Kaival Parikh

Refactor KnnGraphTester

1

`KnnGraphTester` has some drawbacks and needs a refactor because: - Can only test HNSW search time on graphs created by itself (cannot work easily with existing / custom indexes) -...

Add timeout support to AbstractVectorSimilarityQuery

9

### Description Along similar lines of #13202, adding timeout support for [`AbstractVectorSimilarityQuery`](https://github.com/apache/lucene/blob/main/lucene/core/src/java/org/apache/lucene/search/AbstractVectorSimilarityQuery.java) which performs similarity-based vector searches While the graph search happens inside `#scorer`, it may go over the configured...

[DISCUSS] Could we have a different ANN algorithm for Learned Sparse Vectors?

4

### Description Learned Sparse Vectors claim to combine the benefits of sparse (i.e. lexical) and dense (i.e. vector) representations From https://en.wikipedia.org/wiki/Learned_sparse_retrieval: > Learned sparse retrieval or sparse neural search is...

Enable Faiss-based vector format to index larger number of vectors in a single segment

6

### Description I was trying to index a large number of vectors in a single segment, and ran into an error because of the way we [copy vectors to native...

### Description Backport #14178 to 10.x ### Summary of changes - Support for JDK21 via MR-JAR and runtime lookup (similar to other Panama-based classes [here](https://github.com/apache/lucene/tree/f93a0ed318b528229dee91dbbca98cb6607b24ae/lucene/core/src/java21/org/apache/lucene/internal/vectorization)). This was tricky because some...

Support multiple HNSW graphs backed by the same vectors

26

### Description For use-cases of searching different subsets of vectors in the index, where a non-trivial portion of vectors across fields are overlapping. This could be done today by: 1....

Spinoff from #468 -- where we added an option to `KnnGraphTester` to perform "index-time KNN filtering", which simply means creating a new vector field for filters known at index time,...

Align vectors on disk for optimal performance on ARM CPUs

17

### Description Today, float vectors are [aligned to 4 bytes](https://github.com/apache/lucene/blob/8e8e37d9e94c290cf8d02e9f318e601baedf28bc/lucene/core/src/java/org/apache/lucene/codecs/lucene99/Lucene99FlatVectorsWriter.java#L205) in a Lucene index, but with Panama -- we can work with (upto) 512 bits (== 64 bytes, or 16...

Separate Panama and Vector classes

12

Addresses #15284

De-dup raw vectors?

2

### Description Closes #14758 Demonstrating the proposal to de-duplicate raw vectors in Lucene! **_Note: Right now this is very crude, and only for demonstration purposes._**