Why is this version considered as breaking?
We consider this release as breaking because it requires re-indexation of all documents. API usage remains unchanged from v0.22.
Re-indexation happens automatically when migrating your data from a previous version. For more information on how to update to the latest version of MeiliSearch, read our dedicated guide.
Also, we now enforce the Content-Type
header to all the API routes accepting a payload. It was already present for some routes, but not for all of them: we now make it consistent. If you use one of our integrations, it will not impact your code.
New features
- Add Geosearch support: Geosearch allows users to filter and sort results based on geographic criteria. You can refine the precision, distance, and shape of the geographical areas in which the document should be searched. @irevoire
- Support new dataset formats: CSV and NDJSON @MarinPostma
Improvements
- Add
curl
to the Docker image (#1676) @ook - Use tikv-jemallocator instead of jemallocator (#1692) @felixonmars
- Better dependency cache for CI (#1683) @shekhirin
- Unify pest dependency (#1659) @happysalada
- Update dependencies (#1684) @MarinPostma
- Improve CONTRIBUTING.md @curquiza
Fixes
- Fix for ARM binary on RHEL (#1666) @kappa-wingman
- Duplicate IDs when indexing @MarinPostma
- Phrase match error in chinese @ManyTheFish
- Search performance for some queries @ManyTheFish
- M1 Mac binaries available for download via download-latest.sh (#1697) @k-nasa
- Fix minidashboard (#1755) @anirudhRowjee
❤️ Thanks again to our external contributors: @anirudhRowjee, @felixonmars, @happysalada, @kappa-wingman, @k-nasa, and @shekhirin