In T344319, we followed a manual process to unpublish datasets which ended up triggering T344319#9109329 where datasets were removed from the published datasets repo but weren't delisted from the index. Automating this process with a script will help eliminate such mishaps.
Description
Description
Details
Details
Related Changes in Gerrit:
| Subject | Repo | Branch | Lines +/- | |
|---|---|---|---|---|
| Add unpublish script | research/mwaddlink | main | +75 -0 |
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | Trizek-WMF | T304110 [EPIC] Deploy "add a link" to all Wikipedias | |||
| Open | None | T309263 Support languages whose add-a-link models were not published | |||
| Resolved | kevinbazira | T344319 Remove models with poor evaluation metrics from the published datasets repo | |||
| Resolved | kevinbazira | T344799 Automate unpublishing of add-a-link datasets | |||
| Resolved | kevinbazira | T344832 Investigate why the add-a-link training pipeline concludes with missing datasets |
Event Timeline
Comment Actions
Change 951866 had a related patch set uploaded (by Kevin Bazira; author: Kevin Bazira):
[research/mwaddlink@main] Add unpublish script
Comment Actions
This task is currently blocked on T344832: Investigate why the add-a-link training pipeline concludes with missing datasets, which causes CI to fail on the patch here (or any other patch in research/mwaddlink).