Revision Scoring

A generic, machine learning-based revision scoring system designed to be used to automatically differentiate damage from productive contributory behavior on Wikipedia.

Examples

Scoring models:

>>> from mw.api import Session
>>>
>>> from revscoring.extractors import APIExtractor
>>> from revscoring.languages import english
>>> from revscoring.scorers import MLScorerModel
>>>
>>> api_session = Session("https://en.wikipedia.org/w/api.php")
Sending requests with default User-Agent.  Set 'user_agent' on api.Session to quiet this message.
>>> extractor = APIExtractor(api_session, english)
>>>
>>> filename = "models/reverts.halfak_mix.trained.model"
>>> model = MLScorerModel.load(open(filename, 'rb'))
>>>
>>> rev_ids = [105, 642215410, 638307884]
>>> feature_values = [extractor.extract(id, model.features) for id in rev_ids]

>>> scores = model.score(feature_values, probabilities=True)
>>> for rev_id, score in zip(rev_ids, scores):
...     print("{0}: {1}".format(rev_id, score))
...
105: {'probabilities': array([ 0.96441465,  0.03558535]), 'prediction': False}
642215410: {'probabilities': array([ 0.75884553,  0.24115447]), 'prediction': True}
638307884: {'probabilities': array([ 0.98441738,  0.01558262]), 'prediction': False}

Feature extraction:

>>> from mw.api import Session
>>>
>>> from revscoring.extractors import APIExtractor
>>> from revscoring.features import diff, parent_revision, revision, user
>>>
>>> api_extractor = APIExtractor(Session("https://en.wikipedia.org/w/api.php"))
Sending requests with default User-Agent.  Set 'user_agent' on api.Session to quiet this message.
>>>
>>> features = [revision.day_of_week,
...             revision.hour_of_day,
...             revision.has_custom_comment,
...             parent_revision.bytes_changed,
...             diff.chars_added,
...             user.age,
...             user.is_anon,
...             user.is_bot]
>>>
>>> values = api_extractor.extract(
...     624577024,
...     features
... )
>>> for feature, value in zip(features, values):
...     print("{0}: {1}".format(feature, value))
...
<revision.day_of_week>: 6
<revision.hour_of_day>: 19
<revision.has_custom_comment>: True
<(revision.bytes - parent_revision.bytes_changed)>: 3
<diff.chars_added>: 8
<user.age>: 71821407
<user.is_anon>: False
<user.is_bot>: False

Installation

Packages

In order to use this, you need to install a few packages first:

You might need to install some other dependencies depending on your operating system. Try using the packages,

sudo apt-get install python3-dev python3-numpy python3-scipy g++ gfortran liblapack-dev libopenblas-dev myspell-pt myspell-fa myspell-en-au myspell-en-gb myspell-en-us myspell-en-za myspell-fr myspell-es hunspell-vi myspell-he

If you're on Ubuntu, you might also be able to install an Indonesian dictionary:

sudo apt-get install aspell-id

Virtualenv users, please note that you'll have to use the --system-site-packages option if you install scipy and numpy via apt-get. You can also use pip3 within your virtualenv.

Python modules

If you need the Python package installer,

sudo easy_install3 pip

Then, install this module,

pip3 install --user revscoring

You'll need to download NLTK data in order to make use of language features.

python3 -m nltk.downloader wordnet omw stopwords

Authors

Aaron Halfaker:

• http://halfaker.info

Helder:

• https://github.com/he7d3r

Adam Roses Wight:

• https://mediawiki.org/wiki/User:Adamw