Re: [PR] Break the loop when segment is fully deleted by prior delTerms or delQueries [lucene]

2024-06-07 Thread via GitHub
github-actions[bot] commented on PR #13398: URL: https://github.com/apache/lucene/pull/13398#issuecomment-2155718807 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] Fix global score update bug in MultiLeafKnnCollector [lucene]

2024-06-07 Thread via GitHub
benwtrent commented on PR #13463: URL: https://github.com/apache/lucene/pull/13463#issuecomment-2155711517 @gsmiller My directories are: `/candidate` <- Lucene branch `/baseline` <- Lucene main `/util` <- lucene util Once you have the directories all set up:

Re: [PR] TaskExecutor should not fork unnecessarily [lucene]

2024-06-07 Thread via GitHub
original-brownbear commented on code in PR #13472: URL: https://github.com/apache/lucene/pull/13472#discussion_r1631773583 ## lucene/core/src/java/org/apache/lucene/search/TaskExecutor.java: ## @@ -112,15 +102,10 @@ RunnableFuture createTask(Callable callable) { () ->

Re: [PR] Fix global score update bug in MultiLeafKnnCollector [lucene]

2024-06-07 Thread via GitHub
gsmiller commented on code in PR #13463: URL: https://github.com/apache/lucene/pull/13463#discussion_r1631767073 ## lucene/core/src/java/org/apache/lucene/search/knn/MultiLeafKnnCollector.java: ## @@ -103,8 +105,11 @@ public boolean collect(int docId, float similarity) { if

Re: [PR] Fix global score update bug in MultiLeafKnnCollector [lucene]

2024-06-07 Thread via GitHub
gsmiller commented on PR #13463: URL: https://github.com/apache/lucene/pull/13463#issuecomment-2155669750 Ah @benwtrent good catch. Semi-sneaky that `updatesQueue` can have fewer than `k` results when the global update happens, but that makes sense. A couple of things: 1. I brought in yo

[PR] TaskExecutor should not fork unnecessarily [lucene]

2024-06-07 Thread via GitHub
original-brownbear opened a new pull request, #13472: URL: https://github.com/apache/lucene/pull/13472 When executing N tasks and waiting on the result of all of them, we should only fork N - 1 times and execute one task on the calling thread that is getting blocked anyway. This saves at le

Re: [PR] Fix global score update bug in MultiLeafKnnCollector [lucene]

2024-06-07 Thread via GitHub
benwtrent commented on PR #13463: URL: https://github.com/apache/lucene/pull/13463#issuecomment-2155529907 It would be good to have @mayya-sharipova's input here. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL

Re: [PR] Fix global score update bug in MultiLeafKnnCollector [lucene]

2024-06-07 Thread via GitHub
benwtrent commented on PR #13463: URL: https://github.com/apache/lucene/pull/13463#issuecomment-2155513437 https://github.com/mikemccand/luceneutil/pull/270 To fix lucene util + latest lucene changes. -- This is an automated message from the Apache Git Service. To respond to the me

Re: [PR] Fix global score update bug in MultiLeafKnnCollector [lucene]

2024-06-07 Thread via GitHub
benwtrent commented on PR #13463: URL: https://github.com/apache/lucene/pull/13463#issuecomment-2155506996 OK, maybe there is a bigger bug here, or this bug made this even better and we can reduce some constants to improve performance. I benchmarked over 1M 768 vectors, flushing every

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
dweiss commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2155495405 Feel free to push to that PR, Uwe. I think I'm done for the day. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use

Re: [PR] Silence odd test runner warnings after gradle upgrade [lucene]

2024-06-07 Thread via GitHub
dweiss commented on PR #13471: URL: https://github.com/apache/lucene/pull/13471#issuecomment-2155494641 Verify with: ``` gradlew :lucene:codecs:test --max-workers 1 --tests TestSimpleTextDocValuesFormat gradlew :lucene:core:test --max-workers 1 --tests TestCodecLoadingDeadlock ``

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
uschindler commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2155493207 Thanks. The other one which is special is distribution.tests project. This one starts a blank test runner without modules and initializes module system inside. I think we shoul

Re: [PR] Fix global score update bug in MultiLeafKnnCollector [lucene]

2024-06-07 Thread via GitHub
benwtrent commented on PR #13463: URL: https://github.com/apache/lucene/pull/13463#issuecomment-2155488471 @gsmiller I think your patch has a bug. I tried running with Lucene util to benchmark this to see if there is any perf change and got an exception. I am verifying my settings, but want

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
dweiss commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2155485845 I filed https://github.com/apache/lucene/pull/13471 to quiet down those two warnings. -- This is an automated message from the Apache Git Service. To respond to the message, please log

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
dweiss commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2155466471 It appears random because there are multiple runners and messages are proxied back to the console interleaved with everything else. I managed to debug this by running with ``--max-workers

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
uschindler commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2155425720 > These messages are sys-outed from each forked test runner and echoed as 'system.err' by the main gradle process. You can see them in the debug log if you run gradle with -debug:

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
dweiss commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2155396070 I also see these warnings, which you've mentioned: ``` 2024-06-07T14:17:29.1508657Z WARNING: Java vector incubator module is not readable. For optimal vector performance, pass '--add

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
dweiss commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2155392575 And looking at JDK sources, I don't think there is a way to dodge those warnings. ``` > java --add-modules jdk.incubator.vector -version WARNING: Using incubator modules: jdk.incu

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
dweiss commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2155387049 These messages are sys-outed from each forked test runner and echoed as 'system.err' by the main gradle process. You can see them in the debug log if you run gradle with -debug: ```

Re: [PR] WIP expose FlatVectorsFormat [lucene]

2024-06-07 Thread via GitHub
benwtrent commented on code in PR #13469: URL: https://github.com/apache/lucene/pull/13469#discussion_r1631602139 ## lucene/core/src/java/org/apache/lucene/codecs/lucene99/Lucene99FlatVectorsReader.java: ## @@ -217,6 +220,18 @@ public ByteVectorValues getByteVectorValues(String

[PR] Reciprocal Rank Fusion (RRF) in TopDocs [lucene]

2024-06-07 Thread via GitHub
harenlin opened a new pull request, #13470: URL: https://github.com/apache/lucene/pull/13470 ### Description Hello the community, Hank and I just follow the [discussion thread](https://lists.apache.org/thread/j2q4b8dfhoh1x15vo1myt357y5vtjl7z) to implement the RRF function that can

Re: [PR] Add new test case "testGetLines" for lucene/core/analysis/WordlistLoader [lucene]

2024-06-07 Thread via GitHub
hack4chang commented on code in PR #13419: URL: https://github.com/apache/lucene/pull/13419#discussion_r1631543661 ## lucene/core/src/test/org/apache/lucene/analysis/TestWordlistLoader.java: ## @@ -77,4 +82,17 @@ public void testSnowballListLoading() throws IOException { as

Re: [PR] Add new test case "testGetLines" for lucene/core/analysis/WordlistLoader [lucene]

2024-06-07 Thread via GitHub
hack4chang commented on code in PR #13419: URL: https://github.com/apache/lucene/pull/13419#discussion_r1631543661 ## lucene/core/src/test/org/apache/lucene/analysis/TestWordlistLoader.java: ## @@ -77,4 +82,17 @@ public void testSnowballListLoading() throws IOException { as

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
uschindler commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2155292753 > What's the command/ jdk that produces these warnings for you, Uwe? Java 21 with "Gradlew test". You see it also on Policeman Jenkins on every Lucene main build. I can r

Re: [PR] WIP expose FlatVectorsFormat [lucene]

2024-06-07 Thread via GitHub
msokolov commented on code in PR #13469: URL: https://github.com/apache/lucene/pull/13469#discussion_r1631529606 ## lucene/core/src/java/org/apache/lucene/codecs/lucene99/Lucene99FlatVectorsReader.java: ## @@ -217,6 +220,18 @@ public ByteVectorValues getByteVectorValues(String f

Re: [PR] Rewrite newSlowRangeQuery to MatchNoDocsQuery when upper > lower [lucene]

2024-06-07 Thread via GitHub
ioanatia commented on PR #13425: URL: https://github.com/apache/lucene/pull/13425#issuecomment-2155267634 thank you for the merge - I was out for a few days and did not had a chance to bump the version. -- This is an automated message from the Apache Git Service. To respond to the message

Re: [PR] Add new test case "testGetLines" for lucene/core/analysis/WordlistLoader [lucene]

2024-06-07 Thread via GitHub
stefanvodita commented on code in PR #13419: URL: https://github.com/apache/lucene/pull/13419#discussion_r1631490716 ## lucene/core/src/test/org/apache/lucene/analysis/TestWordlistLoader.java: ## @@ -77,4 +82,17 @@ public void testSnowballListLoading() throws IOException {

Re: [I] Expose flat vectors in "user space" [lucene]

2024-06-07 Thread via GitHub
msokolov commented on issue #13468: URL: https://github.com/apache/lucene/issues/13468#issuecomment-2155186004 I was thinking this could be used by PerFieldKnnVectorFormat since with this change a FlatVectorFormat is a KnnVectorFormat -- This is an automated message from the Apache Git Se

Re: [I] Expose flat vectors in "user space" [lucene]

2024-06-07 Thread via GitHub
jpountz commented on issue #13468: URL: https://github.com/apache/lucene/issues/13468#issuecomment-2155181368 Thanks, I had missed the quantization requirement and that you were ok with configuring a codec on the `IndexWriter`. -- This is an automated message from the Apache Git Service.

Re: [I] Expose flat vectors in "user space" [lucene]

2024-06-07 Thread via GitHub
msokolov commented on issue #13468: URL: https://github.com/apache/lucene/issues/13468#issuecomment-2155176509 https://github.com/apache/lucene/pull/13469 is just plumbing things through showing a possible way forward -- This is an automated message from the Apache Git Service. To respond

[PR] WIP expose FlatVectorsFormat [lucene]

2024-06-07 Thread via GitHub
msokolov opened a new pull request, #13469: URL: https://github.com/apache/lucene/pull/13469 just posting this here for discussion purposes; API is up for discussion -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use

Re: [I] Expose flat vectors in "user space" [lucene]

2024-06-07 Thread via GitHub
msokolov commented on issue #13468: URL: https://github.com/apache/lucene/issues/13468#issuecomment-2155145881 What I want to do is index float vectors, have them quantized and scored using the quantized form. -- This is an automated message from the Apache Git Service. To respond to the

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
dweiss commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2155054873 What's the command/ jdk that produces these warnings for you, Uwe? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use

[I] Expose flat vectors in "user space" [lucene]

2024-06-07 Thread via GitHub
msokolov opened a new issue, #13468: URL: https://github.com/apache/lucene/issues/13468 ### Description There are use cases where we want to store medium-dimensional vectors (ie embedding space vectors from ML models), retrieve them, compute distances among them, and perform KNN sear

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
dweiss commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2154978721 the api vs implementation difference is about how configurations are inherited and then exported. See [1]. I'm not sure what's happening with regard to the warning - let me know if you'd

Re: [PR] Document how to make tests run faster in IntelliJ [lucene]

2024-06-07 Thread via GitHub
dweiss commented on PR #13466: URL: https://github.com/apache/lucene/pull/13466#issuecomment-2154966075 If you have the gradle plugin visibile, you can access it from there too: ![image](https://github.com/apache/lucene/assets/199470/95ca3c5d-b21e-48c4-8985-ca3512785584) -- This is

Re: [PR] Document how to make tests run faster in IntelliJ [lucene]

2024-06-07 Thread via GitHub
gsmiller commented on code in PR #13466: URL: https://github.com/apache/lucene/pull/13466#discussion_r1631298811 ## CONTRIBUTING.md: ## @@ -58,7 +58,7 @@ In case your contribution fixes a bug, please create a new test case that fails ### IDE support -- *IntelliJ* - Intelli

Re: [I] Bug in MultiLeafKnnCollector causes #minCompetitiveSimilarity to stay artificially low in some situations [lucene]

2024-06-07 Thread via GitHub
gsmiller commented on issue #13462: URL: https://github.com/apache/lucene/issues/13462#issuecomment-215491 Proposed fix: GH#13463 -- 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

Re: [PR] Document how to make tests run faster in IntelliJ [lucene]

2024-06-07 Thread via GitHub
msokolov commented on code in PR #13466: URL: https://github.com/apache/lucene/pull/13466#discussion_r1631287846 ## CONTRIBUTING.md: ## @@ -58,7 +58,7 @@ In case your contribution fixes a bug, please create a new test case that fails ### IDE support -- *IntelliJ* - Intelli

Re: [PR] Document how to make tests run faster in IntelliJ [lucene]

2024-06-07 Thread via GitHub
msokolov commented on code in PR #13466: URL: https://github.com/apache/lucene/pull/13466#discussion_r1631279623 ## CONTRIBUTING.md: ## @@ -58,7 +58,7 @@ In case your contribution fixes a bug, please create a new test case that fails ### IDE support -- *IntelliJ* - Intelli

Re: [PR] Fix global score update bug in MultiLeafKnnCollector [lucene]

2024-06-07 Thread via GitHub
msokolov commented on code in PR #13463: URL: https://github.com/apache/lucene/pull/13463#discussion_r1631252496 ## lucene/core/src/java/org/apache/lucene/search/knn/MultiLeafKnnCollector.java: ## @@ -103,8 +105,11 @@ public boolean collect(int docId, float similarity) { if

Re: [PR] Document how to make tests run faster in IntelliJ [lucene]

2024-06-07 Thread via GitHub
gsmiller commented on code in PR #13466: URL: https://github.com/apache/lucene/pull/13466#discussion_r1631243472 ## CONTRIBUTING.md: ## @@ -58,7 +58,7 @@ In case your contribution fixes a bug, please create a new test case that fails ### IDE support -- *IntelliJ* - Intelli

[PR] Honor read advice on compound files. [lucene]

2024-06-07 Thread via GitHub
jpountz opened a new pull request, #13467: URL: https://github.com/apache/lucene/pull/13467 This includes the following changes: - New `IndexInput#slice(String, long, long, ReadAdvice)` API that allows creating slices with different advices. - New `CompoundFormat` that aligns start o

Re: [PR] Add int8_hnsw backcompat index creation to dev tools scripts [lucene]

2024-06-07 Thread via GitHub
benwtrent merged PR #13465: URL: https://github.com/apache/lucene/pull/13465 -- 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.a

Re: [PR] Document how to make tests run faster in IntelliJ [lucene]

2024-06-07 Thread via GitHub
msokolov merged PR #13466: URL: https://github.com/apache/lucene/pull/13466 -- 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.ap

[PR] Document how to make tests run faster in IntelliJ [lucene]

2024-06-07 Thread via GitHub
msokolov opened a new pull request, #13466: URL: https://github.com/apache/lucene/pull/13466 Add some info about setting up IntelliJ to use its internal test runner. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the

Re: [PR] WIP - Add minimum number of segments to TieredMergePolicy [lucene]

2024-06-07 Thread via GitHub
carlosdelest commented on code in PR #13430: URL: https://github.com/apache/lucene/pull/13430#discussion_r1631094664 ## lucene/core/src/java/org/apache/lucene/index/TieredMergePolicy.java: ## @@ -522,21 +550,28 @@ private MergeSpecification doFindMerges( final List cand

Re: [PR] Reduce the heap use of BKDReader instances [lucene]

2024-06-07 Thread via GitHub
jpountz merged PR #13464: URL: https://github.com/apache/lucene/pull/13464 -- 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.apa

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
jpountz commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2154632683 FWIW moving my `~/.gradle` directory to a filesystem that is not encrypted worked around the problem. -- This is an automated message from the Apache Git Service. To respond to the mes

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
uschindler commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2154617527 Most error messages regarding the modules are also seen on Gradle 8.6, so its not a new issue. Sorry for false alarm. I will still investigate this. It seems to have indeed some

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
uschindler commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2154584633 I think I know the problem. It looks like Gradle test runner loads those classes for some checks outside of tests. This is why the message appear randomly in the output. In

Re: [PR] Sparse index: optional skip list on top of doc values [lucene]

2024-06-07 Thread via GitHub
iverase commented on code in PR #13449: URL: https://github.com/apache/lucene/pull/13449#discussion_r1631012311 ## lucene/core/src/java/org/apache/lucene/index/FieldInfo.java: ## @@ -289,6 +302,24 @@ static void verifySameDocValuesType( } } + /** + * Verify that the

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
uschindler commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2154578383 When checking jenkisn builds, the message... ``` WARNING: Unknown module: org.apache.lucene.core specified to --enable-native-access ``` ...appears on random gradle

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
uschindler commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2154569035 This seems to be a message out of context. When running expressions tests it has the vector incubator enabled. The message comes from somewhere about the not readable module com

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
uschindler commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2154553906 If the "moduleApi" dependency is wanted then it makes sense to me. But then we should only add the jvmArg if the module is linked at all. Not sure how to test this. So basicall

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
uschindler commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2154550898 Not sure what's the problem with both projects; in the test configuration all looks fine. Maybe we have not seen those messages before as Gradle did not print them. New version seems

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
uschindler commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2154532698 i still see some problems during tests regarding vector module. I have to check what's going on. Possibly some test system properties issue. It looks like the vector module is not alw

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
uschindler commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2154523213 Hi, on my windows pc (which also has limitations on length of path names), I have seen not issues. On my new laptop I havent perl installed at moment, so building docs failed, but

Re: [PR] Update Gradle wrapper to 8.8 - supports Java 22 [lucene]

2024-06-07 Thread via GitHub
jpountz commented on PR #13453: URL: https://github.com/apache/lucene/pull/13453#issuecomment-2154447314 I'm getting errors since upgrading to Gradle 8.8 that I can't figure out how to fix. When running with `--stacktrace`, I see the following error that seems to be the root cause. I was no