Changelog

Breaking changes

• CountMinSketch now has a type parameter for the count type. Possible values are u8 to u64 and i8 to i64.
• HllUnion::get_result is renamed to HllUnion::to_sketch.
• update_f32 and update_f64 are removed from ThetaSketch. Use hash_value's wrapper instead.
• All sketches are now gated by a feature flag. You need to enable the feature flag to use the sketch. For example, to use CountMinSketch, you need to enable the countmin feature.

Notable changes

• The MSRV (Minimum Supported Rust Version) is now 1.86.0.

New features

• New module hash_value provides several value wrappers for matching concrete hashing strategies.
• CountMinSketch with unsigned values now supports halve and decay operations.
• CpcSketch and CpcUnion are now available for cardinality estimation.
• CpcWrapper is now available for reading estimation from a serialized CpcSketch without full deserialization.
• FrequentItemsSketch now supports serde for any value implement FrequentItemValue (builtin supports for i64, u64, and String).
• Expose codec::SketchBytes, codec::SketchSlice, and FrequentItemValue as public API.
• hll::Coupon is now public. You can calculate the coupon and reuse it multiple times avoiding the overhead of hashing multiple times.

Commit history

• docs: document build and test workflows by @PsiACE in #69
• chore: enable more convenient GitHub features by @tisonkun in #74
• chore: correct CountMinSketch::suggest_num_hashes panic docs by @tisonkun in #73
• feat: add halve and decay to countmin sketch of unsigned values by @PsiACE in #71
• chore: update README and add CHANGELOG by @tisonkun in #78
• refactor: remove unnecessary allocations in error formatting by @EricccTaiwan in #79
• test: use snapshots by @tisonkun in #80
• feat: impl CpcSketch by @tisonkun in #75
• refactor!: rename get_result to to_sketch by @tisonkun in #81
• chore: code tidy for FrequentItems and impl for more value type by @tisonkun in #82
• chore: define static lookup table as static variables by @tisonkun in #83
• feat: support cpcsketch serde by @tisonkun in #84
• refactor: sort PairTable values with std sort_unstable by @tisonkun in #86
• chore(tdigest): better handle Centroid adds by @tisonkun in #88
• chore: check seed for CpcSketch and CountMinSketch by @tisonkun in #85
• refactor: expose codec and add centralized Family by @tisonkun in #91
• chore: Remove early development caution warning by @tisonkun in #94
• fix(bloom): align builder parameters with java implementation by @notfilippo in #93
• chore(x): add doc lint by @notfilippo in #95
• feat(theta): compact sketch serde v3/v4 by @ZENOTME in #77
• refactor: export FrequentItemValue and improve docs by @tisonkun in #98
• feat: impl CpcWrapper by @tisonkun in #101
• chore: simplify codec checks code by @tisonkun in #102
• chore: fine tune binomial_bounds.rs code by @tisonkun in #103
• chore(theta): flatten preamble_longs branches for clippy by @Xuanwo in #104
• fix: comment and assertions inconsistency by @YichiZhang0613 in #106
• feat(theta): introduce intersection theta set operation by @ZENOTME in #100
• fix(theta): update compact method to use is_empty for consistency by @ZENOTME in #107
• chore: update copyright year in NOTICE by @caicancai in #110
• Made update_with_coupon public by @fulmicoton-dd in #113
• chore: remove incomplete examples module by @caicancai in #114
• tests: add boundary tests for bit_pack and CpcSketch by @YichiZhang0613 in #108
• feat(hll): expose coupon fn and improve update_with_coupon docs by @notfilippo in #116
• fix(hll): update silently dropped after deserializing a compact List sketch by @notfilippo in #117
• chore: Add back required approving review count to PR settings by @tisonkun in #119
• refactor: introduce hash_value for compatible hashing by @tisonkun in #118
• refactor: gate sketches with feature flags by @tisonkun in #120
• chore: changelog and release docs by @tisonkun in #121
• fix: correct assetion logic by @tabac in #122
• test: for hll::Array4::shift_to_bigger_cur_min by @tisonkun in #123
• chore: bump version to 0.3.0 by @ZENOTME in #125
• chore: fixup signature verification commands in RELEASE.md by @tisonkun in #124
• chore: ensure lockfile up-to-date by @tisonkun in #126

New Contributors

• @EricccTaiwan made their first contribution in #79
• @YichiZhang0613 made their first contribution in #106
• @caicancai made their first contribution in #110
• @fulmicoton-dd made their first contribution in #113
• @tabac made their first contribution in #122

Full Changelog: 0.2.0...0.3.0

What's Changed

• chore: Cargo project bootstrap by @notfilippo in #6
• chore: Add missing license headers by @freakyzoidberg in #7
• ci: Enable squash merge and disable wiki by @Xuanwo in #8
• chore: Fix formatting and wording in README.md by @tisonkun in #9
• feat: HLL backend implementation by @notfilippo in #1
• ci: Bootstrap basic Rust CI by @tisonkun in #13
• ci: install rust toolchain without actions by @tisonkun in #16
• chore: fixup new typos glossary and linters by @tisonkun in #18
• chore: fine tune asfyaml settings by @tisonkun in #19
• chore: code tidy integration tests by @tisonkun in #20
• chore: check InsufficientData before index access by @tisonkun in #24
• feat(hll): implement HllUnion by @notfilippo in #25
• test(hll): HllUnion suite cleanup by @notfilippo in #26
• build: generate snapshots on the fly by @tisonkun in #29
• chore: Prevent compilation on big-endian targets by @tisonkun in #31
• feat: implement T-Digest by @tisonkun in #23
• refactor: implement MurmurHash3 internally by @tisonkun in #36
• chore: rustfmt rules on nightly by @tisonkun in #33
• chore: fix CI failing on main by @notfilippo in #40
• build: set up workspace layout by @tisonkun in #43
• feat: add xxhash64 by @PsiACE in #39
• feat: implement count-min sketch by @PsiACE in #38
• feat: add a generic error type by @tisonkun in #47
• chore: centralize DEFAULT_UPDATE_SEED by @tisonkun in #48
• feat: add theta sketch (part 1) by @ZENOTME in #45
• refactor: replace byteorder with SketchBytes and SketchSlice by @tisonkun in #49
• feat: add frequencies sketches by @PsiACE in #44
• chore: fine tune frequencies and theta sketches by @tisonkun in #51
• refactor: use new codec utils in TDigest and Frequencies sketches by @tisonkun in #52
• feat: implement bloom filter by @notfilippo in #53
• chore: prepare v0.2.0 release by @notfilippo in #57
• docs: document release process by @notfilippo in #61
• docs: add rustdoc examples across sketches by @PsiACE in #58
• ci: generate snaps stably by @tisonkun in #65
• chore: Remove required approving review count from PRs by @tisonkun in #66
• fix: tdigest quantile edge case by @tisonkun in #63
• feat: add theta sketch (part 2) by @ZENOTME in #59

New Contributors

• @notfilippo made their first contribution in #6
• @freakyzoidberg made their first contribution in #7
• @Xuanwo made their first contribution in #8
• @tisonkun made their first contribution in #9
• @PsiACE made their first contribution in #39
• @ZENOTME made their first contribution in #45

Full Changelog: https://github.com/apache/datasketches-rust/commits/0.2.0