As a user I don't want to get an image suggestion that has already been rejected by someone else, so we need to record rejections and remove them from our suggested-image dataset.
When something is rejected, the Growth team will emit an event. This will be stored in the event database in Hive in a new table
The code for gathering suggestions will need to be modified so as to use the contents of the new Hive table to filter out rejected suggestions
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | CBogen | T299781 [EPIC] Image suggestions backend | |||
| Resolved | mfossati | T296814 [EPIC] Article-level image suggestions data pipeline | |||
| Resolved | Cparle | T299890 [M] Exclude previously rejected image suggestions when generating new suggestions |
Once https://gerrit.wikimedia.org/r/c/schemas/event/secondary/+/779052 is deployed, we can gather rejected suggestions through a query similar to this one (pasted from @lbowmaker's advice):
SELECT wiki, page_id, image_filename FROM event_sanitized.image_suggestions_feedback WHERE is_rejected = True
See also T302925#7856220.
Patch written https://gitlab.wikimedia.org/repos/generated-data-platform/datapipelines/-/tree/T299890-exclude-rejections but can't fully test it until the schema is deployed (see https://gitlab.wikimedia.org/repos/generated-data-platform/topics/image-suggestions-feedback/-/merge_requests/1)
Merge request at https://gitlab.wikimedia.org/repos/generated-data-platform/datapipelines/-/merge_requests/67 merged! Closing.