Before it goes live, we need to verify that wbqc can work with the new qlever endpoint while still supporting the blazegraph endpoint. This means reviewing the query templates and identifying which need work.

Templates: the templates in the sparqlhelper (~3)

A/C

• Update any queries which need to be updated, or collect a list of these
• The test in tests/phpunit/Helper/SparqlHelperTest.php are updated for QLever
• Hardcoded queries which are not valid for QLever are updated
• Templates which create queries are updated to work with QLever
• All tests pass with the QLever endpoints
• Create rollout plan (backwards compatible, feature flag?)

(!) Code should not go live. This will happen in T433710: [BLZMI] Set wbqc live on query-next.wikidata.org

Notes:

• Limitation: We’re lacking real data and so CV are difficult to verify.
  • From T433705: [BLZMI] SPIKE - How to have a local setup that uses wikidata.org data, current recommended approach is to run QLever locally. This will allow you to create an environment that can have CV
  • Then T431789: [BLZMI] Create script to generate items with constraint violations will create the CV
• useful tool for rewriting queries: https://to-qlever.toolforge.org/

Timebox: 8h