3866: Update charabia v0.8.0 r=dureuill a=ManyTheFish
# Pull Request
Update Charabia:
- enhance Japanese segmentation
- enhance Latin Tokenization
- words containing `_` are now properly segmented into several words
- brackets `{([])}` are no more considered as context separators so word separated by brackets are now considered near together for the proximity ranking rule
- fixes#3815
- fixes#3778
- fixes [product#151](https://github.com/meilisearch/product/discussions/151)
> Important note: now the float numbers are segmented around the `.` so `3.22` is segmented as [`3`, `.`, `22`] but the middle dot isn't considered as a hard separator, which means that if we search `3.22` we find documents containing `3.22`
Co-authored-by: ManyTheFish <many@meilisearch.com>
3780: Be able to sort facet values by alpha or count r=dureuill a=Kerollmops
This PR introduces a new `sortFacetValuesBy` settings parameter to expose the facet distribution in either count or lexicographic/alpha order.
## Mini Spec of the `sortFacetValuesBy` Settings Parameter
This parameter can be set in the settings to change how the engine returns the facet values. There are two possible values to this parameter.
Please note that the current behavior changed a bit, and keys are returned in lexicographic order instead of undefined order. The previous order wasn't defined as we were using a `HashMap`, which returns entries in hash order (undefined), and we are now using an `IndexMap`, which returns them in insertion order (the order we actually want).
Also, note that there are performance issues when the dataset is enormous. Here are the timings of the engine running on my Macbook Pro M1 (16Go of RAM). [The dataset is 40 million songs file](https://www.notion.so/meilisearch/Songs-from-MusicBrainz-686e31b2bd3845898c7746f502a6e117), and the database size is about 50GiB. Even if you think 800ms is not that high, don't forget that the API is public, and anybody can ask for multiple facets in a single query.
| Search Kind | Get Facets | Max Values per Facet | Time for Alpha | Time for Count | Count but with #3788 |
|------------:|------------|----------------------|:--------------:|----------------|----------------------|
| Placeholder | genres | default (100) | 7ms | 187ms | 122ms |
| Placeholder | genres | 20 | 6ms | 124ms | 75ms |
| Placeholder | album | default (100) | 9ms | 808ms | 677ms |
| Placeholder | album | 20 | 8ms | 579ms | 446ms |
| Placeholder | artist | default (100) | 9ms | 462ms | 344ms |
| Placeholder | artist | 20 | 9ms | 341ms | 246ms |
### Order Values in Alphanumeric Order
This is the default one. Values will be returned by lexicographic order, ascending from A to Z.
```bash
# First, update the settings
curl 'localhost:7700/indexes/movies/settings/facetting' \
-H "Content-Type: application/json" \
-d '{ "sortFacetValuesBy": { "*": "alpha" } }'
# Then, ask for the facet distribution
curl 'localhost:7700/indexes/movies/search?facets=genres'
```
```json5
{
"hits": [
/* list of results */
],
"query": "",
"processingTimeMs": 0,
"limit": 20,
"offset": 0,
"estimatedTotalHits": 1000,
"facetDistribution": {
"genres": {
"Action": 3215,
"Adventure": 1972,
"Animation": 1577,
"Comedy": 5883,
"Crime": 1808,
// ...
}
},
"facetStats": {}
}
```
### Order Values in Count Order
Facet values are sorted by decreasing count. The count is the number of records containing this facet value in the query results.
```bash
# First, update the settings
curl 'localhost:7700/indexes/movies/settings/facetting' \
-H "Content-Type: application/json" \
-d '{ "sortFacetValuesBy": { "*": "count" } }'
# Then, ask for the facet distribution
curl 'localhost:7700/indexes/movies/search?facets=genres'
```
```json5
{
"hits": [
/* list of results */
],
"query": "",
"processingTimeMs": 0,
"limit": 20,
"offset": 0,
"estimatedTotalHits": 1000,
"facetDistribution": {
"genres": {
"Drama": 7337,
"Comedy": 5883,
"Action": 3215,
"Thriller": 3189,
"Romance": 2507,
// ...
}
},
"facetStats": {}
}
```
## Todo List
- [x] Add tests
- [x] Send analytics when a user change the `sortFacetValuesBy`
- [x] Create a prototype and announce it in https://github.com/meilisearch/product/discussions/519.
Co-authored-by: Kerollmops <clement@meilisearch.com>
Co-authored-by: Clément Renault <clement@meilisearch.com>
3864: Remove `/experimental-features` verbs that weren't in the PRD r=dureuill a=dureuill
Removes:
- POST `/experimental-features`
- DELETE `/experimental-features`
keeping only:
- PATCH `/experimental-features`
- GET `/experimental-features`
The two routes that are described in the PRD.
Following `@guimachiavelli's` [question](https://github.com/meilisearch/documentation/issues/2482#issuecomment-1611845372) about the POST route.
Co-authored-by: Louis Dureuil <louis@meilisearch.com>
3699: Search for Facet Values r=Kerollmops a=Kerollmops
This PR introduces the first version of [the _Search for Facet Values_ feature](https://github.com/meilisearch/product/discussions/515) that allows a user to search for facets, by optionally using a prefix string and optionally specifying the `q` and `filter` original search parameters to restrict the candidates to search in.
The steps to merge it into Meilisearch will first start by providing prototype Docker images. This way users will be able to test the prototypes before using them.
The current route to use the _Search for Facet Values_ feature is the `POST /indexes/{index}/facet-search` where the body is a JSON object that looks like the following:
```json5
{
"q": "spiderman", // optional
"filter": "rating > 10", // optional
"facetName": "genres",
"facetQuery": "a" // optional
}
```
## What is missing?
- [x] Send some analytics.
- [x] Support the `matchingStrategy` parameter.
- [x] Make sure that the errors are the right ones.
- [x] Use the [Index typo tolerance settings](https://www.meilisearch.com/docs/learn/configuration/typo_tolerance#minwordsizefortypos) when matching facet values.
- [x] minWordSizeForTypos.oneTypo
- [x] minWordSizeForTypos.twoTypo
- [x] Add tests
- [x] Log the time it took to compute the results.
- [x] Fix the compilation warnings.
- [x] [Create an issue to fix potential performance issues when indexing](https://github.com/meilisearch/meilisearch/issues/3862).
Co-authored-by: Clément Renault <clement@meilisearch.com>
Co-authored-by: Kerollmops <clement@meilisearch.com>
