Wikidossier – Tools to stalk Wikipedia users and pages

You can see the app deployed on Heroku. For example:

• https://wikidossier.herokuapp.com/user/Riceissa
• https://wikidossier.herokuapp.com/usercompare?usernames=Vipul|Riceissa|Ethanbas&ns=0

Usage

For testing/development:

$ flask initdb
$ echo "type the user agent you want to use here" > user-agent.txt
# The user agent is used for making queries to the Wikipedia API
$ flask run
# Then just visit http://127.0.0.1:5000/
# (or whatever the terminal output says) in your browser

For deployment on the server: (idk, it's been many years and i forgot how to do this.)

List of scripts

• sizediff.py: This script uses the MediaWiki API to obtain revisions by username, and stores it as a TSV. The columns are "username", "ns" (namespace number), "timestamp", and "sizediff" (difference in bytes of the page before and after the edit).
• revisions.py: This script uses the MediaWiki API to obtain revisions by page name and stores it as a TSV. The columns are "pagename", "timestamp", and "size".
• afd.py: This script gathers some information from AfDs, most notably the voting information (a list of tuples (v, u), where v is the vote – keep, delete, merge, whatever – and u is the user that cast the vote).
• plot.py: This script takes the data collected using the other scripts and plots them.
• wikidossier.py: The main Flask app

Notes

To filter out "strange" page titles, I used

/[^\d9\d32-\d126“”‘’–—§éñãí¡áóâ°üčō×½êàëżıå]
:g//d

in Vim, to find characters not in a whitelist, and then deleting all lines with characters outside of that list. We can work out something better later.

License

CC0; see LICENSE for the full text.