Please include the license for averaged_perceptron_tagger

[Hiroshiba]

The list of resources available for download from NLTK does not appear to mention the license for averaged_perceptron_tagger.

Could you clarify what license averaged_perceptron_tagger is distributed under?

---

I did some research on my own and hope this information might be helpful:

• It seems likely that the license is MIT, coming from the sloria/textblob-aptagger repository.
• Here are some relevant issues and pull requests:
  • Issue #1110 where POS improvements were discussed and a solution was proposed.
  • Issue #1122 proposing the use of TextBlob PerceptronTagger.
  • Pull Request #1143, which seems related to implementing the above proposal, though it may not have been merged directly or was replaced by a different commit.

From what I can tell, sloria/textblob-aptagger is MIT licensed, and NLTK might be distributing only the necessary pickle files from that repository as part of nltk_data. If that’s correct, it follows that averaged_perceptron_tagger would also be MIT licensed.

If so, it would be helpful to clearly state this in taggers/averaged_perceptron_tagger.xml. For reference on how to include such information in the XML, the cmudict.xml file might be a good example.

Thanks in advance for looking into this!

[ekaf]

Thanks @Hiroshiba. Decompiling the TextBlob pickle reveals that it has different weights than the NLTK package.

Unfortunately, the conversation in PR #1143, and particularly this comment seems to only clarify the license for the Python code, and may not necessarily cover the data package.

[Hiroshiba]

I didn't notice that the comment had been edited and ended up creating a pull request:
#233

[textblob-aptagger](https://github.com/sloria/textblob-aptagger) is under the MIT license, and typically, if nothing else is specified, the binary files included in it are also considered to be under the MIT license. Therefore, I interpreted the pickle files as being covered by the MIT license as well. (Of course, I'm not a legal expert, so I might be mistaken, but that's my general understanding.)