0.9.0
Highlights 🌈
- More ranking factors based on number of typos and exactness 🥳
- New
Loupe::deleteAllDocuments()method 😎 - Order of
searchableAttributesnow influences relevance weighting 💪 - Ranking factors/rules can be customized 😋
- Added support for stop words 🫶
- Changed to Damerau-Levenshtein 🙌
- Performance, Loupe has gotten faster again, both for indexing and searching! 🚀
Changelog
- Usage docs by @daun in #109
- Delete all documents by @daun in #110
- Attribute ranking order by @daun in #111
- Document ranking of results by @daun in #113
- Stop words by @daun in #107
- Update to toflar/state-set-index v3 by @Toflar in #118
- Remove orphaned states after deleting or updating documents by @daun in #117
- Switch to Damerau-Levenshtein by @Toflar in #122
- Benchmark ergonomics by @daun in #126
- Speed up indexing through granular removal of existing data by @daun in #127
- Some more sqlite optimizations by @daun in #129
- Allow to customize the ranking rules by @daun in #128
- Implemented a static cache making sure it's reset on object destruction by @Toflar in #136
- Fix PHP deprecation by @norkunas in #139
- Prevent "endless" queries by @Toflar in #137
- Highlight stop words if they appear in the query by @daun in #132
- Reworked relevance rankers and fixed proximity ranker by @Toflar in #141
- Rank on exactness and typos by @Toflar in #124
- Use binary format for state set cache for better performance and OPcache exclusion by @Toflar in #144
New Contributors
Full Changelog: 0.8.3...0.9.0
Special shoutouts
- To @ausi for working on https://github.com/Toflar/state-set-index enabling Damerau-Levenshtein support ❤️
- To @daun for playing an important role in this release ❤️
Contributors
ausi, Toflar, and 2 other contributors