Page MenuHomePhabricator

Create maintenance script for importing ORES topics
Closed, ResolvedPublic

Description

In production, all edits are scored by ORES and the scores are loaded into the search index, and used by the articletopic: search keyword. In the beta cluster and in developer setups, that doesn't happen, rendering features which rely on ORES topics useless. To ease testing of such features, we want a maintenance script which adds ORES topics to the search index directly, for a large number of articles (ideally in a meaningful way, but completely random topics are still better than nothing).

Since GrowthExperiments is currently the only extension using ORES topics, that seems like the logical place for the script, although it could be somewhere else (like CirrusSearch) if that's deemed useful.

Event Timeline

Change 667579 had a related patch set uploaded (by Gergő Tisza; owner: Gergő Tisza):
[mediawiki/extensions/GrowthExperiments@master] [DNM] Script for adding topics to a test environment

https://gerrit.wikimedia.org/r/667579

Tgr changed the task status from Stalled to Open.Mar 4 2021, 9:48 AM
Tgr moved this task from In Progress to Code Review on the Growth-Team (Sprint 0 (Growth Team)) board.

This is done; works locally, doesn't seem to work in beta. I'll look into that but UpdateWeightedTags.php doesn't seem to work there either so I don't think the reason is specific to this script.

Verified manually that the script puts the correct data in the index on beta. It is search itself that's misbehaving somehow.

Verified manually that the script puts the correct data in the index on beta. It is search itself that's misbehaving somehow.

@Tgr should we ask Discovery-Search team for help with this?

Change 667579 merged by jenkins-bot:
[mediawiki/extensions/GrowthExperiments@master] Script for adding topics to a test environment

https://gerrit.wikimedia.org/r/667579

Change 670175 had a related patch set uploaded (by Gergő Tisza; owner: Gergő Tisza):
[mediawiki/extensions/GrowthExperiments@master] Avoid re-importing the same page in importOresTopics.php

https://gerrit.wikimedia.org/r/670175

Change 670175 merged by jenkins-bot:
[mediawiki/extensions/GrowthExperiments@master] Improve page listing in importOresTopics.php

https://gerrit.wikimedia.org/r/670175

Change 672447 had a related patch set uploaded (by Gergő Tisza; owner: Gergő Tisza):
[mediawiki/extensions/GrowthExperiments@master] Fix topic scores in importOresTopics.php

https://gerrit.wikimedia.org/r/672447

Change 672447 merged by jenkins-bot:
[mediawiki/extensions/GrowthExperiments@master] Fix topic scores in importOresTopics.php

https://gerrit.wikimedia.org/r/672447