Discovery/Status updates/2019-07-15

This is the weekly update for the week starting 2019-07-15



  • Search of wikidata string property values using haswbstatement is case sensitive - after a bit of discussion, we added a patch for adding a case-insenstitive subfield for statement field [1], and re-indexing happened (with a minor issue) to get it into production [2]
  • Implemented support for haslabel:* [3]
  • We received a request to delete search indices for now-deleted zerowiki from production and did the needful [4]
  • Glent work: create oozie workflow for glent m0 prep has been finished [5]
  • The team did some work with the Multimedia team to index captions as description fields not label [6]
  • David worked on adapting IndexLookupFallbackMethod for glent requirements [7]
  • The team worked on evaluating DYM metrics that are available in current search satisfaction logging [8] with a follow up in [9]
    • Metrics we should use moving forward
      • % of search shown a [auto / non-auto] dym
        • Target: Increase % without significantly reducing the other metrics
      • % of people shown non-auto dym that click through to dym results
        • Target: Increase % of clickthrough
      • % of searches shown dym search results [auto / non-auto] dym results that clicked a result
        • Target: Increase % of clickthrough
  • There was a bug where there was an unexpected result set returned by Elasticsearch [10], we went ahead and closed it because it looks like the same root cause is fixed in [11]
  • CirrusSearch will now provide a query dispatcher (once the train finishes week of July 30) [12]
  • We did a spike: load search data into turnilo to test whether exploratory data can do away with some of the dashboards and decided that with respect to our 'did you mean' suggestions and this seems like a plausible path forward [13]
  • WikibaseLexemeCirrusSearch started to fail for no particular reason (seems to be due to missing class exported by WikibaseLexeme) and will be fixed in this week's train (July 30) [14]

Wikidata Query ServiceEdit

  • Finished reindexing database [15] which fixes resource problems with WDQS [16]
  • Wikidata RDF dumps no longer have BETA marker [17]
  • Fixed continuation support for MWAPI in WDQS [18]
  • Implemented support for wdtn: prefix for WDQS GUI [19]
  • Fixed incorrect VIAF URIs in WDQS data [20]
  • Implemented support for ChronologyProtection in WDQS Updater [21]
  • Refactored label service to support more complex queries [22]


  • The weekly update to the portal had been failing for a few weeks, Hashar helped out and fixed it, thanks! [23], but we still have another ticket to be fixed [24]