Milli v0.33.4

Changes

• Revert "Remove Bors required test for Windows" (#631) @curquiza
• Upgrade ubuntu-18.04 to 20.04 (#633) @curquiza
• Make charabia default feature optional (#632) @vincent-herlemont
• We avoid skipping errors in the indexing pipeline (#637) @Kerollmops

Thanks again to @Kerollmops, @curquiza, and @vincent-herlemont! 🎉

Milli v0.33.3

Changes

• Add CI to update the Milli version (#621) @curquiza
• Add dependabot for GHA (#623) @curquiza
• Make sure that long words are ignored (#628) @Kerollmops

Thanks again to @Kerollmops, @curquiza! 🎉

Milli v0.33.2

Changes

• Fix word criterion (#620) @ManyTheFish

Thanks again to @ManyTheFish ! 🎉

Milli v0.33.1

Changes

• Remove the artifacts of the past (#615) @irevoire
• Accept integers as document ids again (#617) @Kerollmops

Thanks again to @Kerollmops, and @irevoire! 🎉

Milli v0.33.0

API breaking

• Add EXISTS filter (#556) @loiclec
• Better threshold (#607) @irevoire
• Filter operators: NOT + IN[..] (#596) @loiclec
• Matching query terms policy (#598) @ManyTheFish
• Upgrade charabia v0.6.0 (#611) @ManyTheFish

DB breaking

• Add EXISTS filter (#556) @loiclec

Changes

• fix: Remove whitespace trimming during document id validation (#599) @ManyTheFish
• Stop reindexing already indexed documents (#579) @irevoire
• Simplify some unit tests (#600) @loiclec
• Introduce snapshot tests (#601) @loiclec
• Make binaries faster on release profile through better compile options (#606) @loiclec
• Use mimalloc as the default allocator (#602) @loiclec
• Optimise facets indexing (#590) @loiclec
• Fix(Search): Fix phrase search candidates computation (#594) @ManyTheFish
• Fix soft deleted documents (#608) @ManyTheFish
• Word prefix pair proximity docids indexation refactor (#587) @loiclec
• Remove Bors required test for Windows (#612) @curquiza

Thanks again to @ManyTheFish, @curquiza, @irevoire and @loiclec! 🎉

Milli v0.32.0

DB breaking

• Fasten documents deletion and update (#557) @irevoire

Changes

• Fix not equal filter when field contains both number and strings (#568) @GraDKh
• Bump uuid to 1.1.2 (#578) @Kerollmops
• Use BufReader to read datasets in benchmarks (#583) @loiclec
• Fix name of "release_date" facet in movies benchmarks (#588) @loiclec
• Enriched documents batch reader (#561) @Kerollmops

Thanks again to @GraDKh, @Kerollmops, @ManyTheFish, @bigstream-dmytro, @irevoire and @loiclec! 🎉

Milli v0.31.2

Release done on releaes-v0.31.2 branch, starting from v0.31.1 tag, and not from main. Will be used for Meilisearch v0.28.1 (hotfix)

Fixes

• Ensure that matches are all highlighted (#592) @ManyTheFish

Milli v0.31.1

Changes

• Introduce the copy_to_path method on the Index (#566) @Kerollmops

Thanks again to @Kerollmops ! 🎉

Milli v0.31.0

DB breaking

• Rename the limitedTo parameter into maxTotalHits (#564) @Kerollmops

Changes

• Improve the estimatedNbHits when a distinctAttribute is specified (#563) @Kerollmops

Thanks again to @Kerollmops ! 🎉

Milli v0.30.0

Fixes

• Avoid having an ending separator before crop marker (#559) @ManyTheFish

API breaking

• Ensure that Index methods are not bypassed by Meilisearch (#554) @ManyTheFish

Changes

• Deletion benchmarks (#558) @ManyTheFish

Thanks again to @ManyTheFish! 🎉