3215: Use nightly in cargo fmt r=curquiza a=curquiza
Discussed with `@Kerollmops,` needs this change
Co-authored-by: Clémentine Urquizar - curqui <clementine@meilisearch.com>
729: Fix distincted exhaustive hits r=Kerollmops a=ManyTheFish
This PR changes the name and behavior of `bucket_candidates`:
- `bucket_candidates` become `initial_candidates` that is less confusing
- `initial_candidates` is no more a simple `RoaringBitmap` but an enum allowing us to precise if the candidates are exhaustive or not
- this enum ensures that any modification is allowed only if the candidates are not already exhaustive.
The bug occurred because `initial_candidates` are modified during the bucket sort allowing the estimation to be more and more precise along the search, and this was an issue when the `initial_candidates` were already exhaustive, now, if candidates are exhaustive, then no modifications are made.
Co-authored-by: ManyTheFish <many@meilisearch.com>
727: Fix bug in filter search r=Kerollmops a=loiclec
# Pull Request
## Related issue
Fixes (partially, until merged into meilisearch) https://github.com/meilisearch/meilisearch/issues/3178
## What does this PR do?
The most important change is this one:
```rust
// in milli/src/search/facet/facet_range_search.rs, line 239
let should_stop = {
match self.right {
Bound::Included(right) => right < previous_key.left_bound,
Bound::Excluded(right) => right <= previous_key.left_bound,
Bound::Unbounded => false,
}
};
```
where the operations `<` and `<=` between the two branches were switched. This caused (very few) documents to be missing from filter results.
The second change is a simplification of the algorithm for filters such as `field = value`, where we now perform a direct query into the "Level 0" of the facet db to retrieve the docids instead of invoking the full facet search algorithm. This change is done in `milli/src/search/facet/filter.rs`.
I have added yet more insta-snapshot tests, rechecked the content of the snapshots, and added some integration tests as well.
This is purely a fix in the search algorithms. Based on this PR alone, a dump will not be necessary to switch from v0.30.1 (where this bug is present) to v0.30.2 (where this PR is merged).
Co-authored-by: Loïc Lecrenier <loic.lecrenier@me.com>
By creating snapshots and updating the format of the existing
snapshots. The next commit will apply the fix, which will show
its effects cleanly on the old and new snapshot tests
3208: Stop snapshotting the version of meilisearch in the dump r=Kerollmops a=irevoire
It might change, and we don't want to update this test every time we make a new release.
Co-authored-by: Tamo <tamo@meilisearch.com>
3207: Add release check when starting latest CI r=Kerollmops a=curquiza
Adding this to have the same kind of check before starting to move the latest tag
<img width="737" alt="Capture d’écran 2022-12-07 à 12 18 33" src="https://user-images.githubusercontent.com/20380692/206165868-18a2be7c-78ec-48c9-acb9-d7f60797c2e3.png">
Also, removing an un-unused script
Co-authored-by: curquiza <clementine@meilisearch.com>
3128: Bumps cargo_toml version to most up to date r=curquiza a=colbsmcdolbs
# Pull Request
## Related issue
Fixes#3127
## What does this PR do?
- The README of this repository declares that one package is not up to date. In order to ensure Due Diligence, I have bumped the version number of the package. No test failures running on Windows.
## PR checklist
Please check if your PR fulfills the following requirements:
- [X] Does this PR fix an existing issue, or have you listed the changes applied in the PR description (and why they are needed)?
- [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: Colby Allen <colbyjayallen@gmail.com>
723: Fix bug in handling of soft deleted documents when updating settings r=Kerollmops a=loiclec
# Pull Request
## Related issue
Fixes (partially, until merged into meilisearch) https://github.com/meilisearch/meilisearch/issues/3021
## What does this PR do?
This PR fixes the bug where a `missing key in documents database` internal error message could appear when indexing documents.
When updating the settings, before clearing the database and before creating the transform output, we now modify the `ExternalDocumentsIds` structure to get rid of all references to soft deleted document ids in its FSTs.
It used to be that updating the settings would clear the soft-deleted document ids, but keep the original `ExternalDocumentsIds` structure. As a consequence of this, when processing a future document addition, we could wrongly believe that a document was being replaced when, in fact, it was a completely new document. See the tests `bug_3021_first`, `bug_3021_second`, and `bug_3021` for a minimal test case that would have reproduced the issue.
We need to take special care to:
- evaluate how users should update to v0.30.1 (containing this fix): dump? reimporting all documents from scratch?
- understand IF/HOW this bug could have caused duplicate documents to be returned
- and evaluate the correctness of the fix, of course :)
Co-authored-by: Loïc Lecrenier <loic.lecrenier@me.com>
3190: Fix the dump date-import of the dumpv4 r=irevoire a=irevoire
# Pull Request
After merging https://github.com/meilisearch/meilisearch/pull/3012 I realized that the tests on the date of the dump-v4 were still ignored, thus, I fixed them and then noticed #3012 wasn't working properly.
## Related issue
Fixes https://github.com/meilisearch/meilisearch/issues/2987 a second time
`@funilrys` since you wrote most of the code you might be interested, but don't feel obligated to review this code.
Someone from the team will double-check it works 😁
Co-authored-by: Tamo <tamo@meilisearch.com>
719: Add more members of `filter_parser` to `milli::` & `From<&str>` implementation for `Token` r=Kerollmops a=GregoryConrad
## What does this PR do?
The current `milli::Filter` and `milli::FilterCondition` APIs require working with some members of `filter_parser` directly that `milli::` does *not* re-export to its users (at least when not parsing input using `parse`). Also, using `filter_parser` does not make sense when using milli from an embedded context where there is no query to parse.
Instead of reworking `milli::Filter` and `milli::FilterCondition`, this PR adds two non-breaking changes that ease the use of milli:
- Re-exports more members of the dependent version of `filter_parser` in `milli`
- Implements `From<&str>` for `filter_parser::Token`
- This will also allow some basic tests that need to create a `Token` from a string to avoid some boilerplate.
In conjunction, both of these will allow milli users to easily create a `Token` from a `&str` without needing to add `filter_parser` as an extra dependency.
Note: I wanted to use `FromStr` for the `From` implementation; however, it requires returning a `Result` which is not needed for the conversion. Thus, I just left it as `From<&str>`.
Co-authored-by: Gregory Conrad <gregorysconrad@gmail.com>
3197: Revert "Upgrade alpine 3.16 to 3.17" r=irevoire a=curquiza
Reverts meilisearch/meilisearch#3189
Because `rust:alpine3.17` does not exist, and our scheduled CI failed: https://github.com/meilisearch/meilisearch/actions/runs/3626327181
`@ivanionut` for your information, I'm sorry I should have better checked before accepting the PR, this is my bad
Co-authored-by: Clémentine Urquizar - curqui <clementine@meilisearch.com>
3175: Rename dump command from --dumps-dir to --dump-dir r=dureuill a=dureuill
# Pull Request
## Related issue
Fixes#3132
## What does this PR do?
- Rename the dump commands, env variables and default config
## PR checklist
Please check if your PR fulfills the following requirements:
- [x] Does this PR fix an existing issue, or have you listed the changes applied in the PR description (and why they are needed)?
- [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: Louis Dureuil <louis@meilisearch.com>
722: Geosearch for zero radius r=irevoire a=amab8901
# Pull Request
## Related issue
Fixes#3167 (https://github.com/meilisearch/meilisearch/issues/3167)
## What does this PR do?
- allows Geosearch with zero radius to return the specified location when the coordinates match perfectly (instead of returning nothing). See link for more details.
- new attempt on https://github.com/meilisearch/milli/pull/713
## PR checklist
Please check if your PR fulfills the following requirements:
- [ X ] Does this PR fix an existing issue, or have you listed the changes applied in the PR description (and why they are needed)?
- [ 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: amab8901 <amab8901@protonmail.com>
Co-authored-by: Tamo <irevoire@protonmail.ch>
720: Make soft deletion optional in document addition and deletion + add lots of tests r=irevoire a=loiclec
# Pull Request
## What does this PR do?
When debugging recent issues, I created a few unit tests in the hopes reproducing the bugs I was looking for. In the end, I didn't find any, but I thought it would still be good to keep those tests.
More importantly, I added a field to the `DeleteDocuments` and `IndexDocuments` builders, called `disable_soft_deletion`. If set to `true`, the indexing/deletion will never add documents to the `soft_deleted_documents_ids` and instead perform a real deletion of the documents from the databases.
For the new tests, I have:
- Improved the insta-snapshot format of the `external_documents_ids` structure
- Added more tests for the facet DB indexing, deletion, and search algorithms, making sure to test them when the facet DB contains strings (instead of numbers) as well.
- Added more tests for the incremental indexing of the prefix proximity databases. For example, to see if documents are replaced correctly and if common prefixes are deleted correctly.
- Added tests that mix soft deletion and hard deletion, including when processing batches of document updates.
Co-authored-by: Loïc Lecrenier <loic.lecrenier@me.com>
