v1.0.0

What's Changed

Bug Fixes

• Remove liburing-devel from pre-cxx11 Dockerfile by @wxyucs with @Copilot in #1566
• fix(hnsw): fix concurrent access to max_level_ in HNSW search operations by @inabao in #1573
• fix(iter_filter): add bound check for iter filter context by @inabao in #1588
• fix(memory_usage): conditionally call cal_memory_usage only on actual capacity changes by @LHT129 in #1587
• fix(sindi): fix error when neither doc nor query is relevant by @inabao in #1589
• fix(build): fix ASAN build failure in diskann by @inabao in #1604
• fix(datacell): restore total_count_ during deserialization by @inabao in #1621
• fix(datacell): fix resize logic to update size after expansion by @inabao in #1642
• fix(hgraph): restore total_count after deserialization by @inabao in #1647
• fix(quantization): add missing quantization types to IsValidQuantizationType by @LHT129 in #1668
• fix(quantization): add null check in TransformQuantizer Computer destructor by @LHT129 in #1664
• fix(datacell): fix memory leak in FlattenDataCell query on exception by @wxyucs with @Copilot in #1680
• fix(hgraph): fix potential out-of-bounds access in Merge function by @inabao in #1626
• fix: correct GetStats unsupported error message by @wxyucs in #1699
• fix(quantizer): prevent division by zero in ScalarQuantizer::EncodeOneImpl by @LHT129 in #1666
• fix(hgraph): serialize ignore_reorder field in ToJson() by @LHT129 in #1711
• fix(pyramid): add missing RaBitQ parameter mappings by @inabao in #1703
• fix(diskann): fix memory leak in search paths by using RAII by @LHT129 in #1733
• fix: add INVALID_ENTRY_POINT guard to iterative KnnSearch and RangeSearch by @wxyucs with @Copilot in #1742
• fix(io): correct IO type name in AsyncIOParameter JSON constructor by @LHT129 in #1723
• fix(hgraph): add null check in GetVectorByInnerId to prevent crash during Tune by @LHT129 in #1761
• fix(quantization): add missing constexpr and static_assert in INT8Quantizer::ComputeImpl by @LHT129 in #1719
• fix: make VsagException public inherit from std::exception by @inabao in #1776
• perf(build): skip boost library compilation, only copy headers by @inabao in #1773
• fix(build): make clean targets remove hidden files in build directories by @LHT129 in #1781
• fix(io): check mremap return value to prevent potential crash by @LHT129 in #1725
• fix(build): filter sanitizer flags for external projects by @LHT129 in #1794
• fix(json_wrapper): fix undefined behavior in copy constructor by @LHT129 in #1715
• fix: prevent crash when KnnSearch and Tune run concurrently by @wxyucs with @Copilot in #1764
• fix(hnsw): synchronize parallel add graph updates by @inabao in #1802
• fix: propagate iterator filter init failures in hnsw by @wxyucs in #1844
• fix(build): avoid sanitizer linker flags in external projects by @wxyucs in #1849
• Fix Check Compatibility CI Failure by @LHT129 in #1862
• fix(simd): link simd_test against simd by @wxyucs in #1857
• fix(ci): publish pyvsag wheels on release by @jac0626 in #1888
• test(transform): fix flaky Basic Hadamard Test by raising minimum threshold by @LHT129 in #1893
• fix(quantization): correct FP16/BF16 conversion in QuantizerAdapter::ProcessQueryImpl by @LHT129 in #1898
• fix(ci): use pr url to obtain contributors PR number by @wxyucs in #1919
• fix(ci): fetch remote branch before force-with-lease push by @wxyucs in #1921
• fix: make lint parse .clang-tidy Checks correctly by @jac0626 in #1928
• fix: fallback async_io param and warn when libaio is unavailable by @jac0626 in #1932
• fix: prevent memory leak when MemoryIO reallocate fails by @LHT129 in #1853
• fix: correct FP16 conversion and add BF16 support in search reasoning; exempt FP16 from NEED_TRAIN flag by @LHT129 in #1944
• test: stabilize hnsw feedback functest by @jac0626 in #1954
• fix: avoid memcmp on dataset path strings by @jac0626 in #1952
• fix: make filtered graph search deterministic by @wxyucs in #1970
• fix: keep HGraph count consistent after add OOM by @wxyucs in #1975
• fix: CalDistanceById batch interface returns -1 for invalid IDs by @LHT129 in #2006
• fix: avoid NaN recall in HGraph duplicate range-search functest by @jac0626 in #2013
• fix: alloc-dealloc-mismatch in 110_index_warp example by @LHT129 in #2011
• fix: compilation failure on newer gcc version by @pkusunjy in #1982
• fix: avoid empty fast bitset memcpy by @jac0626 in #1983
• fix: avoid OOM in release packaging by @wxyucs in #2042
• fix: guard scalar quantizer NaN encoding by @jac0626 in #1997
• fix: normalize missing CalcDistanceById results by @LHT129 in #2044
• fix(hgraph): make repeated force remove idempotent by @LHT129 in #2080
• fix: add null-check before buf_ allocation in ProcessQueryImpl by @mukejane in #2165
• fix: guard extra_info pointer arithmetic against nullptr in BruteForce::Add by @LHT129 in #2168
• fix(datacell): prevent uint32 multiplication overflow in FlattenDataCell::query by @LHT129 in #2188
• fix(allocator): use throw instead of throw e to prevent exception slicing by @LHT129 in #2178
• fix(rabitq): use fabs to skip near-zero dims in EncodeExtendRaBitQ by @LHT129 in #2197
• fix(ci): install deps for AArch64 example tests by @wxyucs in #2192
• fix(io): zero-initialize MemoryBlockIO blocks to prevent reverse edge corruption by @LHT129 in #2212
• fix(pyramid): set duplicate_query_id in add_one_point for duplicate detection by @LHT129 in #2208
• fix(cluster): correct x_sqr vector offset for non-first batches in KMeans BLAS path by @LHT129 in #2201
• fix(c_api): return error when index is NULL in vsag_index_add and vsag_index_train by @LHT129 in #2183
• fix(ci): add shell: bash to Run Examples step to fix silent skip by @jac0626 in #2218
• fix(io): zero-initialize buffer in BasicIO::Resize fallback path by @LHT129 in #2221
• fix(dataset): handle SourceID in DeepCopy, Append, and destructor by @LHT129 in #2199
• fix(datacell): avoid integer overflow in FlattenDataCell::MergeOther offset by @LHT129 in #2210
• fix: prevent int16 accumulator overflow in SQ4UniformComputeCodesIPImpl by @LHT129 in #2229
• fix(io): guard resource cleanup on error paths in AsyncIO and MMapIO by @jfeng18 in #2205
• fix: defer NumElements update in DatasetImpl::Append for exception safety by @LHT129 in #2190
• fix(datacell): memory leak in SparseVectorDataCell query and ComputePairVectors on exception path by @LHT129 in #2242
• fix(quantizer): use correct copy length in FP32Quantizer::DecodeBatchImpl by @LHT129 in #2181
• fix(ci): strip -Ofast/-ffast-math from third-party build flags by @LHT129 in #2246
• fix(datacell): validate ID ranges in DenseDuplicateTracker deserialization by @LHT129 in #2258
• fix(hgraph): add shared_lock in CalDistanceById to prevent crash with concurrent Tune by @LHT129 in #2247
• fix(hgraph): avoid undefined behavior on nullptr arithmetic in get_data by @LHT129 in #2279
• fix: correct bucket bias calculation in IVF merge when buckets_per_data > 1 by @LHT129 in #2261
• fix(hgraph): add force_remove_mutex_ shared lock in UpdateVector to prevent data race by @LHT129 in #2272
• fix(quantizer): add nullptr check in SQ4UniformQuantizer::ProcessQueryImpl by @LHT129 in #2269
• fix(io): block until all in-flight async reads complete in MultiReadImpl by @LHT129 in #2287
• fix(searcher): correct candidate_set sign and lower_bound in iterator drain path by @LHT129 in #2293
• fix(io): add nullptr check in AsyncIO::ReadImpl before memcpy by @LHT129 in #2231
• fix(label_table): eliminate TOCTOU race in MarkRemove by @LHT129 in #2271
• fix(hgraph): update parallel_searcher_ mutex in SetImmutable() by @LHT129 in #2329
• fix(c-api): add result cap to range_search and validate file open status by @LHT129 in #2320
• fix: reuse existing BlackListFilter on repeated Run() calls to prevent memory leak by @LHT129 in #2317
• fix(hgraph): add radius filtering after reorder in RangeSearch by @LHT129 in #2321
• fix(bruteforce): RangeSearch must exclude MarkRemove-deleted documents by @LHT129 in #2330
• fix(storage): guard footer parse for short streams by @LHT129 in #2326
• bug: filter check in visit stage breaks graph connectivity and reduces recall by @LHT129 in #2314
• fix(hgraph): avoid GetStats crash for RabitQ split by @LightWant in #2360
• fix(simq): remove AddMode parameter from SIMQ::Add by @LHT129 in #2373
• fix: add minimum sample count check in PCATransformer::Train by @LHT129 in #2173
• fix: compact SINDI sparse term lists by @Roxanne0321 in #2351
• fix: guard inner index binary deserialization by @jac0626 in #1984
• [codex] fix main lint failures by @Roxanne0321 in #2384
• fix: support RabitQ bucket quantization in IVF by @jac0626 in #2323
• fix(hgraph): expose skip_ratio and skip_strategy search parameters by @LHT129 in #2367
• fix: avoid null write during skip deserialize by @LHT129 in #2403
• fix(heap): bound non-fixed MemmoveHeap Push to prevent buffer overflow (topk<10) by @jac0626 in #2299
• fix: resolve clang-tidy linter failures by @wxyucs in #2430
• fix: avoid clang-tidy shared pointer leak reports by @wxyucs in #2433

New Features

• feat(diskann): support calculate distance by id in diskann by @inabao in #1553
• feat(remove): refactor Remove with mark remove mode by @LHT129 in #1551
• feat(platform): Add support for macOS (Darwin) by @stuBirdFly in #1439
• feat(io): add Resize API for efficient storage allocation by @LHT129 in #1571
• Add Copilot instructions for repository coding standards by @wxyucs with @Copilot in #1586
• feat: upgrade yaml-cpp from 0.8.0 to 0.9.0 by @wxyucs in #1591
• feat: support to build vsag with system OpenBLAS by @wxyucs in #1594
• feat: support downloading third parties from custom URLs by @wxyucs in #1606
• feat(index): add ShrinkAndRepair API with timeout by @LHT129 in #1613
• feat: replace CPR with cpp-httplib and add HTTP monitor server by @wxyucs in #1616
• feat(datacell): add CheckIdExists interface to Graph DataCell classes by @LHT129 in #1653
• feat(label_table): add hole list for reusable inner_id by @LHT129 in #1683
• feat(hgraph): implement ShrinkAndRepair for deleted node reclamation by @LHT129 in #1706
• python:modernize python build system with native CMake and support Python 3.13 & 3.14 by @jac0626 in #1599
• feat(hgraph): add reverse edge support for HGraph index by @LHT129 in #1747
• feat(dataset): add FP16/BF16 input vector support by @LHT129 in #1731
• feat(api): add reasoning support to SearchRequest, Dataset and QueryContext by @LHT129 in #1838
• feat: add Node.js/TypeScript bindings and quickstart examples by @jiaweizone in #1812
• feat(reasoning): implement ReasoningContext data collector by @LHT129 in #1839
• feat(hgraph): add remove support for graph indexes by @LHT129 in #1810
• feat(warp): add multi-vector interface and brute force implementation by @xfmeng17 in #1945
• feat(sindi): add built-in term ID remapping for sparse vocabularies by @liric24 in #1961
• feat: add MultiVector dataset support and migrate WARP to MultiVectors API by @xfmeng17 in #1995
• [skip ci] feat: add AI-agent-assisted issue drafting workflow by @wxyucs in #2007
• feat(hgraph): integrate reasoning mechanism for search diagnostics by @LHT129 in #1911
• feat: support Pyramid ExportModel by @wxyucs in #2034
• feat(python): add sparse HDF5 dataset helpers for pyvsag by @wxyucs in #2036
• feat(simd): add AMX backends — SQ8U INT8 IP and BF16 GEMM for KMeans by @wxyucs in #2032
• Add SINDI analyze support by @Roxanne0321 in #2018
• feat: support parallel brute force range search by @wxyucs in #2040
• feat(eval): support multi-vector datasets by @LHT129 in #2031
• feat: add RabitQ split base reorder support by @LightWant in #1980
• feat: add MultiVectorDataCell for ColBERT-style multi-vector storage by @xfmeng17 in #2076
• feat: support disabling reorder through search params by @LHT129 in #2078
• feat: add Source ID support and cache export/import for HGraph by @LHT129 in #2093
• feat(hgraph): switch to brute-force search on low valid_ratio by @wxyucs in #2115
• feat: support system-installed OpenBLAS via VSAG_USE_SYSTEM_DEPS by @wxyucs in #2117
• feat(hgraph): accelerate Build by warm-starting from an imported cache by @LHT129 in #2114
• feat(hgraph): add support_force_remove switch by @LHT129 in #2086
• feat: use tl::unexpected in Index default implementations by @LHT129 in #2141
• feat: define pyramid named hierarchy API parameters by @jac0626 in #2104
• feat(pyramid): implement multi-hierarchy Build/Add/Search by @jac0626 in #2226
• feat: add SIMQ multi-vector retrieval index by @Sia-Sheerland in #2357
• feat: integrate reasoning into IVF search by @LHT129 in #2345
• feat: expose mrle_dim parameter to HGraph public API by @LHT129 in #2371
• feat(common): add repr field to IndexCommonParam by @LHT129 in #2379
• feat(sindi): add low-memory immutable build by @Roxanne0321 in #2424
• feat(serialization): add streaming serialization foundation by @wxyucs in #2256

Improvements

• refactor(inner_index): remove default thread pool during index creation by @inabao in #1558
• refactor(api): update public API comments by @wxyucs in #1554
• refactor(smart_pointer): replace shared_ptr with unique_ptr for better performance by @LHT129 in #1570
• feat(example): add HGraph extra_info usage example by @inabao in #1577
• refactor(duplicate): refactor record array for duplicate by @inabao in #1563
• refactor(example): simplify sindi example by @inabao in #1584
• enhance(label_table): use robin_pg_map instead of unordered_map by @inabao in #1592
• refactor(pyramid): modify default value of pyramid thread pool by @inabao in #1595
• refactor(test): unify filter logic based on ID and extra-info by @inabao in #1600
• feat(pyramid): move topk_factor to IndexSearchParameter base class by @inabao in #1605
• refactor(tests): migrate Python tests to pytest framework by @LHT129 in #1619
• refactor(quantizer): deduplicate batch methods in CRTP base class by @LHT129 in #1624
• feat(hgraph): add hops_limit search parameter by @inabao in #1620
• chore(ci): remove pull_request trigger from coverage workflow by @LHT129 in #1675
• fix: add missing namespaces for stream serialization helpers by @wxyucs in #1669
• perf(simd): optimize FP32 distance computation with reduce and templa… by @LHT129 in #1625
• refactor(hnsw,io): correct error type for memory allocation by @inabao in #1658
• fix(quantization): add null pointer check in Computer destructor by @LHT129 in #1662
• chore(ci): migrate compatibility indexes from git branch to GitHub Releases by @LHT129 in #1692
• test: add unit coverage for utils module by @wxyucs in #1700
• fix(ci): improve index upload logic to compatibility-indexes release by @LHT129 in #1708
• refactor(io): replace C-style cast with C++ style cast by @LHT129 in #1641
• refactor(io): extract platform-specific syscalls to IOSyscall by @LHT129 in #1629
• fix(datacell): add virtual destructor to ExtraInfoInterface by @LHT129 in #1717
• perf(label_table): add TryGetIdByLabel to avoid exception overhead by @inabao in #1701
• feat(analyzer): implement duplicate detection in GetDuplicate by @inabao in #1705
• chore(build): isolate MKL and OpenBLAS paths by @wxyucs in #1709
• perf(pyramid): trim flat search path overhead by @wxyucs in #1713
• fix: use quoted project includes in search_param.h by @wxyucs in #1712
• refactor: decentralize factory and python bindings by @wxyucs in #1697
• feat(ci): merge workflows into unified PR CI with path filtering by @LHT129 in #1744
• refactor: replace spdlog with internal default logger by @wxyucs in #1750
• refactor: reorganize and clean up cmake configuration by @wxyucs in #1762
• perf(cluster): optimize KMeans algorithm with K-means++ and parallel improvement by @LHT129 in #1737
• test(impl): add unit tests for untested components in impl module by @LHT129 in #1728
• refactor(hnsw): remove static HNSW implementation by @LHT129 in #1784
• perf(build): reduce diskann static library size by @LHT129 in #1792
• style: unify 'constexpr static' to 'static constexpr' by @LHT129 in #1782
• fix: unify virtual/override keyword usage in destructors by @LHT129 in #1763
• ci(mergify): upgrade configuration to current format by @mergify[bot] in #1799
• chore: change default value of ENABLE_INTEL_MKL to OFF and trigger Python Build X86 on CMake changes by @wxyucs with @Copilot in #1806
• refactor(scripts): dynamic compatibility index list from release files by @LHT129 in #1813
• docs(clang-tidy): Switched from > to | in .clang-tidy to pr… by @pkusunjy in #1828
• refactor(tests): modularize fixtures and split functest/unittest include by @LHT129 in #1821
• build: slim default build targets by @wxyucs in #1796
• chore: enforce kind and version labels on PRs via Mergify by @wxyucs in #1850
• ci(mergify): upgrade configuration to current format by @mergify[bot] in #1854
• chore: replace all-contributors with contributors-readme-action by @wxyucs in #1856
• refactor(simd): unify SIMD dispatch cascades via shared macros by @LHT129 in #1871
• ci: skip heavy builds on docs-only changes by @wxyucs in #1885
• chore(ci): allow manual dispatch for docs sync workflow by @wxyucs in #1897
• chore(ci): skip docs-only changes on push-triggered workflows by @wxyucs in #1901
• refactor(quantizer): eliminate code redundancy in FP16/BF16 quantizers using template by @LHT129 in #1855
• test(fixtures): use fixed seed for reproducible test data generation by @LHT129 in #1872
• test: add timestamp to UT log output by @LHT129 in #1909
• chore: make fmt and lint compatible with macOS (Homebrew llvm@15) by @jac0626 in #1906
• Unify exception types to use VsagException by @LHT129 in #1852
• test: normalize functest tag system by @LHT129 in #1902
• perf(index): replace stats_mutex_ with lock-free atomic counter in WindowResultQueue by @LHT129 in #1721
• test: move expression_visitor test from functests to unittests by @LHT129 in #1895
• chore: remove ShrinkAndRepair API and related shrink code by @LHT129 in #1924
• refactor(quantization): remove redundant DataType alias by @LHT129 in #1938
• fix: reduce PGUnorderedMap overhead in label indexes by @LHT129 in #1940
• ci: skip tsan for markdown-only pull requests by @wxyucs in #1972
• chore: harden C++ release workflow by @wxyucs in #1971
• refactor: remove RemoveMode backward-compatible aliases by @LHT129 in #1978
• fix: make public headers self-contained by @LHT129 in #1987
• feat(simd): add SIMD implementations for NormalizeWithCentroid by @LHT129 in #1926
• refactor: deduplicate hgraph and ivf functests by @LHT129 in #1993
• ci: correctly skip tsan for markdown-only pull requests by @wxyucs in #2000
• feat: allow selecting inner index label remap map by @LHT129 in #1981
• refactor: split typing aliases and slim common param includes by @LHT129 in #1991
• ci: skip coverage workflow for markdown-only pushes by @wxyucs in #2038
• ci: require linked issue for kind/bug and kind/feature PRs by @wxyucs in #2039
• feat: add HGraph duplicate distance threshold by @LHT129 in #2051
• test: split TestIndex implementation into focused files by @LHT129 in #2058
• refactor(hgraph): create hgraph directory and split cpp files by @LHT129 in #2030
• build: install CMake package config and set library SOVERSION by @wxyucs in #2106
• chore: add Apache 2.0 license headers to remaining source files by @wxyucs in #2107
• refactor(label_table): extract LabelRemap class and remove Hole functionality by @LHT129 in #2089
• chore(docker): split ci-x86 image into base / mkl / openblas variants by @wxyucs in #2119
• feat(sparse): add optional token_sequences to sparse HDF5 datasets by @Roxanne0321 in #2111
• fix: honor system dependency setting in Makefile by @wxyucs in #2122
• ci: switch fast-path jobs to system OpenBLAS and split-image tags by @wxyucs in #2120
• refactor(algorithm): move scattered index implementations into dedicated subdirectories by @LHT129 in #2127
• feat(python): unify FP16/BF16 dense binding interfaces by @LHT129 in #1946
• chore: bundle dependencies for release artifacts by @wxyucs in #2142
• refactor(algorithm): extract common search helpers to InnerIndexInterface base class by @LHT129 in #2139
• fix(hgraph): correct build-cache source_id mapping and improve warm-start performance by @misaka0714 in #2125
• build: install OpenBLAS in dependency scripts by @wxyucs in #2143
• chore: align coverage threshold check by @wxyucs in #2149
• test: add coverage tests for rabitq_split, mmap_io, warp, and c_api by @LHT129 in #2148
• refactor(parameter): introduce macros to eliminate CheckCompatibility boilerplate by @LHT129 in #2153
• fix: expand Pyramid ef search for large topk by @jac0626 in #2162
• fix(ci): keep docs sync branch aligned by @wxyucs in #2159
• build: support system fmt dependency by @wxyucs in #2169
• perf(simd): optimize AVX-512 reductions and remove dead allocation by @jfeng18 in #2155
• perf(hgraph): skip mutex acquisition on immutable index search by @jfeng18 in #2154
• perf: avoid unnecessary copies in ConjugateGraph::Serialize loop by @LHT129 in #2161
• feat: add LazyHGraph index by @Roxanne0321 in #2151
• refactor(simd): unify cross-ISA implementations via traits + kernel templates by @LHT129 in #2156
• ci: add aarch64 CI docker images and run arm jobs in containers by @wxyucs in #2198
• perf(sindi): grow sparse term list safely by @Roxanne0321 in #2195
• perf(simd): use dual accumulators in SQ4 kernel templates for better ILP by @jfeng18 in #2227
• feat(hgraph): Refactor HGraph node refinement with immediate neighbor writeback by @misaka0714 in #2174
• Remove SO version suffix from libvsag.so by @LHT129 in #2224
• refactor: extract total_count_ to InnerIndexInterface base class by @LHT129 in #2157
• feat(hgraph): expose ImportCache build hit-rate via GetStats by @LHT129 in #2238
• feat(pyramid,sindi,warp): add MARK_REMOVE support by @LHT129 in #2136
• refactor(searcher): remove dead to_be_visited_rid allocation and writes by @LHT129 in #2251
• test(reasoning): add comprehensive test coverage for ReasoningContext by @LHT129 in #2267
• refactor(algorithm): merge WARP into BruteForce as multi-vector mode by @LHT129 in #2275
• feat(rabitq): support split hybrid IO and 3+5 base reorder by @LightWant in #2112
• test: enhance quantization test coverage and quality by @LHT129 in #2316
• ci: reuse ASan build artifacts for compatibility check by @LHT129 in #2342
• test: optimize TSAN CI test performance by @LHT129 in #2344
• ci: run clang-tidy only on changed files in PR CI by @LHT129 in #2339
• fix: resolve src/ circular dependencies and establish unidirectional layering by @LHT129 in #2348
• refactor(index): unify KnnSearch/RangeSearch into SearchWithRequest for IVF/HGraph/BruteForce by @LHT129 in #2341
• refactor: remove unused AddMode parameter from Index::Add by @LHT129 in #2366
• feat(sindi): support immutable reads and sparse fp16 values by @Roxanne0321 in #2296
• feat(bruteforce): integrate reasoning mechanism for search diagnostics by @LHT129 in #2328
• Fix/sindi auto heap insert by @Roxanne0321 in #2147
• feat(lazy_hgraph): support extra info filtering by @Roxanne0321 in #2263
• feat(sindi): use sparse vector datacell and remove sparse index by @Roxanne0321 in #2056
• [codex] Add macOS build and PR CI support by @jac0626 in #1930
• perf: optimize rabitq split 3-bit filtering by @LightWant in #2390
• fix(hgraph): optimize rabitq split 2bit search by @LightWant in #2407
• refactor(io): split io files into subdirectories by @LHT129 in #2393
• test(diskann): reduce pq_dim test runs from 3 to 1 in PR CI by @LHT129 in #2404
• fix: rename MMapIO::start_ to mapped_ptr_ to avoid shadowing BasicIO::start_ by @LHT129 in #2398
• feat(logger): support VSAG_LOG_LEVEL by @wxyucs in #2414
• feat: unify memory statistics API signatures by @LHT129 in #2388
• feat(simq): add search statistics by @LHT129 in #2418

Other Changes

• [skip ci] docs: add SINDI algorithm documentation by @inabao in #1572
• fix(reader_io): fix missing parameter when deserialize index by @inabao in #1597
• [skip ci]Update Roadmap in README.md by @wxyucs in #1560
• fix(io): add MAP_FAILED error check for mmap by @LHT129 in #1655
• chore(io): remove unused io_read_object.h file by @LHT129 in #1640
• optimize: avoid repeated align param calculation in DirectIOObject.Set by @LHT129 in #1638
• feat(python): extend Python bindings with additional index operations by @LHT129 in #1678
• [skip ci] docs: add Claude guidance for coding agents by @wxyucs in #1686
• [skip ci] chore: complete P0 workflow hygiene by @wxyucs in #1693
• refactor(quantizer): extract mapper array as static class member by @LHT129 in #1671
• chore: simplify CODEOWNERS to require all four maintainers for every … by @wxyucs in #1702
• docs(quantization): add layout diagrams for all quantizer types by @LHT129 in #1688
• fix(diskann): normalize data for COSINE metric in PQ quantization by @inabao in #1745
• fix(hgraph): enable continuous shrink processing in HGraphShrinkContext by @LHT129 in #1790
• [skip ci] docs(readme): refresh publications list by @inabao in #1803
• fix: narrowing conversion warnings by @pkusunjy in #1800
• feat(pyramid): add PyramidAnalyzer for index statistics and diagnostics by @inabao in #1766
• refactor(reverse_edge): move reverse edge logic from HGraph to graph datacell by @LHT129 in #1805
• refactor(hgraph): move duplicate tracking to graph layer by @inabao in #1797
• docs: add arXiv badges for all publications in README by @wxyucs in #1811
• docs(io): add Doxygen comments for src/io directory by @LHT129 in #1817
• chore(docs): refactor docs/ dir and add GH Action to sync docs by @wxyucs in #1333
• docs(contributor): contributors readme action update by @github-actions[bot] in #1859
• docs: add AGENTS.md for openai codex and gitignore .DS_Store by @xfmeng17 in #1866
• [skip ci] docs: restructure docs/blog for mdbook and rewrite sync workflow by @wxyucs in #1882
• docs: correct 'last updated' date in contributors list by @wxyucs in #1883
• docs(contributor): contributors readme action update by @github-actions[bot] in #1884
• [skip ci] docs: refresh zh/en guides and mirror full en docs by @wxyucs in #1886
• docs(contributor): contributors readme action update by @github-actions[bot] in #1903
• docs(landing): add copy buttons, refine indexes table, update ecosystem by @wxyucs in #1910
• docs(indexes): add dedicated pages for HGraph, IVF, SINDI, Pyramid by @wxyucs in #1914
• fix(ci): stabilize contributors readme updates by @wxyucs in #1913
• docs(contributor): contributors readme action update by @github-actions[bot] in #1922
• docs: add language toggle button on docs, blogs and landing by @wxyucs in #1925
• docs: reorganize documentation navigation by @wxyucs in #1934
• docs: add Index Analysis page for AnalyzeIndexBySearch and analyze_index tool by @wxyucs in #1941
• docs(contributor): contributors readme action update by @github-actions[bot] in #1935
• docs: fix CMake quickstart example by @wxyucs in #1948
• docs(contributor): contributors readme action update by @github-actions[bot] in #1956
• docs: update eval_performance tool documentation by @wxyucs in #1967
• docs(contributor): contributors readme action update by @github-actions[bot] in #1988
• docs: clarify CalcDistanceById/CalDistanceById overload usage by @LHT129 in #1992
• docs: update sindi parameter descriptions by @Roxanne0321 in #1998
• [skip ci] docs: align AI sign-off rules with Linux kernel policy by @wxyucs in #2001
• docs: add user guide for extra_info feature by @wxyucs in #2004
• docs(contributor): contributors readme action update by @github-actions[bot] in #2003
• docs(site): add Google Analytics (gtag.js) to vsag.io by @wxyucs in #2010
• docs: add filtered search and attribute (hybrid) search guides by @wxyucs in #2016
• docs: add HDF5 dataset format page to Resources by @wxyucs in #2019
• docs: add Index Lifecycle Management page by @wxyucs in #2020
• docs: add BruteForce index page (en + zh) by @wxyucs in #2023
• docs: expand memory page and add Index Introspection page by @wxyucs in #2021
• docs: document non-fp32 input types for HGraph by @wxyucs in #2022
• docs: add Build and Train page (Train API + ODescent) by @wxyucs in #2024
• docs(contributor): contributors readme action update by @github-actions[bot] in #2035
• docs: reorganize agent guidance by @wxyucs in #2033
• docs: add Per-Search Allocator advanced page by @wxyucs in #2025
• docs: add extensibility guide for advanced features by @wxyucs in #2026
• docs: add Quantization Transform advanced page by @wxyucs in #2027
• docs(contributor): contributors readme action update by @github-actions[bot] in #2041
• docs: add metric semantics note by @wxyucs in #2043
• docs(contributor): contributors readme action update by @github-actions[bot] in #2048
• docs: update Google Analytics tracking tag by @wxyucs in #2065
• docs: document CalcDistanceById and CalDistanceById APIs by @wxyucs in #2067
• docs: refresh site logo across landing page and mdbook subsites by @wxyucs in #2075
• docs: add Quantization chapter with per-quantizer reference pages by @wxyucs in #2074
• docs: add agent guide for filing GitHub issues by @wxyucs in #2082
• docs(indexes): add overview SVG diagrams for each index by @wxyucs in #2091
• docs: add per-language README to examples/{cpp,python,typescript}/ by @wxyucs in #2109
• docs: add iterator search page (EN/ZH) by @wxyucs in #2110
• docs(contributor): contributors readme action update by @github-actions[bot] in #2084
• docs(hgraph): document brute_force_threshold and add example by @wxyucs in #2124
• refactor: remove tombstone recovery functionality by @LHT129 in #2134
• docs: consolidate documentation entry points by @wxyucs in #2150
• docs(contributor): contributors readme action update by @github-actions[bot] in #2146
• docs(contributor): contributors readme action update by @github-actions[bot] in #2184
• docs: use HGraph in example snippets instead of deprecated HNSW by @wxyucs in #2203
• docs(contributor): update contributors list and sync zh page by @wxyucs in #2191
• docs(contributor): contributors readme action update by @github-actions[bot] in #2225
• docs(contributor): contributors readme action update by @github-actions[bot] in #2248
• docs(pyramid): add multi-hierarchy example and documentation by @jac0626 in #2252
• docs: add offline / internal-mirror third-party build guide by @wxyucs in #2307
• docs: remove hnsw/diskann references from user-facing docs by @wxyucs in #2343
• docs: add Doxygen comments to InnerIndex subclass headers by @LHT129 in #2338
• docs: add FAQ documentation by @wxyucs in #2362
• refactor: migrate test_index_old tests to target files and remove redundant file by @LHT129 in #2361
• docs: add HGraph-based disk index best practices guide by @wxyucs in #2382
• docs: add C++ public API reference (en + zh) by @wxyucs in #2391

New Contributors

• @mergify[bot] made their first contribution in #1799
• @pkusunjy made their first contribution in #1800
• @jiaweizone made their first contribution in #1812
• @xfmeng17 made their first contribution in #1866
• @liric24 made their first contribution in #1961
• @LightWant made their first contribution in #1980
• @mukejane made their first contribution in #2165
• @jfeng18 made their first contribution in #2155
• @Sia-Sheerland made their first contribution in #2357

Full Changelog: v0.18.0...v1.0.0