What's Changed
Version 2.0.0 changes the default MinHash permutation scheme to "affine32", which fixes a similarity over-estimation bias on large sets (#212), halves sketch memory, and speeds up updates by roughly 4x. A 64-bit "affine64" scheme is available for billion-scale sets. Hash values differ from earlier versions: rebuild persisted sketches and LSH indexes, or pass MinHash(..., scheme="legacy") to interoperate with existing data. See the MinHash documentation for details.
- Fix WeightedMinHash input mutation by @dipeshbabu in #317
- Preserve custom hash functions in sketch unions by @dipeshbabu in #318
- Bump actions/checkout from 6 to 7 by @dependabot[bot] in #322
- Add affine MinHash permutation schemes; refresh benchmarks and docs by @ekzhu in #323
- Bump codecov/codecov-action from 6 to 7 by @dependabot[bot] in #321
- Bump pillow from 11.3.0 to 12.2.0 in the uv group across 1 directory by @dependabot[bot] in #324
Full Changelog: v1.10.0...v2.0.0
Contributors
ekzhu, dependabot, and dipeshbabu