I am working with a text in utf-8. I want to tokenize it and then convert it into a list. However I get the following error.
import nltk, jieba, re, os
with open('file.txt') as f:
tokenized_text = jieba.cut(f,cut_all=True)
type(tokenized_text)
generator
word_list = list(tokenized_text)
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
<ipython-input-5-16b25477c71d> in <module>()
----> 1 list(new)
~/anaconda3/lib/python3.6/site-packages/jieba/__init__.py in cut(self, sentence, cut_all, HMM)
280 - HMM: Whether to use the Hidden Markov Model.
281 '''
--> 282 sentence = strdecode(sentence)
283
284 if cut_all:
~/anaconda3/lib/python3.6/site-packages/jieba/_compat.py in strdecode(sentence)
35 if not isinstance(sentence, text_type):
36 try:
---> 37 sentence = sentence.decode('utf-8')
38 except UnicodeDecodeError:
39 sentence = sentence.decode('gbk', 'ignore')
AttributeError: '_io.TextIOWrapper' object has no attribute 'decode'
I understand the problem lies somewhere in the jieba package. I also tried to change the code into
with open('file.txt') as f:
new = jieba.cut(f,cut_all=False)
but got the same result.
jieba.cut
takes a string, not a file. This is explained in the readme.