Closed treyhunner closed 9 years ago
Let's keep it simple first, do it on each save. Especially if we're saving each similarity separately, it will only be averaging across those records, one query for the calculation to take place, and a second to save. Two extra queries per similarity added or modified.
@macro1 that seems fine at first
I'm not sure if this should happen on every
UserSimilarity
save, as a queued task shortly after saves, or as a scheduled task.