8000 chore: update changelog for 4.0.0-beta.9 by github-actions[bot] · Pull Request #2 · tlwillke/jvector · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

chore: update changelog for 4.0.0-beta.9 #2

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

chore: update changelog for 4.0.0-beta.9 #2

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
192 changes: 189 additions & 3 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,192 @@
# Changelog
### Changelog

## [4.0.0-beta.5](https://github.com/datastax/jvector/tree/4.0.0-beta.5) (2025-05-23)
All notable changes to this project will be documented in this file. Dates are displayed in UTC.

[Full Changelog](https://github.com/datastax/jvector/compare/4.0.0-beta.4...4.0.0-beta.5)
Generated by [`auto-changelog`](https://github.com/CookPete/auto-changelog).

#### 4.0.0-beta.9

- Fix issue when calling cleanup while concurrently executing searches [`#483`](https://github.com/tlwillke/jvector/pull/483)
- Improve the efficiency and memory usage of GraphIndexBuilder.cleanup [`#477`](https://github.com/tlwillke/jvector/pull/477)
- add PQ training benchmark [`#482`](https://github.com/tlwillke/jvector/pull/482)
- Remove extraneous character from datasets.yml [`#484`](https://github.com/tlwillke/jvector/pull/484)
- Upgrade YAML files to v5 after the format was introduced in the last update [`#478`](https://github.com/tlwillke/jvector/pull/478)
- Add a new graph node using a search score [`#473`](https://github.com/tlwillke/jvector/pull/473)
- chore(release): Bump tag version and update changelog [`#471`](https://github.com/tlwillke/jvector/pull/471)
- Bench improvements with YAML config files [`#462`](https://github.com/tlwillke/jvector/pull/462)
- avx512 test runners, jobs, and assertions [`#469`](https://github.com/tlwillke/jvector/pull/469)
- Find insertion point before performing copy in ConcurrentNeighborMap [`#468`](https://github.com/tlwillke/jvector/pull/468)
- Factorize the computation of diverse edges [`#466`](https://github.com/tlwillke/jvector/pull/466)
- Random vector index build jmh + setup scripts [`#424`](https://github.com/tlwillke/jvector/pull/424)
- Better control over setting string formats for the benchmark metrics [`#461`](https://github.com/tlwillke/jvector/pull/461)
- Check values of clusterCount in PQ [`#464`](https://github.com/tlwillke/jvector/pull/464)
- Perf metrics improvement v1.1 [`#460`](https://github.com/tlwillke/jvector/pull/460)
- Simplify OnDiskGraphIndex.View to avoid code duplication [`#458`](https://github.com/tlwillke/jvector/pull/458)
- Improve perf metrics measurement and summarization [`#459`](https://github.com/tlwillke/jvector/pull/459)
- Creating starting point for changelog tracking. [`#456`](https://github.com/tlwillke/jvector/pull/456)
- chore(release) adjust changelog generation steps [`#437`](https://github.com/tlwillke/jvector/pull/437)
- Workflow/tag release update [`#452`](https://github.com/tlwillke/jvector/pull/452)
- Workflow/tag release update [`#450`](https://github.com/tlwillke/jvector/pull/450)
- Workflow/tag release update [`#448`](https://github.com/tlwillke/jvector/pull/448)
- Workflow/tag release update [`#446`](https://github.com/tlwillke/jvector/pull/446)
- Workflow/tag release update [`#444`](https://github.com/tlwillke/jvector/pull/444)
- Workflow/tag release update [`#442`](https://github.com/tlwillke/jvector/pull/442)
- Workflow/tag release update [`#440`](https://github.com/tlwillke/jvector/pull/440)
- Fixes for GHA tag-release workflow [`#438`](https://github.com/tlwillke/jvector/pull/438)
- Fix minor bug in getNodes. [`#434`](https://github.com/tlwillke/jvector/pull/434)
- Fix/refactor NodeScoreIterator, BoundedLongHeap, and GrowableLongHeap bulk addition implementations [`#433`](https://github.com/tlwillke/jvector/pull/433)
- Remove extra prefix of v from tag version [`#432`](https://github.com/tlwillke/jvector/pull/432)
- Update only the root level pom.xml as part of the GHA workflow [`#431`](https://github.com/tlwillke/jvector/pull/431)
- New GHA workflow to create new tag and update changelog [`#428`](https://github.com/tlwillke/jvector/pull/428)
- bugfix for 429 - eliminate maven-resources-plugin warning [`#430`](https://github.com/tlwillke/jvector/pull/430)
- Update test resume [`#422`](https://github.com/tlwillke/jvector/pull/422)
- Fix calls to deprecated GraphIndex.size() [`#426`](https://github.com/tlwillke/jvector/pull/426)
- Fix NPE in GraphIndexBuilder.load [`#425`](https://github.com/tlwillke/jvector/pull/425)
- Reduce the number of vector allocations in BuildScoreProvider.pqBuilderScoreProvider [`#419`](https://github.com/tlwillke/jvector/pull/419)
- Improve the computation of accuracy [`#408`](https://github.com/tlwillke/jvector/pull/408)
- Merge latest commits from hnsw-3 [`#423`](https://github.com/tlwillke/jvector/pull/423)
- Fix native implementations of PQ assembleAndSum and pqDecodedCosineSimilarity [`#420`](https://github.com/tlwillke/jvector/pull/420)
- Fix FusedADC.writeInline [`#417`](https://github.com/tlwillke/jvector/pull/417)
- Implement NodeQueue#pushAll and AbstractLongHeap#addAll [`#415`](https://github.com/tlwillke/jvector/pull/415)
- Count expanded nodes [`#406`](https://github.com/tlwillke/jvector/pull/406)
- Search pruning & fix the reported number of visited nodes [`#405`](https://github.com/tlwillke/jvector/pull/405)
- Fix flaky tests and eliminate console output [`#404`](https://github.com/tlwillke/jvector/pull/404)
- Remove query-time usage of ByteSequence::slice to reduce object allocations [`#403`](https://github.com/tlwillke/jvector/pull/403)
- add index construction benchmark [`#398`](https://github.com/tlwillke/jvector/pull/398)
- Add jmh benchmarks [`#396`](https://github.com/tlwillke/jvector/pull/396)
- Fix MutableBQVectors parameterization. Add basic test coverage. [`#395`](https://github.com/tlwillke/jvector/pull/395)
- make examples use index view [`#392`](https://github.com/tlwillke/jvector/pull/392)
- Update Test2DThreshold to control for averages instead of worst-case statistics [`#391`](https://github.com/tlwillke/jvector/pull/391)
- Change variable names to improve readability [`#388`](https://github.com/tlwillke/jvector/pull/388)
- Fix NVQ distance computations in Native provider [`#389`](https://github.com/tlwillke/jvector/pull/389)
- Improved use ScoreTracker to avoid wasteful searching for very large k [`#387`](https://github.com/tlwillke/jvector/pull/387)
- Use ScoreTracker to avoid wasteful searching for very large k [`#384`](https://github.com/tlwillke/jvector/pull/384)
- Fix CI on Windows due to missing posix_madvise support [`#383`](https://github.com/tlwillke/jvector/pull/383)
- add MADV_RANDOM [`#382`](https://github.com/tlwillke/jvector/pull/382)
- Make ravv usage thread-safe [`#381`](https://github.com/tlwillke/jvector/pull/381)
- Non-uniform vector quantization [`#374`](https://github.com/tlwillke/jvector/pull/374)
- Hand-unroll the SIMD dot product loop [`#380`](https://github.com/tlwillke/jvector/pull/380)
- Fix regression in assembleAndSum PQ decoder performance [`#379`](https://github.com/tlwillke/jvector/pull/379)
- Store compressed vectors in dense ByteSequence for PQVectors [`#370`](https://github.com/tlwillke/jvector/pull/370)
- Reenable SimdOps.assembleAndSum; implement Panama/Native equivalent for CosineDecoder acceleration [`#368`](https://github.com/tlwillke/jvector/pull/368)
- Use fma in SimdOps.cosineSimilarity sum vector [`#363`](https://github.com/tlwillke/jvector/pull/363)
- Improve performance of reconnectOrphanedNodes [`#359`](https://github.com/tlwillke/jvector/pull/359)
- Use float in cosine metric final calculation in default vectorization provider [`#358`](https://github.com/tlwillke/jvector/pull/358)
- approximateMediod returns a random node when the graph is too disconnected to search for the centroid [`#356`](https://github.com/tlwillke/jvector/pull/356)
- Remove check for VBMI on CPU. With Fused ADC using shorts rather than bytes, we no longer need vpermi2b. [`#352`](https://github.com/tlwillke/jvector/pull/352)
- Set IdentityMapper maxOrdinal correctly in Grid/SiftSmall. [`#351`](https://github.com/tlwillke/jvector/pull/351)
- add support for non-sequential remapped ordinals [`#349`](https://github.com/tlwillke/jvector/pull/349)
- cache reranked scores [`#341`](https://github.com/tlwillke/jvector/pull/341)
- extract RandomAccessWriter interface from BRAW [`#340`](https://github.com/tlwillke/jvector/pull/340)
- Clear scratch structures if search terminates exceptionally [`#337`](https://github.com/tlwillke/jvector/pull/337)
- Reduce tendency of reconnectOrphanedNodes to leave orphaned nodes [`#335`](https://github.com/tlwillke/jvector/pull/335)
- Ecapsulate NodeArray internals [`#328`](https://github.com/tlwillke/jvector/pull/328)
- Remove on-disk reranking [`#327`](https://github.com/tlwillke/jvector/pull/327)
- standardize ReaderSupplier implementations as inner classes of their respective RandomAccessReaders, and add a Supplier for SimpleReader [`#323`](https://github.com/tlwillke/jvector/pull/323)
- Fix InlineVectorValues.size/LvqVectorValues.size [`#320`](https://github.com/tlwillke/jvector/pull/320)
- Fix writing jvector2-compatible indexes incrementally [`#313`](https://github.com/tlwillke/jvector/pull/313)
- OrdinalMapper [`#299`](https://github.com/tlwillke/jvector/pull/299)
- Add a memory-mapped RandomAccessReader using MemorySegment api [`#296`](https://github.com/tlwillke/jvector/pull/296)
- tweaks to make life easier for upgraders [`#290`](https://github.com/tlwillke/jvector/pull/290)
- remove (broken) concurrency support from removeDeletedNodes [`#273`](https://github.com/tlwillke/jvector/pull/273)
- Reduce allocation by pooling GraphSearcher objects [`#270`](https://github.com/tlwillke/jvector/pull/270)
- track visited nodes using IntHashSet instead of different BitSets [`#269`](https://github.com/tlwillke/jvector/pull/269)
- make Test2DThreshold less fragile [`#266`](https://github.com/tlwillke/jvector/pull/266)
- Generify new abstracted ODGI, such that the caching hierarchy doesn't require casting back to view/cachednode types. Fix any leakage of return types outside of visibility scope. [`#262`](https://github.com/tlwillke/jvector/pull/262)
- Support building indexes using compressed vectors [`#244`](https://github.com/tlwillke/jvector/pull/244)
- Enable JDK 22 CI [`#258`](https://github.com/tlwillke/jvector/pull/258)
- Use on-heap MemorySegments for native vectors/sequences [`#246`](https://github.com/tlwillke/jvector/pull/246)
- Merge 3.0-alpha [`#256`](https://github.com/tlwillke/jvector/pull/256)
- Replace NormalDistributionTracker with TwoPhaseTracker [`#247`](https://github.com/tlwillke/jvector/pull/247)
- Replace NormalDistributionTracker with TwoPhaseTracker [`#247`](https://github.com/tlwillke/jvector/pull/247)
- Anisotropic PQ [`#201`](https://github.com/tlwillke/jvector/pull/201)
- add a diverseBefore marker to avoid recomputing diversity that hasn't changed [`#242`](https://github.com/tlwillke/jvector/pull/242)
- Cleanup for 3.0 release [`#182`](https://github.com/tlwillke/jvector/pull/182)
- Release 2.0.5 [`#171`](https://github.com/tlwillke/jvector/pull/171)
- Remove ThreadPooling self reference in Pooled object to prevent memory leak [`#169`](https://github.com/tlwillke/jvector/pull/169)
- PhysicalCoreExecutor can exit gracefully; CachingGraphIndex's cacheDistance can be customized [`#160`](https://github.com/tlwillke/jvector/pull/160)
- Add optional FJP args for indexing and quantization [`#162`](https://github.com/tlwillke/jvector/pull/162)
- Fix some bugs [`#156`](https://github.com/tlwillke/jvector/pull/156)
- Use fma in SIMD Euclidean/cosine [`#153`](https://github.com/tlwillke/jvector/pull/153)
- Fix usage of null acceptOrds in SiftSmall example [`#152`](https://github.com/tlwillke/jvector/pull/152)
- vectorsEncountered not always in sync with resultQueue, causing NPE when breaking out of loop due to threshold probability [`#150`](https://github.com/tlwillke/jvector/pull/150)
- Run verify phase in CI (which includes license checks) [`#149`](https://github.com/tlwillke/jvector/pull/149)
- add upgrade guide [`#147`](https://github.com/tlwillke/jvector/pull/147)
- Split LongHeap into Growable and Bounded flavors [`#146`](https://github.com/tlwillke/jvector/pull/146)
- Release 2.0.1 [`#144`](https://github.com/tlwillke/jvector/pull/144)
- add getOriginalSize and getCompressedSize to CompressedVectors interface [`#143`](https://github.com/tlwillke/jvector/pull/143)
- Cherry-pick various bench improvements from PR #76. [`#133`](https://github.com/tlwillke/jvector/pull/133)
- Fix running single test using Maven commandline [`#141`](https://github.com/tlwillke/jvector/pull/141)
- Reconnect orphaned nodes in cleanup() [`#138`](https://github.com/tlwillke/jvector/pull/138)
- Add binary quantization [`#135`](https://github.com/tlwillke/jvector/pull/135)
- Updated downloadhelper [`#134`](https://github.com/tlwillke/jvector/pull/134)
- downloads wikipedia fvec files for 100k, switched to squad based query vectors [`#130`](https://github.com/tlwillke/jvector/pull/130)
- Addresses issue #36 by adding license header checks. Added headers on… [`#129`](https://github.com/tlwillke/jvector/pull/129)
- Ipcexample [`#127`](https://github.com/tlwillke/jvector/pull/127)
- Deletes [`#117`](https://github.com/tlwillke/jvector/pull/117)
- CI improvement [`#131`](https://github.com/tlwillke/jvector/pull/131)
- Adds DenseIntMap for building graph with much less contention. back to zero dependency [`#128`](https://github.com/tlwillke/jvector/pull/128)
- fix mergeNeighbors to not add duplicate nodes, and fix test to check for duplicates [`#119`](https://github.com/tlwillke/jvector/pull/119)
- Mt index build fixes [`#113`](https://github.com/tlwillke/jvector/pull/113)
- Fork test VM per core [`#111`](https://github.com/tlwillke/jvector/pull/111)
- Add improved test coverage for on-disk graph caching [`#109`](https://github.com/tlwillke/jvector/pull/109)
- Fix SimpleMappedReader to respect offset [`#106`](https://github.com/tlwillke/jvector/pull/106)
- KMeansPlusPlusClusterer optimizations [`#100`](https://github.com/tlwillke/jvector/pull/100)
- Add simd approach for summing the cached PQ products of each encoded vector [`#104`](https://github.com/tlwillke/jvector/pull/104)
- README tweaks [`#101`](https://github.com/tlwillke/jvector/pull/101)
- Remove triangle inequality from k means plus plus [`#98`](https://github.com/tlwillke/jvector/pull/98)
- Clean up all build warnings related to multimodule versioning. [`#97`](https://github.com/tlwillke/jvector/pull/97)
- Fragment cache [`#94`](https://github.com/tlwillke/jvector/pull/94)
- wikipedia datasets in readme [`#95`](https://github.com/tlwillke/jvector/pull/95)
- Add decodedCosine fast path [`#91`](https://github.com/tlwillke/jvector/pull/91)
- Implement optimized decoded square distance [`#89`](https://github.com/tlwillke/jvector/pull/89)
- Fix code coverage in IntelliJ [`#88`](https://github.com/tlwillke/jvector/pull/88)
- Fix recall regression for centered PQ with non-dot product metrics [`#84`](https://github.com/tlwillke/jvector/pull/84)
- Refactor tests into jvector-tests module. Set up configurations to be able to run tests with JDK11 features and JDK20 features. [`#75`](https://github.com/tlwillke/jvector/pull/75)
- move sharing from annotation to method; use that in PQ [`#83`](https://github.com/tlwillke/jvector/pull/83)
- Minor documentation updates [`#77`](https://github.com/tlwillke/jvector/pull/77)
- Change package from com.github.jbellis to io.github.jbellis [`#71`](https://github.com/tlwillke/jvector/pull/71)
- Release 0.9.0 [`#69`](https://github.com/tlwillke/jvector/pull/69)
- Build rework for release. Split examples out into separate project. [`#68`](https://github.com/tlwillke/jvector/pull/68)
- Build updates to prep for release [`#66`](https://github.com/tlwillke/jvector/pull/66)
- Javadoc fixes and produce javadoc as part of the build [`#64`](https://github.com/tlwillke/jvector/pull/64)
- Revert changes to hdf5 path in Bench, set up mvn exec to use parent basedir [`#65`](https://github.com/tlwillke/jvector/pull/65)
- Migrate to Maven multi-module project and produce multirelease jar [`#61`](https://github.com/tlwillke/jvector/pull/61)
- Swap Maven jdk11 build to use -release 11, which also checks for too-new APIs [`#59`](https://github.com/tlwillke/jvector/pull/59)
- Remove MacOS runner [`#58`](https://github.com/tlwillke/jvector/pull/58)
- Fix Windows tests [`#57`](https://github.com/tlwillke/jvector/pull/57)
- Switch to maven profiles, add gh workflow for tests and add package f… [`#55`](https://github.com/tlwillke/jvector/pull/55)
- Switch to ThreadLocal cached searcher in GraphIndexBuilder [`#56`](https://github.com/tlwillke/jvector/pull/56)
- Fix npe [`#53`](https://github.com/tlwillke/jvector/pull/53)
- Track numVisited when performing graph searches [`#51`](https://github.com/tlwillke/jvector/pull/51)
- Replace NBHM with CHM, remove NBHM dependency [`#49`](https://github.com/tlwillke/jvector/pull/49)
- Perf improvements [`#47`](https://github.com/tlwillke/jvector/pull/47)
- JVECTOR-31 [`#46`](https://github.com/tlwillke/jvector/pull/46)
- JVECTOR-8 [`#42`](https://github.com/tlwillke/jvector/pull/42)
- Fix offset math in testing MappedRandomAccessReader readFloatsAt [`#44`](https://github.com/tlwillke/jvector/pull/44)
- Issue 38 multiple jdk builds [`#40`](https://github.com/tlwillke/jvector/pull/40)
- JVECTOR-20 [`#41`](https://github.com/tlwillke/jvector/pull/41)
- Faster on-disk index flush in Bench [`#37`](https://github.com/tlwillke/jvector/pull/37)
- JVector 24: DiskANN [`#28`](https://github.com/tlwillke/jvector/pull/28)
- Integrate PQ to recall Bench [`#14`](https://github.com/tlwillke/jvector/pull/14)
- Add PanamaVectorSupport for Java 20+ [`#16`](https://github.com/tlwillke/jvector/pull/16)
- Added details for running both example classes via maven exec plugin [`#18`](https://github.com/tlwillke/jvector/pull/18)
- Integrate recall benchmark functionality [`#10`](https://github.com/tlwillke/jvector/pull/10)
- Add mvn wrapper [`#7`](https://github.com/tlwillke/jvector/pull/7)
- Initial clean up of pom.xml for packaging and release [`#5`](https://github.com/tlwillke/jvector/pull/5)
- Use float in cosine metric final calculation in default vectorization provider (#358) [`#357`](https://github.com/jbellis/jvector/issues/357)
- only compute aMagnitude for cosines once, since it is independent of the query vector [`#220`](https://github.com/tlwillke/jvector/issues/220)
- only compute aMagnitude for cosines once, since it is independent of the query vector [`#220`](https://github.com/tlwillke/jvector/issues/220)
- replace optimistic locking with pessimistic to prevent size() inconsistencies [`#218`](https://github.com/tlwillke/jvector/issues/218)
- Fix GraphIndexBuilder.reconnectOrphanedNodes [`#214`](https://github.com/tlwillke/jvector/issues/214)
- Improve GraphIndexBuilder#cleanup doc for concurrent searches [`#188`](https://github.com/tlwillke/jvector/issues/188)
- Make GraphIndexBuilder.markNodeDeleted thread-safe [`#174`](https://github.com/tlwillke/jvector/issues/174)
- rename BFS_DISTANCE to CACHE_distance; fixes #96 [`#96`](https://github.com/tlwillke/jvector/issues/96)
- make View extend AutoCloseable. Fixes #78 [`#78`](https://github.com/tlwillke/jvector/issues/78)
- update PQ to take RAVV. fixes #74 [`#74`](https://github.com/tlwillke/jvector/issues/74)
- fix dotProduct-with-offsets in Default provider. fixes #60 [`#60`](https://github.com/tlwillke/jvector/issues/60)
- apply alpha-diversity incrementally in enforceMaxConnLimit, like we do in insertDiverse. This significantly improves the quality of connections retained, vs just starting at the back with max alpha. Closes #33 [`#33`](https://github.com/tlwillke/jvector/issues/33)
- Created initial README.md [`#6`](https://github.com/tlwillke/jvector/issues/6)
- import hnsw + pq [`9f5b9c6`](https://github.com/tlwillke/jvector/commit/9f5b9c6a4f4f8d8bc0ba441246dfe265bfde4cc1)
- Squashed merge of PR #402: Add hierarchical structure to the graph index [`00a13a8`](https://github.com/tlwillke/jvector/commit/00a13a8a793009ffd13bb61374725653e0c50b87)
- Introduce OnDiskGraphIndex feature abstraction to embed additional content in graph (e.g., fused ADC or LVQ). Add the ability to incrementally write inline content to a graph index. Caching index only implements caching of edges. [`09f6850`](https://github.com/tlwillke/jvector/commit/09f68507b495a0dfe7269e85da74943e509f804e)
Loading
0