gi11es / articlequality

A library for performing automatic detection of assessment classes of Wikipedia article text.

Wikipedia article quality classification

This library provides a set of utilities for performing automatic detection of assessment classes of Wikipedia articles. For more information, see the full documentation at http://pythonhosted.org/wikiclass.

Compatible with Python 3.x only. Sorry.

• Install: pip install wikiclass
• Models: https://github.com/wikimedia/articlequality/tree/master/models
• Documentation: http://pythonhosted.org/wikiclass

Basic usage

>>> import articlequality
>>> from revscoring import Model
>>>
>>> scorer_model = Model.load(open("models/enwiki.nettrom_wp10.gradient_boosting.model", "rb"))
>>>
>>> text = "I am the text of a page.  I have a <ref>word</ref>"
>>> articlequality.score(scorer_model, text)
{'prediction': 'stub',
 'probability': {'stub': 0.27156163795807853,
                 'b': 0.14707452309674252,
                 'fa': 0.16844898943510833,
                 'c': 0.057668704007171959,
                 'ga': 0.21617801281707663,
                 'start': 0.13906813268582238}}

Authors

• Aaron Halfaker -- https://github.com/halfak
• Morten Warncke-Wang -- http://www-users.cs.umn.edu/~morten