemmatizer for now that simply returns the same string
|
|
|
|
# until the morphology is done for German
|
2016-09-25 15:49:53 +03:00
|
|
|
return Vocab.load(self.path, lex_attr_getters=lex_attr_getters, vectors=vectors,
|
2016-09-24 21:26:17 +03:00
|
|
|
lemmatizer=False)
|
2016-05-02 17:04:53 +03:00
|
|
|
|
2016-09-24 21:26:17 +03:00
|
|
|
stop_words = set()
|