[GitHub] lucene-solr pull request #514: LUCENE-8591: add LegacyBM25Similarity#getDisc...
Github user javanna closed the pull request at: https://github.com/apache/lucene-solr/pull/514 --- - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[GitHub] lucene-solr issue #514: LUCENE-8591: add LegacyBM25Similarity#getDiscountOve...
Github user javanna commented on the issue: https://github.com/apache/lucene-solr/pull/514 This has been merged. --- - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[GitHub] lucene-solr pull request #514: LUCENE-8591: add LegacyBM25Similarity#getDisc...
GitHub user javanna opened a pull request: https://github.com/apache/lucene-solr/pull/514 LUCENE-8591: add LegacyBM25Similarity#getDiscountOverlaps You can merge this pull request into a Git repository by running: $ git pull https://github.com/javanna/lucene-solr enhancement/legacy_bm25_get_discount_overlaps Alternatively you can review and apply these changes as the patch at: https://github.com/apache/lucene-solr/pull/514.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #514 commit 05772af49b0980ed9661c4fb7e9cd1d4bb00603c Author: Luca Cavanna Date: 2018-12-05T15:08:13Z LUCENE-8591: add LegacyBM25Similarity#getDiscountOverlaps --- - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[GitHub] lucene-solr issue #511: LUCENE-8563: Remove k1+1 from the numerator of BM25S...
Github user javanna commented on the issue: https://github.com/apache/lucene-solr/pull/511 Merged. --- - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[GitHub] lucene-solr pull request #511: LUCENE-8563: Remove k1+1 from the numerator o...
Github user javanna closed the pull request at: https://github.com/apache/lucene-solr/pull/511 --- - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[GitHub] lucene-solr pull request #511: LUCENE-8563: Remove k1+1 from the numerator o...
Github user javanna commented on a diff in the pull request: https://github.com/apache/lucene-solr/pull/511#discussion_r237392155 --- Diff: lucene/MIGRATE.txt --- @@ -150,3 +150,11 @@ in order to support ToParent/ToChildBlockJoinQuery. Normalization is now type-safe, with CharFilterFactory#normalize() returning a Reader and TokenFilterFactory#normalize() returning a TokenFilter. + +## k1+1 constant factor removed from BM25 similarity numerator --- End diff -- Sure! --- - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[GitHub] lucene-solr pull request #511: LUCENE-8563: Remove k1+1 from the numerator o...
Github user javanna commented on a diff in the pull request: https://github.com/apache/lucene-solr/pull/511#discussion_r237392120 --- Diff: lucene/MIGRATE.txt --- @@ -150,3 +150,11 @@ in order to support ToParent/ToChildBlockJoinQuery. Normalization is now type-safe, with CharFilterFactory#normalize() returning a Reader and TokenFilterFactory#normalize() returning a TokenFilter. + +## k1+1 constant factor removed from BM25 similarity numerator + +Scores computed by the BM25 similarity are lower than previously as the k1+1 +constant factor was removed from the numerator of the scoring formula. +Ordering of results is preserved unless scores are computed from multiple +fields using different similarities. The previous behaviour is now exposed +through the LegacyBM25Similarity class. --- End diff -- yes I wasn't sure how to phrase that. Will add. --- - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[GitHub] lucene-solr pull request #511: Remove k1+1 from the numerator of BM25Similar...
GitHub user javanna opened a pull request: https://github.com/apache/lucene-solr/pull/511 Remove k1+1 from the numerator of BM25Similarity Patch for https://issues.apache.org/jira/browse/LUCENE-8563. This PR removes the k1+1 factor from the numerator of `BM25Similarity` and adds a new `LegacyBM25Similarity` under misc that exposes the old behaviour. Note that I haven't found a way to easily reproduce the previous behaviour in the explain method, so I left that part out of `LegacyBM25Similarity` for now. You can merge this pull request into a Git repository by running: $ git pull https://github.com/javanna/lucene-solr lucene-8563_bm25_k1_numerator Alternatively you can review and apply these changes as the patch at: https://github.com/apache/lucene-solr/pull/511.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #511 commit c665ed04b94f330fe27deec781cc1c59d45fddb5 Author: javanna Date: 2018-11-14T09:51:48Z Remove k1+1 constant factor from BM25 formula numerator commit 1b3714410771b53d30f29c78f27c37c617dea85c Author: javanna Date: 2018-11-14T11:15:39Z adapt TestFunctionQuery commit 7f56226906ededdc285877c429a00dae4f5a6d10 Author: javanna Date: 2018-11-14T14:08:38Z adapt TestPayloadScoreQParserPlugin commit 870d1d79afef7800ec348fabbe7b14b1a11ae1c0 Author: javanna Date: 2018-11-14T16:32:53Z add migrate note commit 5f0a0b0fb0ec7ebad8094f7fa97a09c877d3a1cb Author: javanna Date: 2018-11-27T21:15:12Z add LegacyBM25Similarity --- - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org