Summary

We're getting very close to having SPARQL endpoints for the WDQS main and scholarly subgraphs. We're working on documentation and communication.

We are drafting a plan on replacing our aging Elasticsearch 7.x, which besides providing a supported platform for search should also allow us to experiment with vector search.

What we've accomplished

WDQS graph splitting

Drafted a guide about federation at https://www.wikidata.org/wiki/Wikidata:SPARQL_query_service/WDQS_graph_split/Internal_Federation_Guide

Search backend replacement

A decision record for the target technology for the search backend replacement was drafted for team review. After this initial review it will be circulated further - https://www.mediawiki.org/wiki/Draft:ABaso_(WMF)/Wikimedia_Search_Platform/Decision_Records/Search_backend_replacement_technology

Misc

Investigate incident causing instabilities on the search clusters - https://phabricator.wikimedia.org/T370621
Additional edits to https://techblog.wikimedia.org/2024/07/18/how-many-languages-does-wikimedia-search-support/ to clarify that we do support all languages communities, even if the level of support for specific languages varies.
Error: Call to a member function audienceCan() on null (error related to optimization in Wikidata indexation) - https://phabricator.wikimedia.org/T370770