(转载) 安装nltk遇到的坑:下载失败+安装成功但导入不了

作者:野指针呀 

地址:https://blog.csdn.net/mjj1024/article/details/105618784

 

在jupyter notebook上运行代码时:

 

1     import nltk
2     paragraph = "i am a good boy ! are you ok? hahaha i am fine"
3     words_list = nltk.word_tokenize(paragraph)
4     print(words_list)

出现错误:

1     ModuleNotFoundError Traceback (most recent call last)
2     <ipython-input-4-55bf564de021> in <module>
3     ----> 1 import nltk
4     2 paragraph = "i am a good boy ! are you ok? hahaha i am fine"
5     3 words_list = nltk.word_tokenize(paragraph)
6     4 print(words_list)
7      
8     ModuleNotFoundError: No module named 'nltk'

 

显示没有nltk这个模块。

然后在cmd和conda里分别运行pip list和conda list,发现nltk都已经安装好。之后搜了一个博客,才知道还要下载nltk语料包。

尝试了一下自动下载:

在idle 3.7(按照自己电脑上的版本)中运行代码:

 

1     >>> import nltk
2     >>> nltk.download()

跳出来NlTK Downloder框,然后出现了getaddrinfo failed错误,貌似是服务器的问题。

之后看博客说是把NlTK Downloder里的Server Index的内容:

https://raw.githubusercontent.com/nltk/nltk_data/gh-pages/index.xml

换成:http://www.nltk.org/nltk_data/

点击下载时又出现错误相同错误:getaddrinfo failed

搜了一堆博客尝试后无果,没办法只能乖乖手动安装nltk。

手动安装有点麻烦,但也没办法。

不过有看到一个大佬写了个代码装的,看起来很厉害:

手动安装Python NLTK语言包

我自己是手动下载,然后解压。

在github上下载语料库:https://github.com/nltk/nltk_data

下载之后把里面的packages文件名改成nltk_data(里面的压缩包都要解压),然后放在该放的路径下。
查看该放的路径:可以先运行一段代码(在idle中运行),错误提示里会给出路径,比如下面:

 1     >>> import nltk
 2     >>> paragraph = "i am a good boy ! are you ok? hahaha i am fine"
 3     >>> words_list = nltk.word_tokenize(paragraph)
 4     Traceback (most recent call last):
 5     File "<pyshell#5>", line 1, in <module>
 6     words_list = nltk.word_tokenize(paragraph)
 7     File "C:\Program Files\Python37\lib\site-packages\nltk\tokenize\__init__.py", line 144, in word_tokenize
 8     sentences = [text] if preserve_line else sent_tokenize(text, language)
 9     File "C:\Program Files\Python37\lib\site-packages\nltk\tokenize\__init__.py", line 105, in sent_tokenize
10     tokenizer = load('tokenizers/punkt/{0}.pickle'.format(language))
11     File "C:\Program Files\Python37\lib\site-packages\nltk\data.py", line 868, in load
12     opened_resource = _open(resource_url)
13     File "C:\Program Files\Python37\lib\site-packages\nltk\data.py", line 993, in _open
14     return find(path_, path + ['']).open()
15     File "C:\Program Files\Python37\lib\site-packages\nltk\data.py", line 701, in find
16     raise LookupError(resource_not_found)
17     LookupError:
18     **********************************************************************
19     Resource [93mpunkt[0m not found.
20     Please use the NLTK Downloader to obtain the resource:
21      
22     [31m>>> import nltk
23     >>> nltk.download('punkt')
24     [0m
25     For more information see: https://www.nltk.org/data.html
26      
27     Attempted to load [93mtokenizers/punkt/english.pickle[0m
28      
29     Searched in:
30     - 'C:\\Users\\马静静/nltk_data'
31     - 'C:\\Program Files\\Python37\\nltk_data'
32     - 'C:\\Program Files\\Python37\\share\\nltk_data'
33     - 'C:\\Program Files\\Python37\\lib\\nltk_data'
34     - 'C:\\Users\\马静静\\AppData\\Roaming\\nltk_data'
35     - 'C:\\nltk_data'
36     - 'D:\\nltk_data'
37     - 'E:\\nltk_data'
38     - ''
39     **********************************************************************

 

这一部分就是可以放nltk_data的路径:

 1     Searched in:
 2     - 'C:\\Users\\马静静/nltk_data'
 3     - 'C:\\Program Files\\Python37\\nltk_data'
 4     - 'C:\\Program Files\\Python37\\share\\nltk_data'
 5     - 'C:\\Program Files\\Python37\\lib\\nltk_data'
 6     - 'C:\\Users\\马静静\\AppData\\Roaming\\nltk_data'
 7     - 'C:\\nltk_data'
 8     - 'D:\\nltk_data'
 9     - 'E:\\nltk_data'
10     - ''

 

我解压完后,直接把文件夹 ( packages文件名改成nltk_data的文件夹 ) 放在C:\\Users\\马静静\\下。
再运行代码就ok了。

1     >>> import nltk
2     >>> paragraph = "i am a good boy ! are you ok? hahaha i am fine"
3     >>> words_list = nltk.word_tokenize(paragraph)
4     >>> print(words_list)
5     ['i', 'am', 'a', 'good', 'boy', '!', 'are', 'you', 'ok', '?', 'hahaha', 'i', 'am', 'fine']

 

但是,又发现在idle中和pycharm中运行都没有问题,在jupyter notebook中运行依然是找不到模块。

又搜了一堆博客,大概明白了是因为jupyter和python下载包的路径不同,导致在jupyter运行时找不到包。

尝试过重新install nltk、修改环境变量等操作依然无果,后来看到一个博客(python Jupyter不能导入外部包----解决方案)找到了解决方法,感谢这位小可爱。

posted @ 2020-10-29 19:59  liweikuan  阅读(2828)  评论(0编辑  收藏  举报