lahwaacz/wiki-scripts

statistics.py: revise caching of user properties

Closed this issue · 0 comments

Since wiki-scripts now provide quite robust base for caching interface, it should be exploited by the statistics.py script for caching user properties:

  • initial query based on list=allusers (this is the slowest part of the script)
  • use list=recentchanges-based query to find active users in the past month (also used to calculate recent edits count)
  • use list=logevents to find new users and groups changes since the last update of the database
  • use list=users to update total edit count and other properties selectively for the found users
  • (possibly) if the database is older than the interval for active users, fall back to list=allusers to preserve integrity of the database

This should eliminate the need to parse the existing table, which depends on specific formatting of each column.