[PR] Fixed bit set vector [lucene]

2024-09-25 Thread via GitHub
risdenk opened a new pull request, #13827: URL: https://github.com/apache/lucene/pull/13827 ### Description -- 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 u

Re: [PR] Expose more parameters in MultiLeafKnnCollector ctor [lucene]

2024-09-25 Thread via GitHub
dungba88 commented on code in PR #13826: URL: https://github.com/apache/lucene/pull/13826#discussion_r1776165179 ## lucene/core/src/java/org/apache/lucene/search/knn/MultiLeafKnnCollector.java: ## @@ -58,7 +58,32 @@ public final class MultiLeafKnnCollector implements KnnCollect

Re: [I] Allow MultiLeafKnnCollector.greediness to be configurable [lucene]

2024-09-25 Thread via GitHub
dungba88 commented on issue #13699: URL: https://github.com/apache/lucene/issues/13699#issuecomment-2375535710 > TopKnnCollectorManager is so simple, it just doesn't seem worth it. I would just assume somebody provide their own manager which returns the appropriate collector with the option

Re: [PR] Add AbstractKnnVectorQuery.seed for seeded HNSW [lucene]

2024-09-25 Thread via GitHub
github-actions[bot] commented on PR #13635: URL: https://github.com/apache/lucene/pull/13635#issuecomment-2375493422 This PR has not had activity in the past 2 weeks, labeling it as stale. If the PR is waiting for review, notify the d...@lucene.apache.org list. Thank you for your contributi

Re: [PR] Update QueryUtils to use CollectorManager [lucene]

2024-09-25 Thread via GitHub
github-actions[bot] commented on PR #13748: URL: https://github.com/apache/lucene/pull/13748#issuecomment-2375493347 This PR has not had activity in the past 2 weeks, labeling it as stale. If the PR is waiting for review, notify the d...@lucene.apache.org list. Thank you for your contributi

Re: [PR] Add factory methods for IndexWriterConfig. [lucene]

2024-09-25 Thread via GitHub
github-actions[bot] commented on PR #13765: URL: https://github.com/apache/lucene/pull/13765#issuecomment-2375493322 This PR has not had activity in the past 2 weeks, labeling it as stale. If the PR is waiting for review, notify the d...@lucene.apache.org list. Thank you for your contributi

Re: [PR] Drop final modifier on the public DataInput.readGroupVInts method [lucene]

2024-09-25 Thread via GitHub
rmuir commented on PR #13825: URL: https://github.com/apache/lucene/pull/13825#issuecomment-2375369876 Isn't the root cause of the issue the `protected` method? The `final` method just invokes other methods, so alternatively we could fix the visibility of the protected method to be public l

Re: [I] Allow MultiLeafKnnCollector.greediness to be configurable [lucene]

2024-09-25 Thread via GitHub
benwtrent commented on issue #13699: URL: https://github.com/apache/lucene/issues/13699#issuecomment-2375131941 I opened a PR with that approach. As for > Allow it to be configured only in TopKnnCollectorManager, and users would need to override AbstractKnnVectorQuery.getKnnColle

[PR] Expose more parameters in MultiLeafKnnCollector ctor [lucene]

2024-09-25 Thread via GitHub
benwtrent opened a new pull request, #13826: URL: https://github.com/apache/lucene/pull/13826 There is currently no way to configure two parameters for the multi-leaf collector. For expert extensibility, this commit adds another ctor for advance usage: closes: https://github.com/apac

Re: [I] DataInput class can't be used with delegation pattern [lucene]

2024-09-25 Thread via GitHub
dweiss commented on issue #13820: URL: https://github.com/apache/lucene/issues/13820#issuecomment-2375077701 I linked the suggested minimal PR above. I am not convinced the test in TestFilterIndexInput.testOverrides is correct. The problem with delegation pattern ("FilterXyz" classes) is th

Re: [I] Allow MultiLeafKnnCollector.greediness to be configurable [lucene]

2024-09-25 Thread via GitHub
mikemccand commented on issue #13699: URL: https://github.com/apache/lucene/issues/13699#issuecomment-2375058087 > As for how to set these tunables, its possible now to create a sub-class of the knn queries and override `getKnnCollectorManager`. Maybe `MultiLeafKnnCollector` just needs to b

Re: [I] Allow MultiLeafKnnCollector.greediness to be configurable [lucene]

2024-09-25 Thread via GitHub
benwtrent commented on issue #13699: URL: https://github.com/apache/lucene/issues/13699#issuecomment-2374932998 > So the workaround for users now is to just ask for a larger top N and discard all but the top K in the end? Yes, that is the current solution. > +1 it'd be nice to

Re: [I] Allow MultiLeafKnnCollector.greediness to be configurable [lucene]

2024-09-25 Thread via GitHub
mikemccand commented on issue #13699: URL: https://github.com/apache/lucene/issues/13699#issuecomment-2374865202 > This is a similar conversation around folks adding `efSearch` as a parameter. Or possibly custom kNN collector behavior (your own early stopping logic...). Is `efSearch`

Re: [I] Add test backward compatibility indices for 8.11.4 [lucene]

2024-09-25 Thread via GitHub
ChrisHegarty closed issue #13821: Add test backward compatibility indices for 8.11.4 URL: https://github.com/apache/lucene/issues/13821 -- 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 c

Re: [PR] [9.12] Add support and compat testing for 8.11.4 [lucene]

2024-09-25 Thread via GitHub
ChrisHegarty commented on PR #13823: URL: https://github.com/apache/lucene/pull/13823#issuecomment-2374390742 > So I'm looking at the last time I did this for 8.11.3 ([521c26e](https://github.com/apache/lucene/commit/521c26e4fb547bb0aeafdbea838aec038ceb1538)), and it looks like there were m

Re: [PR] Add 8.11.4 back compat indices [lucene]

2024-09-25 Thread via GitHub
ChrisHegarty merged PR #13824: URL: https://github.com/apache/lucene/pull/13824 -- 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...@lucen

Re: [PR] [9.12] Add support and compat testing for 8.11.4 [lucene]

2024-09-25 Thread via GitHub
ChrisHegarty merged PR #13823: URL: https://github.com/apache/lucene/pull/13823 -- 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...@lucen

Re: [PR] [9.12] Add support and compat testing for 8.11.4 [lucene]

2024-09-25 Thread via GitHub
HoustonPutman commented on PR #13823: URL: https://github.com/apache/lucene/pull/13823#issuecomment-2374363412 So I'm looking at the last time I did this for 8.11.3 (https://github.com/apache/lucene/commit/521c26e4fb547bb0aeafdbea838aec038ceb1538), and it looks like there were more things a

[PR] [9.12] Add support and compat testing for 8.11.4 [lucene]

2024-09-25 Thread via GitHub
ChrisHegarty opened a new pull request, #13823: URL: https://github.com/apache/lucene/pull/13823 This commit adds support and compat testing for 8.11.4, to 9.12 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL a

Re: [I] Allow MultiLeafKnnCollector.greediness to be configurable [lucene]

2024-09-25 Thread via GitHub
benwtrent commented on issue #13699: URL: https://github.com/apache/lucene/issues/13699#issuecomment-2374032761 I have been thinking more and more of how we can pass "custom" or "complex" parameters via the API. FWIW, this greediness is honestly focused on graph based indices only. I

Re: [I] Allow MultiLeafKnnCollector.greediness to be configurable [lucene]

2024-09-25 Thread via GitHub
mikemccand commented on issue #13699: URL: https://github.com/apache/lucene/issues/13699#issuecomment-2373996504 [Disclaimer: I work with @dungba88 at Amazon Product Search team, and I also suggested he open this issue ;)] > Lucene has tens of parameters like this one, exposing them a

Re: [I] DataInput class can't be used with delegation pattern [lucene]

2024-09-25 Thread via GitHub
dweiss commented on issue #13820: URL: https://github.com/apache/lucene/issues/13820#issuecomment-2373976921 The test added as part of #11868 is checking whether only abstract methods are overridden in filter classes. It's a different case here. -- This is an automated message from the Ap

Re: [I] DataInput class can't be used with delegation pattern [lucene]

2024-09-25 Thread via GitHub
dweiss commented on issue #13820: URL: https://github.com/apache/lucene/issues/13820#issuecomment-2373937582 There is FilterIndexInput which extends IndexInput (extends DataInput). Interesting this hasn't been caught. -- This is an automated message from the Apache Git Service. To respond

Re: [I] DataInput class can't be used with delegation pattern [lucene]

2024-09-25 Thread via GitHub
mikemccand commented on issue #13820: URL: https://github.com/apache/lucene/issues/13820#issuecomment-2373932348 +1 for API symmetry. I thought we had reflection based unit tests to validate that our delegating `FilterXX` classes delegate all methods? Maybe we don't have a `FilterDa

Re: [PR] Bump the codec version to 10.0. [lucene]

2024-09-25 Thread via GitHub
mikemccand commented on PR #13815: URL: https://github.com/apache/lucene/pull/13815#issuecomment-2373926685 I was in a meeting yesterday and someone told me "oh we are using a very recent Lucene release, 9.1". And I asked them to double check ... and it was 9.10 they were using. This conf

Re: [I] TestTieredMergePolicy.testSimulateAppendOnly fails with AssertionError [lucene]

2024-09-25 Thread via GitHub
mikemccand commented on issue #13818: URL: https://github.com/apache/lucene/issues/13818#issuecomment-2373898281 Whether it's test only issue or bug in TMP, this is quite minor corner case -- it should not be a blocker for 9.12.0. -- This is an automated message from the Apache Git Servic

[PR] Add mergeProgress into MergeState for abort in mergeMiddle [lucene]

2024-09-25 Thread via GitHub
luyuncheng opened a new pull request, #13822: URL: https://github.com/apache/lucene/pull/13822 ### Description we see the scenarios like #13354, it would make abort waiting for merge finished, like #13354 and https://github.com/elastic/elasticsearch/issues/107513 elasticsearch #`rem

Re: [I] Add test backward compatibility indices for 8.11.4 [lucene]

2024-09-25 Thread via GitHub
ChrisHegarty commented on issue #13821: URL: https://github.com/apache/lucene/issues/13821#issuecomment-2373681027 I tried, but was unsuccessful in building the old _lucene-solr_ repo. :-( @HoustonPutman can you please run `dev-tools/scripts/addBackcompatIndexes.py 8.11.4`, and add t

[I] Add test backward compatibility indices for 8.11.4 [lucene]

2024-09-25 Thread via GitHub
ChrisHegarty opened a new issue, #13821: URL: https://github.com/apache/lucene/issues/13821 Nightly smoke tester failed last night: https://github.com/apache/lucene/actions/runs/11025229258/job/30619746011 Releases that don't seem to be tested: 8.11.4 Traceback (most recent

[I] DataInput class can't be used with delegation pattern [lucene]

2024-09-25 Thread via GitHub
dweiss opened a new issue, #13820: URL: https://github.com/apache/lucene/issues/13820 ### Description The change in #12841 brought this pair of methods to DataInput: ``` public final void readGroupVInts(long[] dst, int limit) throws IOException { ... } protected void readGrou