(Some of) these specialized tokenizations include huge dictionaries
that currently account for 90% (!) of the meilisearch binary size.
This commit adds chinese, hebrew, japanese, and thai feature flags
that are propagated via milli down to the charabia crate. To keep it
backward compatible, they are enabled by default.
Related to meilisearch/milli#632
2523: Improve the tasks error reporting when processed in batches r=irevoire a=Kerollmops
This fixes#2478 by changing the behavior of the task handler when there is an error in a batch of document addition or update.
What changes is that when there is a user error in a task in a batch we now report this task as failed with the right error message but we continue to process the other tasks. A user error can be when a geo field is invalid, a document id is invalid, or missing.
fixes#2582, #2478
Co-authored-by: Kerollmops <clement@meilisearch.com>
Co-authored-by: ManyTheFish <many@meilisearch.com>
2466: index resolver tests r=MarinPostma a=MarinPostma
add more index resolver tests
depends on #2455
followup #2453
Co-authored-by: ad hoc <postma.marin@protonmail.com>
2414: Improve index uid validation upon API key creation r=Kerollmops a=pierre-l
- ~Use an IndexUid newtype to enforce stronger constraints~
- ~`cargo update -p vergen`~ (`rustup update` was the proper fix for this)
- Add a new `meilisearch_types` crate
- Move `meilisearch_error` to `meilisearch_types::error`
- Move `meilisearch_lib::index_resolver::IndexUid` to `meilisearch_types::index_uid`
- Add a new `InvalidIndexUid` error in `meilisearch_types::index_uid`
- Move `meilisearch_http::routes::StarOr` to `meilisearch_types::star_or`
- Use the `IndexUid` and `StarOr` in `meilisearch_auth::Key`
Fixes#2158
Co-authored-by: pierre-l <pierre.larger@gmail.com>
Move `meilisearch_error` to `meilisearch_types::error`
Move `meilisearch_lib::index_resolver::IndexUid` to `meilisearch_types::index_uid`
Add a new `InvalidIndexUid` error in `meilisearch_types::index_uid`
2494: Introduce the new faceting and pagination settings r=ManyTheFish a=Kerollmops
This PR introduces two new settings following the newly created spec https://github.com/meilisearch/specifications/pull/157:
- The `faceting.max_values_per_facet` one describes the maximum number of values (each with a count) associated with a value in a facet distribution query.
- The `pagination.limited_to` one describes the maximum number of documents that a search query can ever return.
Co-authored-by: Kerollmops <clement@meilisearch.com>
2460: Create custom error types for `TaskType`, `TaskStatus`, and `IndexUid` r=Kerollmops a=walterbm
# Pull Request
## What does this PR do?
Fixes#2443 by making the following changes:
- Add custom `TaskTypeError` for `TaskType::from_str`
- Add custom `TaskStatusError` for `TaskStatus::from_str`
- Add custom `IndexUidFormatError` for `IndexUid::from_str`
- Implement `From<IndexUidFormatError> for IndexResolverError` to convert between errors
- Replace all usages of `IndexUid::new` with `IndexUid::from_str`
- **NOTE** I am relatively new to Rust and I struggled a lot with this final part. This PR ended up with a messy error conversion which does not seem ideal. Please let me know if you have any suggestions for how to make this better and I'll be happy to make any updates!
## PR checklist
Please check if your PR fulfills the following requirements:
- [x] Does this PR fix an existing issue?
- [x] Have you read the contributing guidelines?
- [x] Have you made sure that the title is accurate and descriptive of the changes?
Thank you so much for contributing to Meilisearch!
Co-authored-by: walter <walter.beller.morales@gmail.com>
2473: fix blocking in dumps r=irevoire a=MarinPostma
This PR fixes two blocking calls in the dump process.
Co-authored-by: ad hoc <postma.marin@protonmail.com>
