4938: Remove default embedder r=ManyTheFish a=dureuill
# Pull Request
## Related issue
Fixes#4738
## What does this PR do?
[See public usage](https://meilisearch.notion.site/v1-11-AI-search-changes-0e37727193884a70999f254fa953ce6e#1044b06b651f80edb9d4ef6dc367bad0)
- Remove `hybrid.embedder` boolean from analytics because embedder is now mandatory and so the boolean would always be `true`
- Rework search kind so that a search without query but with vector is a vector search regardless of (non-zero) semantic ratio
Co-authored-by: Louis Dureuil <louis@meilisearch.com>
4888: bring back v1.10.0 into main r=Kerollmops a=ManyTheFish
Co-authored-by: Louis Dureuil <louis@meilisearch.com>
Co-authored-by: meili-bors[bot] <89034592+meili-bors[bot]@users.noreply.github.com>
Co-authored-by: Tamo <tamo@meilisearch.com>
Co-authored-by: ManyTheFish <many@meilisearch.com>
4845: Fix perf regression facet strings r=ManyTheFish a=dureuill
Benchmarks between v1.9 and v1.10 show a performance regression of about x2 (+3dB regression) for most indexing workloads (+44s for hackernews).
[Benchmark interpretation in the engine weekly meeting](https://www.notion.so/meilisearch/Engine-weekly-4d49560d374c4a87b4e3d126a261d4a0?pvs=4#98a709683276450295fcfe1f8ea5cef3).
- Initial investigation pointed to #4819 as the origin of the regression.
- Further investigation points towards the hypernormalization of each facet value in `extract_facet_string_docids`
- Most of the slowdown is in `normalize_facet_strings`, and precisely in `detection.language()`.
This PR improves the situation (-10s compared with `main` for hackernews, so only +34s regression compared with `v1.9`) by skipping normalization when it can be skipped.
I'm not sure how to fix the root cause though. Should we skip facet locale normalization for now? Cc `@ManyTheFish`
---
Tentative resolution options:
1. remove locale normalization from facet. I'm not sure why this is required, I believe we weren't doing this before, so maybe we can stop doing that again.
2. don't do language detection when it can be helped: won't help with the regressions in benchmark, but maybe we can skip language detection when the locales contain only one language?
3. use a faster language detection library: `@Kerollmops` told me about https://github.com/quickwit-oss/whichlang which bolsters x10 to x100 throughput compared with whatlang. Should we consider replacing whatlang with whichlang? Now I understand whichlang supports fewer languages than whatlang, so I also suggest:
4. use whichlang when the list of locales is empty (autodetection), or when it only contains locales that whichlang can detect. If the list of locales contains locales that whichlang *cannot* detect, **then** use whatlang instead.
---
> [!CAUTION]
> this PR contains a commit that adds detailed spans, that were used to detect which part of `extract_facet_string_docids` was taking too much time. As this commit adds spans that are called too often and adds 7s overhead, it should be removed before landing.
Co-authored-by: Louis Dureuil <louis@meilisearch.com>
Co-authored-by: ManyTheFish <many@meilisearch.com>
4846: Add OpenAI tests r=dureuill a=dureuill
# Pull Request
## Related issue
Part of fixing #4757
## What does this PR do?
- OpenAI embedder: don't pass apiKey when it is empty (slightly improves error messages)
- rest embedder and rest-based embedders: specialize the authorization denied error message depending on the configuration source
- fix existing tests
- Adds assets containing prerecorded texts to embed and the embeddings obtained from OpenAI
- Adds an asset containing a tokenized long document and the embedding obtained from OpenAI for this token
- Uses the wiremock crate to mock the OpenAI API: parse the openai request, lookup the response in assets, craft an openai response
Co-authored-by: Louis Dureuil <louis@meilisearch.com>
