Page MenuHomePhabricator

Newcomer tasks: create test pages on beta enwiki
Closed, ResolvedPublic

Description

Testing newcomer tasks involves many test pages (hundreds, if we also want to test the 200 cutoff), preferably with enough content for page images and text extracts. That can't really be set up by hand so we should do it via bot.

Event Timeline

Alternatively we could make pageimages, textextracts, the page protection check etc. all use the remote wiki. Which would be a significant change to the architecture but maybe worth it in the long term.

Mentioned in SAL (#wikimedia-cloud) [2019-10-31T01:14:27Z] <tgr> importing a bunch of pages from production cswiki via importDump.php for T236823

Mentioned in SAL (#wikimedia-cloud) [2019-10-31T20:29:26Z] <tgr> importing a bunch of pages from production cswiki via importDump.php for T236823 (for reals now)

I followed @Etonkovidova's advice and took a simpler approach:

  • went to Special:Export on cswiki, added some maintenance templates (Kdy?, Kdo?, Pravopis, Wikifikovat) and the contents of the categories they put articles into (via the "add pages from category" feature), and downloaded the dump
  • (tried Special:Import on beta but unsurprisingly that timed out)
  • uploaded the dump file to beta tin (deployment-deploy01) and ran mwscript importDump.php enwiki --report 100 --username-prefix cs --no-local-users /tmp/<filename> (user directories are not public so had to move it to tmp)
  • ran mwscript rebuildrecentchanges.php enwiki and mwscript initSiteStats.php enwiki --update to keep things consistent.
  • went through the templates and fixed category references (e.g. replaced {{Kategorie|Údržba:Články s vyhýbavými slovy}} with [[Category:Údržba:Články s vyhýbavými slovy]])

Change 547648 had a related patch set uploaded (by Gergő Tisza; owner: Gergő Tisza):
[operations/mediawiki-config@master] GrowthExperiments (beta-only): make GE use local search on beta enwiki

https://gerrit.wikimedia.org/r/547648

Wouldn't it make sense to request beta cswiki for the purpose of our experiments?

Wouldn't it make sense to request beta cswiki for the purpose of our experiments?

I have no idea how much effort it is to set up and maintain a beta wiki.
@Etonkovidova what do you think?

Wouldn't it make sense to request beta cswiki for the purpose of our experiments?

I have no idea how much effort it is to set up and maintain a beta wiki.

About the same as production wiki.

@Etonkovidova what do you think?

That does not sound reassuring :)

Wouldn't it make sense to request beta cswiki for the purpose of our experiments?

I have no idea how much effort it is to set up and maintain a beta wiki.

About the same as production wiki.

@Etonkovidova what do you think?

Since we have more than 40 lang wiki in beta cluster, I don't think it''d be too difficult to maintain. There are arwiki and kowiki, so it might be beneficial to have cswiki. Another benefit of cswiki in betalabs, it's to keep enwiki close to production and not tweak to accommodate new GrowthExperiments features.

Change 547648 merged by jenkins-bot:
[operations/mediawiki-config@master] GrowthExperiments (beta-only): make GE use local search on beta enwiki

https://gerrit.wikimedia.org/r/547648

Text extracts still don't seem to work, although the extracts seems to be working fine for the imported pages. Page images don't work because none of the articles seem to have images (other than infobox images which have not been transferred as the export was done without templates) - that's surprising but seems legit. Pageviews mostly don't work because those pull data from enwiki - we'd really have to set up beta cswiki to fix that.

It will be useful for testing instrumentation and local search, though.

Pageviews mostly don't work because those pull data from enwiki - we'd really have to set up beta cswiki to fix that.

Yes, it's a good argument for setting cswiki beta.

Beta cswiki has been created. I'm importing same pages as tgr into cswiki as initial content.

Mentioned in SAL (#wikimedia-releng) [2019-11-02T17:52:12Z] <Urbanecm> deployment-prep: mwscript importDump.php --wiki=cswiki /tmp/Wikipedie-20191031004003.xml (T236823)

Mentioned in SAL (#wikimedia-releng) [2019-11-02T17:52:56Z] <Urbanecm> deployment-prep: mwscript rebuildrecentchanges.php --wiki=cswiki (T236823)

Mentioned in SAL (#wikimedia-releng) [2019-11-02T17:55:04Z] <Urbanecm> deployment-prep: mwscript initSiteStats.php --wiki=cswiki --update (T236823)

Checked for cs.wikipedia.beta.wmflabs.org

  • articles are present (~1700 pages)
  • Homepage can be enabled with Suggested edits module

@Urbanecm Some additional config changes are needed?

The Suggested edits module and difficulty layers look as the following:

Screen Shot 2019-11-03 at 3.05.53 PM.png (547×391 px, 44 KB)

Screen Shot 2019-11-03 at 3.06.30 PM.png (477×785 px, 86 KB)

Change 548224 had a related patch set uploaded (by Kosta Harlan; owner: Kosta Harlan):
[operations/mediawiki-config@master] [beta] Working configuration for newcomer tasks on cswiki

https://gerrit.wikimedia.org/r/548224

Change 548224 abandoned by Kosta Harlan:
[beta] Working configuration for newcomer tasks on cswiki

Reason:
in favor of 547897

https://gerrit.wikimedia.org/r/548224

Checked for cs.wikipedia.beta.wmflabs.org

  • articles are present (~1700 pages)
  • Homepage can be enabled with Suggested edits module

@Urbanecm Some additional config changes are needed?

Yes, tracked with T237167: Configure beta cswiki for GrowthExperiments.

The Suggested edits module and difficulty layers look as the following:

Screen Shot 2019-11-03 at 3.05.53 PM.png (547×391 px, 44 KB)

Screen Shot 2019-11-03 at 3.06.30 PM.png (477×785 px, 86 KB)

Etonkovidova claimed this task.

Thanks, @Urbanecm - looks as all is Resolved.