dweiss opened a new pull request, #12881: URL: https://github.com/apache/lucene/pull/12881
This patch provides a number of small improvements aimed at improving performance of MatchHighlighter (and MatchRegionRetriever), especially in corner cases like: * queries that result in a large number of hits, especially in long fields. This causes hit passage scoring to be time-consuming, only to result in a few "best" passages. A configurable `maxHitsPerField` limit is added to allow capping the number of matches retrieved (and scored) to a reasonable number. * queries that require highlighting of hundreds of documents. The major performance bottleneck here is field value loading. MatchRegionRetriever now allows specifying which fields to load unconditionally, as well as filtering fields that contain hits (to skip computing highlights for fields which are used for filtering or are never displayed). Another improvement here is that highlights are now computed in parallel (using index searcher's task executor). There are a few minor API changes so I think it should be targeted for 10. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
