Once T294067 is deployed (probably in MediaWiki_1.39/wmf.23), we can reindex the relevant wikis, to activate the new analyzer!
Current counts are: Bengali (7 wikis)
Acceptance Criteria
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Open | None | T219550 [EPIC] Harmonize language analysis across languages | |||
| Resolved | Gehel | T272606 [EPIC] Unpack all Elasticsearch analyzers | |||
| Resolved | TJones | T294067 Install and unpack Bengali analyzer | |||
| Resolved | TJones | T315265 Reindex Bengali wikis to enable new analyzer |
Reindexing is done. Write up on Mediawiki.
Summary: Bengali Wikipedia had a very high zero-results rate (49.0%), and introducing stemming (and other changes—but mostly stemming) provided results for about ⅐ of zero-results queries, lowering the zero-results rate to 42.3%—which is still very high, but definitely better.