Page MenuHomePhabricator

Patroller Tasks: Filters screen for List of Edits
Closed, ResolvedPublic

Description

This is to build the comprehensive Filters activity, which will allow the user to select the types of edits they want to see, and from which Wiki.


Designs (Figma)
7. Default filters8. Ability to reset filters9. Selected filters10. No filter results
Default filters.png (720×4,592 px, 232 KB)
Reset filters.png (720×4,592 px, 235 KB)
No results.png (720×1,440 px, 28 KB)
No results.png (720×1,440 px, 28 KB)
  1. Default filters
  • The UI is similar to ‘Watchlist’ filters. Some filters are borrowed from ‘Watchlist’ filters and others are new.
  • Filters include:
    • Wiki (language filters) with the ability to ‘Update app languages’ (the following titles and descriptions borrowed from web’s ‘Recent changes’ filters)
    • User registration and experience
      • Title: Unregistered
        • Description: Editors who aren’t logged in.
      • Title: Registered
        • Description: Logged-in editors.
      • Title: Newcomers
        • Description: Registered editors who have fewer than 10 edits of 4 days of activity.
      • Title: Learners
        • Description: Registered editors whose experience falls between “Newcomers” and “Experienced users.”
      • Title: Experienced users
        • Description: Registered editors with more than 500 edits and 30 days of activity.
    • Latest revisions
      • All
      • Latest revision
    • Not the latest revision
      • Automated contributions
      • All
      • Bot
      • Human (not bot)
    • Contribution quality predictions
      • Title: Very likely good
        • Description: Highly accurate at finding almost all problem-free edits.
      • Title: May have problems
        • Description: Finds most flawed or damaging edits but with lower accuracy.
      • Title: Likely to have problems
        • Description: With medium accuracy, finds as intermediate fraction of problem edits.
      • Title: Very likely to have problems
        • Description: Very highly accurate at finding the most obviously flawed or damaging edits.
    • User intent predictions
      • Title: Very likely good faith
        • Description: Highly accurate at finding almost all good-faith edits.
      • Title: May be bad faith
        • Description: Finds most bad-faith edits but with a lower accuracy.
        • Title: Likely to bad faith
        • Description: With medium accuracy, finds an intermediate fraction of bad-faith edits.
        • Title: Very likely bad faith
        • Description: Very highly accurate at finding the most obvious bad-faith edits.
    • Significance
      • All
      • Minor edits
      • Non-minor edits
  1. Reset filters
  • When new filters are selected a ‘Reset’ button will appear in the top corner to allow patrollers to reset all filters to default state.
  1. Selected filters
  • When filters are selected and patroller returns to ‘List of edits’ page the number of selected filters is indicated by a value next to the ‘filter’ icon. The icon and the value become blue.
  1. No filter results
  • If a set of filters doesn’t produce results, an illustration & message will appear.
  • Copy: Try removing [x] filters to see more edits

APK: https://github.com/wikimedia/apps-android-wikipedia/pull/4176

Event Timeline

Dbrant triaged this task as Medium priority.Aug 8 2023, 5:38 PM

Hi @OTichonova

Per discussion from engineers (cc @Dbrant and @Sharvaniharan), would it be possible to remove New comers, Learners, and Experienced users from the User registration and experience category?

The three options will require extra API calls for each valid user and will increase the server load when trying to fetch the information. Also, the local filtering process might slow down the scrolling experience.

And most importantly, removing the three options will align the filter with the Watchlist filter in the app.

cc @JTannerWMF

Hi @OTichonova

The implementation is completed. Please download the APK from the ticket description to test it. Thanks!

I have included the User experience filter here but this will need to be reviewed by other engineers to see if it causes any performance or server load balance issues. T346192