Tobi_WMDE_SW added a comment.

Outcome of the meeting with @aude @janzerebecki @thiemowmde @daniel

(1) uniqueness constraints:
-> PROBLEM: currently done by a self-JOIN on the huge table -> bad idea
-> solution would be a separate table with hashes
-> AGREEMENT moving uniqueness constraints away from terms table
-> AGREEMENT the linked suggestion (hash based) on how to do this should be 
discussed again before starting to implement it (and needs to be discussed with 
WMF (performance))

(2) finding props by label:
-> PROBLEM: timeout for short prefixes (caused by the fact that we sort the 
huge result set by weight in memory)
-> AGREEMENT: use elastic but keep current structure of terms table to be able 
to fall back
-> AGREEMENT: there should be a fallback for users that don’t want to depend on 
elastic
-> AGREEMENT: split off interfaces from term index (for each type of query, and 
one for updating)
-> AGREEMENT: once we have the elastic solution, we can think of moving away 
from the fall back (or find a different solution as fallback)

(3) finding labels by id:
-> AGREEMENT: keep it as it is for now until we have solved problem 1 and 2 and 
then re-visit to discuss this again

general AGREEMENT: out of those 3 problem 2 has highest priority.


TASK DETAIL
  https://phabricator.wikimedia.org/T86530

REPLY HANDLER ACTIONS
  Reply to comment or attach files, or !close, !claim, !unsubscribe or !assign 
<username>.

EMAIL PREFERENCES
  https://phabricator.wikimedia.org/settings/panel/emailpreferences/

To: Tobi_WMDE_SW
Cc: thiemowmde, Tobi_WMDE_SW, JeroenDeDauw, JanZerebecki, aude, Aklapper, 
Liuxinyu970226, daniel, hoo, Wikidata-bugs, GWicke



_______________________________________________
Wikidata-bugs mailing list
Wikidata-bugs@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs

Reply via email to