python - 使用 TreeTagger 进行标记时必须使用 *unicode* 字符串作为标记文本吗？

Question

来自TreeTagger's website我创建了一个目录并下载了指定的文件。然后treetaggerwrapper ，因此从 documentation我尝试测试并尝试如何标记一些文本，如下所示:

In [40]:

import treetaggerwrapper

tagger = treetaggerwrapper.TreeTagger(TAGLANG='en')

tags = tagger.TagText("This is a very short text to tag.")

print tags

然后我收到以下警告:

WARNING:TreeTagger:Abbreviation file not found: english-abbreviations
WARNING:TreeTagger:Processing without abbreviations file.
ERROR:TreeTagger:Must use *unicode* string as text to tag, not <type 'str'>.

---------------------------------------------------------------------------
TreeTaggerError                           Traceback (most recent call last)
<ipython-input-40-37b912126580> in <module>()
      1 import treetaggerwrapper
      2 tagger = treetaggerwrapper.TreeTagger(TAGLANG='en')
----> 3 tags = tagger.TagText("This is a very short text to tag.")
      4 print tags

/usr/local/lib/python2.7/site-packages/treetaggerwrapper.pyc in TagText(self, text, numlines, tagonly, prepronly, tagblanks, notagurl, notagemail, notagip, notagdns, encoding, errors)
   1236         return self.tag_text(text, numlines=numlines, tagonly=tagonly,
   1237                  prepronly=prepronly, tagblanks=tagblanks, notagurl=notagurl,
-> 1238                  notagemail=notagemail, notagip=notagip, notagdns=notagdns)
   1239 
   1240     # --------------------------------------------------------------------------

/usr/local/lib/python2.7/site-packages/treetaggerwrapper.pyc in tag_text(self, text, numlines, tagonly, prepronly, tagblanks, notagurl, notagemail, notagip, notagdns, nosgmlsplit)
   1302             # Raise exception now, with an explicit message.
   1303             logger.error("Must use *unicode* string as text to tag, not %s.", type(text))
-> 1304             raise TreeTaggerError("Must use *unicode* string as text to tag.")
   1305 
   1306         if isinstance(text, six.text_type):

TreeTaggerError: Must use *unicode* string as text to tag.

哪里可以下载英语和西类牙语的缩写文件？以及如何正确安装treetaggerwrapper？

Answer 1

该方法仅接受 unicode 字符串，在字符串中添加 u 使其成为 unicode 字符串:

tags = tagger.TagText(u"This is a very short text to tag.")

“这是要标记的非常短的文本。” 是 str 类型，一旦添加 u 它就是 unicode:

In [12]: type("This is a very short text to tag.")
Out[12]: str

In [13]: type(u"This is a very short text to tag.")
Out[13]: unicode

如果您从其他来源获取 str，则需要解码:

In [15]: s = "This is a very short text to tag."

In [16]: type(s)
Out[16]: str

In [17]: type(s.decode("utf-8"))
Out[17]: unicode

可以下载标记脚本here