Currently we get 200 items from ApiQueryGrowthTasks. But we only have text extracts to show with the first 20 cards. So, we need to implement code which will load batches of 20 text extracts as the user navigates through the stack of suggested edits.
This task could optionally involve completing T236522 but it doesn't necessarily need to.
| Subject | Author | Repo | Branch | Lines +/- | |
|---|---|---|---|---|---|
| Suggested Edits: Use RESTBase for loading text extracts | Kosta Harlan | mediawiki/extensions/GrowthExperiments | master | +44 -16 |
Not sure if we should try to do that for 1.0, but eventually, life will become much simpler if we replace random sort with pseudorandom (based on user id, for example). ElasticSearch supports that, exposing it in CirrusSearch should be trivial. After that, we can just use API continuation (or batches of 20 in the first place).
Change 548278 had a related patch set uploaded (by Kosta Harlan; owner: Kosta Harlan):
[mediawiki/extensions/GrowthExperiments@master] (wip) Suggested Edits: Use RESTBase for loading text extracts
Change 548278 merged by jenkins-bot:
[mediawiki/extensions/GrowthExperiments@master] Suggested Edits: Use RESTBase for loading text extracts