The new Chinese segmenter is splitting words in smaller parts.
The words `小化妆包` was previously seegmented as `小 / 化妆包` and is now segmented as `小 / 化妆 / 包`,
which changes the tests results.
Test suite / Tests on ubuntu-22.04 (push) Failing after 12s
Test suite / Tests almost all features (push) Has been skipped
Test suite / Test with Ollama (push) Failing after 8s
Test suite / Test disabled tokenization (push) Has been skipped
Test suite / Run tests in debug (push) Failing after 11s
Test suite / Run Clippy (push) Successful in 7m1s
Test suite / Run Rustfmt (push) Successful in 2m44s
Run the indexing fuzzer / Setup the action (push) Successful in 1h5m40s
Indexing bench (push) / Run and upload benchmarks (push) Has been cancelled
Benchmarks of indexing (push) / Run and upload benchmarks (push) Has been cancelled
Benchmarks of search for geo (push) / Run and upload benchmarks (push) Has been cancelled
Benchmarks of search for songs (push) / Run and upload benchmarks (push) Has been cancelled
Benchmarks of search for Wikipedia articles (push) / Run and upload benchmarks (push) Has been cancelled
Test suite / Tests on macos-13 (push) Has been cancelled
Test suite / Tests on windows-2022 (push) Has been cancelled
5371: Composite embedders r=irevoire a=dureuill
# Pull Request
## Related issue
Fixes#5343
## What does this PR do?
- Implement [public usage](https://www.notion.so/meilisearch/Composite-embedder-usage-14a4b06b651f81859dc3df21e8cd02a0)
- Refactor the way we check if a parameter is mandatory/allowed/disallowed for a given source
- Take the "nesting context" into account for computer if a parameter is mandatory/allowed/disallowed
- Add tests checking all parameters with all sources, and made sure the results didn't change compared with v1.13
## Dumpless Upgrade
- This adds a new value for an existing parameter => compatible without change
- This adds new optional parameters => compatible without change
Co-authored-by: Louis Dureuil <louis@meilisearch.com>
**Changes:**
The transform structure is now relying on FieldIdMapWithMetadata and AttributePatterns to prepare
the obkv documents during a settings reindexing.
The InnerIndexSettingsDiff and InnerIndexSettings structs are now relying on FieldIdMapWithMetadata, FilterableAttributesRule and AttributePatterns to define the field and the databases that should be reindexed.
The faceted_fields_ids, localized_searchable_fields_ids and localized_faceted_fields_ids have been removed in favor of the FieldIdMapWithMetadata.
We are now relying on the FieldIdMapWithMetadata to retain vectors_fids from the facets and the searchables.
The searchable database computing is now relying on the FieldIdMapWithMetadata to know if a field is searchable and retrieve the locales.
The facet database computing is now relying on the FieldIdMapWithMetadata to compute the facet databases, the facet-search and retrieve the locales.
The facet level database computing is now relying on the FieldIdMapWithMetadata and the facet level database are cleared depending on the settings differences (clear_facet_levels_based_on_settings_diff).
The vector point extraction uses the FieldIdMapWithMetadata instead of FieldsIdsMapWithMetadata.
**Impact:**
- Dump import
- Settings update
Look for flaky tests / flaky (push) Failing after 1s
Indexing bench (push) / Run and upload benchmarks (push) Has been cancelled
Benchmarks of indexing (push) / Run and upload benchmarks (push) Has been cancelled
Benchmarks of search for geo (push) / Run and upload benchmarks (push) Has been cancelled
Benchmarks of search for songs (push) / Run and upload benchmarks (push) Has been cancelled
Benchmarks of search for Wikipedia articles (push) / Run and upload benchmarks (push) Has been cancelled
Publish binaries to GitHub release / Check the version validity (push) Failing after 5s
Test suite / Tests almost all features (push) Failing after 13s
Test suite / Tests on ubuntu-22.04 (push) Failing after 19s
Test suite / Test with Ollama (push) Failing after 7s
Test suite / Test disabled tokenization (push) Failing after 10s
Test suite / Run tests in debug (push) Failing after 15s
Test suite / Run Rustfmt (push) Failing after 16s
Test suite / Run Clippy (push) Successful in 9m39s
SDKs tests / define-docker-image (push) Failing after 5s
SDKs tests / .NET SDK tests (push) Has been skipped
SDKs tests / Dart SDK tests (push) Has been skipped
SDKs tests / Go SDK tests (push) Has been skipped
SDKs tests / Java SDK tests (push) Has been skipped
SDKs tests / JS SDK tests (push) Has been skipped
SDKs tests / PHP SDK tests (push) Has been skipped
SDKs tests / Python SDK tests (push) Has been skipped
SDKs tests / Ruby SDK tests (push) Has been skipped
SDKs tests / Rust SDK tests (push) Has been skipped
SDKs tests / Swift SDK tests (push) Has been skipped
SDKs tests / meilisearch-js-plugins tests (push) Has been skipped
SDKs tests / meilisearch-rails tests (push) Has been skipped
SDKs tests / meilisearch-symfony tests (push) Has been skipped
Test suite / Tests on macos-13 (push) Has been cancelled
Test suite / Tests on windows-2022 (push) Has been cancelled
5355: Support fetching the pooling method from the model configuration r=Kerollmops a=dureuill
# Pull Request
## Related issue
Fixes#5354
## What does this PR do?
- Fetches the pooling configuration from the model repository
- Use a pooling method that depends on the pooling configuration of that model.
- Allow overriding the pooling method with a new huggingFace embedder parameter `pooling`
- for backward-compatibility with Meilisearch v1.13
- for compatibility with embedders that exhibit the same behavior as Meilisearch v1.13
- Handle the default value of that new parameter
- for compatibility, when importing a db/a dump, it should be set to `forceMean`
- when (re)set from the settings for an embedder, it should be set to `useModel`
Co-authored-by: Louis Dureuil <louis@meilisearch.com>
