我正在尝试做预处理功能,只允许我传递文本语料库,然后我就可以得到干净的文本
from contractions import CONTRACTION_MAP
import re
import nltk
from nltk.corpus import wordnet
from pattern.en import suggest
from nltk.stem import PorterStemmer
data_cleaning():
punkt_token = nltk.PunktSentenceTokenizer()
punkt_token_output = punkt_token.tokenize(sample_text)
word_tokenize = nltk.word_tokenize
word_tokenize_ouput = word_tokenize(sentence)
#Removing Special Character
pattern = re.compile("[{}]".format(re.escape(string.punctuation)))
filter_text = filter(None,[pattern.sub('' ,c) for c in corpus])
output = " ".join(filter_text)
def expand_contractions(text) :
pattern = re.compile(({})".format("|".join(CONTRACTION_MAP.keys())),flags=
re.DOTALL| re.IGNORECASE)
def replace_text(t):
txt = t.group(0)
if txt.lower() in CONTRACTION_MAP.keys():
return CONTRACTION_MAP[txt.lower()]
expand_text = pattern.sub(replace_text,text)
return expand_text
corpus = expand_contraction = [expand_contractions(c) for c in corpus]
目前没有回答
相关问题 更多 >
编程相关推荐