190: Make bucket candidates optionals r=Kerollmops a=LegendreM
Before the bucket candidates were the result of the facet filters or result of the query tree.
They will now be only the result of the query tree, making the number of candidates more consistent between the same request with or without facet filters.
Fix some clippy warnings.
Fix#186
Co-authored-by: many <maxime@meilisearch.com>
187: Fix fields distribution after documents merge r=Kerollmops a=shekhirin
Resolves https://github.com/meilisearch/milli/issues/174
The problem was with calculation of fields distribution before the merge in `output_from_sorter()`. So if you'd import two documents with the same primary key value, fields distribution will count it as two documents, while `output_from_sorter()` will merge these documents into one.
---
```console
➜ Downloads cat short_movies.json
[
{"id":"47474","title":"The Serpent's Egg","poster":"https://image.tmdb.org/t/p/w500/n7z0doFkXHcvo8QQWHLFnkEPXRU.jpg","overview":"The Serpent's Egg follows a week in the life of Abel Rosenberg, an out-of-work American circus acrobat living in poverty-stricken Berlin following Germany's defeat in World War I.","release_date":246844800,"genres":["Thriller","Drama","Mystery"]},
{"id":"47474","title":"The Serpent's Egg","poster":"https://image.tmdb.org/t/p/w500/n7z0doFkXHcvo8QQWHLFnkEPXRU.jpg","overview":"The Serpent's Egg follows a week in the life of Abel Rosenberg, an out-of-work American circus acrobat living in poverty-stricken Berlin following Germany's defeat in World War I.","release_date":246844800,"genres":["Thriller","Drama","Mystery"]}
]
➜ Downloads curl -X POST -H "Content-Type: text/json" --data-binary @short_movies.json 127.0.0.1:7700/indexes/movies/documents
{"updateId":0}
```
## Before
```console
➜ Downloads curl -s 127.0.0.1:7700/indexes/movies/stats | jq
{
"numberOfDocuments": 1,
"isIndexing": false,
"fieldsDistribution": {
"release_date": 2,
"poster": 2,
"title": 2,
"overview": 2,
"genres": 2,
"id": 2
}
}
```
## After
```console
➜ Downloads curl -s 127.0.0.1:7700/indexes/movies/stats | jq
{
"numberOfDocuments": 1,
"isIndexing": false,
"fieldsDistribution": {
"poster": 1,
"release_date": 1,
"title": 1,
"genres": 1,
"id": 1,
"overview": 1
}
}
```
Co-authored-by: Alexey Shekhirin <a.shekhirin@gmail.com>
183: remove tests on main r=Kerollmops a=MarinPostma
remove testing on main since we now use bors for merging.
Co-authored-by: Marin Postma <postma.marin@protonmail.com>
- pass excluded document to criteria to remove them in higher levels of the bucket-sort
- merge already returned document with excluded documents to avoid duplicas
Related to #125 and #112Fix#170
