Page MenuHomePhabricator

enwiki article count is severely wrong on www.wikipedia.org
Closed, ResolvedPublicBUG REPORT

Description

Steps to replicate the issue (include links if applicable):

What happens?:

  • See it say 657 000+

What should have happened instead?:

  • It should say 6 579 000+

Software version (skip for WMF-hosted wikis like Wikipedia):
WMF production

Other information (browser name/version, screenshots, etc.):
See article count in https://en.wikipedia.org/wiki/Special:Statistics

image.png (266×495 px, 18 KB)

Event Timeline

rWPRD1c5457bac701: Assets build - 2022-11-21 09:23:50+00:00 looks to have broken it.

<small><bdi dir="ltr">6&nbsp;458&nbsp;000+</bdi> <span>articles</span></small>

became

<small><bdi dir="ltr">657&nbsp;000+</bdi> <span>articles</span></small>

Root cause seems to be that the build process draws the statistics from https://pagecounts.toolforge.org/pagecounts.json, which hasn't been updated to the sharding introduced in T306589, and so only gets one of the shards (which is also why the number is almost exactly one order of magnitude too small). https://github.com/MaxSem/pagecounts/pull/2 updates the query to deal with this; @Jdrewniak and @MaxSem are maintainers of the Toolforge tool.

Legoktm triaged this task as High priority.Nov 24 2022, 4:51 AM

oh boy, that looks bad. I've somehow lost access to the pagecounts repository on Github, but in the interest of making that code more visible, I think it'd be a good idea to merge it into the main Portals repository. @Vahurzpu thank you for the fix, I'm going to update the toolforge instance with that code shortly, followed by a patch that merges that script in the main portals repository.

Change 861426 had a related patch set uploaded (by Jdrewniak; author: Jdrewniak):

[operations/mediawiki-config@master] Bumping portals to master

https://gerrit.wikimedia.org/r/861426

Change 861426 merged by jenkins-bot:

[operations/mediawiki-config@master] Bumping portals to master

https://gerrit.wikimedia.org/r/861426

Confirming the count was fixed. @Jdrewniak do you want to close this or was there still followup work to do?

Ladsgroup assigned this task to Jdrewniak.