TileDB-Vector-Search 0.0.4

What's Changed

• Clean-up ingestion.py and reduce non-verbose log messages. by @teo-tsirpanis in #82
• Reduce the minimum accuracy in test_ingestion.py and move it to a constant. by @teo-tsirpanis in #84
• Remove unit tests that were really utility programs by @lums658 in #85
• Add tiling optimization and refactor to single function by @lums658 in #86
• Enable all C++ tests in CI by @ihnorton in #87
• Set vacuum and consolidation mode in ingestion by @Shelnutt2 in #83
• Update Python API to use blocked implementations from PR #86 by @ihnorton in #88
• Avoid creating contexts to pass the consolidation/vacuum parameters. by @teo-tsirpanis in #89
• Distributed query implementation by @NikolaosPapailiou in #90

Full Changelog: 0.0.3...0.0.4

TileDB-Vector-Search 0.0.3

What's Changed

• Wrap partition_ivf_index and tdbPartitionedMatrix by @ihnorton in #70
• Update python api by @NikolaosPapailiou in #79
• Use groups instead of os.path.join and register arrays with relative … by @NikolaosPapailiou in #81
• Fix segfault in finite ram queries by @NikolaosPapailiou in #80

Full Changelog: 0.0.2...0.0.3

TileDB-Vector-Search 0.0.2

What's Changed

• Update path in Dockerfile by @ihnorton in #66
• Run all python tests, they no longer depend on a token by @ihnorton in #69
• Log stats on array openings and query submissions. by @teo-tsirpanis in #52
• Use native vector ingestion in Python taskgraphs by @NikolaosPapailiou in #67
• New interfaces in support of distributed execution by @lums658 in #68
• Cleanup for release / Support for distributed execution by @lums658 in #71
• Pin pydantic because machow/quartodoc#209 by @ihnorton in #72
• Fix non-void return path warning by @lums658 in #73
• Install OpenBLAS in Python CI. by @teo-tsirpanis in #74
• Fix empty read slices in ingestion by @NikolaosPapailiou in #75
• Call load() for tdbColMajorMatrix in ingestion by @NikolaosPapailiou in #78

Full Changelog: 0.0.1...0.0.2

TileDB-Vector-Search 0.0.1

Initial tag of the TileDB-Vector-Search repo.

What's Changed

• Restructure repo by @NikolaosPapailiou in #1
• Delete leftover files. by @teo-tsirpanis in #5
• Initial k-means query support in C++. by @lums658 in #6
• Small fixes by @teo-tsirpanis in #8
• Prepare library for integration with Python by @lums658 in #12
• lums/tmp/queries by @lums658 in #14
• Python ingestion by @NikolaosPapailiou in #11
• Add CI. by @teo-tsirpanis in #9
• Use Accelerate to fix build on macOS after PR9, add USE_MKL_CBLAS to … by @ihnorton in #16
• lums/tmp/async ivf by @lums658 in #15
• lums/tmp/benchmarking by @lums658 in #18
• Python bindings and FlatIndex API by @ihnorton in #7
• Add flat_index demo notebook by @ihnorton in #22
• Fix CI by skipping failing tests. by @teo-tsirpanis in #23
• Avoid creating contexts inside library functions. by @teo-tsirpanis in #20
• Ctx updates: keep_alive; zero-arg constructor by @ihnorton in #25
• Add option for running ingestion using local mode by @NikolaosPapailiou in #27
• pip support by @ihnorton in #26
• Add Dockerfile by @NikolaosPapailiou in #28
• Vector ingestion by @NikolaosPapailiou in #31
• lums/tmp/out of core benchmarking by @lums658 in #32
• Reformat all files with Black by @ihnorton in #33
• Improve ivf_flat ingestion by @NikolaosPapailiou in #34
• Add IVF_Flat index by @NikolaosPapailiou in #35
• Install libtiledb and set rpath for usage with packaged extension by @ihnorton in #37
• Minor fixes by @ihnorton in #38
• lums/tmp/kmeans by @lums658 in #36
• Rename experimental -> src by @ihnorton in #39
• Fix IVF_Flat by @NikolaosPapailiou in #40
• Add support for fine-grained timing by @lums658 in #41
• Separate open and use for infinite ram [skip ci] by @lums658 in #43
• Bump Catch2 version to 3.3.2 to fix GCC 13 build by @ihnorton in #44
• Only install catch2 if tests are enabled by @Shelnutt2 in #45
• Support setting cmake options in python via env by @Shelnutt2 in #46
• Remove .idea folder by @Shelnutt2 in #48
• Fix a number of warnings by @Shelnutt2 in #50
• Remove unused global variable global_region by @Shelnutt2 in #53
• Enable Quarto rendering for docs and reference by @Shelnutt2 in #51
• Add benchmarks README by @ihnorton in #49
• Fix README installation link by @Shelnutt2 in #54
• Py fixes; Run Python tests and pip builds by @ihnorton in #30
• Use preloaded matrices as input for kmeans query call by @ihnorton in #47
• Clean up/harmonize READMEs by @ihnorton in #55
• Don't use editable pip install: broken; require pip>22 for in-source … by @ihnorton in #56
• Add .gitattributes for packaging by @Shelnutt2 in #58
• Minor cleanup by @ihnorton in #59
• Set cmake flags based on selected processor by @Shelnutt2 in #60
• Infrastructure for benchmarking and cleanup for release by @lums658 in #61
• Remove array_types.h by @lums658 in #62
• Fix MKL builds on Linux. by @teo-tsirpanis in #63
• Add support for sizes arrays by @lums658 in #64
• Use std::vector as opaque type to avoid copy by @ihnorton in #65

New Contributors

• @NikolaosPapailiou made their first contribution in #1
• @lums658 made their first contribution in #6
• @ihnorton made their first contribution in #16
• @Shelnutt2 made their first contribution in #45

Full Changelog: https://github.com/TileDB-Inc/TileDB-Vector-Search/commits/0.0.1-alpha