3781: Revert "Improve docker cache" r=Kerollmops a=curquiza
Reverts meilisearch/meilisearch#3566 because does not work as expected, and so I want to remove useless complexity from the CI and Dockerfile
Co-authored-by: Clémentine U. - curqui <clementine@meilisearch.com>
3779: Add a cron test with disabled tokenization (with @roy9495) r=Kerollmops a=curquiza
Replaces https://github.com/meilisearch/meilisearch/pull/3746 because of bors issue
Co-authored-by: TATHAGATA ROY <98920199+roy9495@users.noreply.github.com>
Co-authored-by: Clémentine U. - curqui <clementine@meilisearch.com>
3775: Last error code changes on the new get/delete documents routes r=dureuill a=irevoire
# Pull Request
## Related issue
Fixes#3774
## What does this PR do?
Following the specification: https://github.com/meilisearch/specifications/pull/236
1. Get rid of the `invalid_document_delete_filter` and always use the `invalid_document_filter`
2. Introduce a new `missing_document_filter` instead of returning `invalid_document_delete_filter` (that’s consistent with all the other routes that have a mandatory parameter)
3. Always return the `original_filter` in the details (potentially set to `null`) instead of hiding it if it wasn’t used
Co-authored-by: Tamo <tamo@meilisearch.com>
3768: Fix bugs in graph-based ranking rules + make `words` a graph-based ranking rule r=dureuill a=loiclec
This PR contains three changes:
## 1. Don't call the `words` ranking rule if the term matching strategy is `All`
This is because the purpose of `words` is only to remove nodes from the query graph. It would never do any useful work when the matching strategy was `All`. Remember that the universe was already computed before by computing all the docids corresponding to the "maximally reduced" query graph, which, in the case of `All`, is equal to the original graph.
## 2. The `words` ranking rule is replaced by a graph-based ranking rule.
This is for three reasons:
1. **performance**: graph-based ranking rules benefit from a lot of optimisations by default, which ensures that they are never too slow. The previous implementation of `words` could call `compute_query_graph_docids` many times if some words had to be removed from the query, which would be quite expensive. I was especially worried about its performance in cases where it is placed right after the `sort` ranking rule. Furthermore, `compute_query_graph_docids` would clone a lot of bitmaps many times unnecessarily.
2. **consistency**: every other ranking rule (except `sort`) is graph-based. It makes sense to implement `words` like that as well. It will automatically benefit from all the features, optimisations, and bug fixes that all the other ranking rules get.
3. **surfacing bugs**: as the first ranking rule to be called (most of the time), I'd like `words` to behave the same as the other ranking rules so that we can quickly detect bugs in our graph algorithms. This actually already happened, which is why this PR also contains a bug fix.
## 3. Fix the `update_all_costs_before_nodes` function
It is a bit difficult to explain what was wrong, but I'll try. The bug happened when we had graphs like:
<img width="730" alt="Screenshot 2023-05-16 at 10 58 57" src="https://github.com/meilisearch/meilisearch/assets/6040237/40db1a68-d852-4e89-99d5-0d65757242a7">
and we gave the node `is` as argument.
Then, we'd walk backwards from the node breadth-first. We'd update the costs of:
1. `sun`
2. `thesun`
3. `start`
4. `the`
which is an incorrect order. The correct order is:
1. `sun`
2. `thesun`
3. `the`
4. `start`
That is, we can only update the cost of a node when all of its successors have either already been visited or were not affected by the update to the node passed as argument. To solve this bug, I factored out the graph-traversal logic into a `traverse_breadth_first_backward` function.
Co-authored-by: Loïc Lecrenier <loic.lecrenier@me.com>
Co-authored-by: Louis Dureuil <louis@meilisearch.com>
3757: Adjust the cost of edges in the `position` ranking rule by bucketing positions more aggressively r=loiclec a=loiclec
This PR significantly improves the performance of the `position` ranking rule when:
1. a query contains many words
2. the `position` ranking rule needs to be called many times
3. the score of the documents according to `position` is high
These conditions greatly increase:
1. the number of edge traversals that are needed to find a valid path from the `start` node to the `end` node
2. the number of edges that need to be deleted from the graph, and therefore the number of times that we need to recompute all the possible costs from START to END
As a result, a majority of the search time is spent in `visit_condition`, `visit_node`, and `update_all_costs_before_node`. This is frustrating because it often happens when the "universe" given to the rule consists of only a handful of document ids.
By limiting the number of possible edges between two nodes from `20` to `10`, we:
1. reduce the number of possible costs from START to END
2. reduce the number of edges that will be deleted
3. make it faster to update the costs after deleting an edge
4. reduce the number of buckets that need to be computed
In terms of relevancy, I don't think we lose or gain much. We still prefer terms that are in a lower positions, with decreasing precision as we go further. The previous choice of bucketing wasn't chosen in a principled way, and neither is this one. They both "feel" right to me.
Co-authored-by: Loïc Lecrenier <loic.lecrenier@me.com>
Co-authored-by: meili-bors[bot] <89034592+meili-bors[bot]@users.noreply.github.com>
3738: Add analytics on the get documents resource r=dureuill a=irevoire
# Pull Request
## Related issue
Fixes https://github.com/meilisearch/meilisearch/issues/3737
Related spec https://github.com/meilisearch/specifications/pull/234
## What does this PR do?
Add the analytics for the following routes:
- `GET` - `/indexes/:uid/documents`
- `GET` - `/indexes/:uid/documents/:doc_id`
- `POST` - `/indexes/:uid/documents/fetch`
These analytics are aggregated between two events:
- `Documents Fetched GET`
- `Documents Fetched POST`
That shares the same payload:
Property name | Description | Example |
|---------------|-------------|---------|
| `requests.total_received` | Total number of request received in this batch | 325 |
| `per_document_id` | `false` | false |
| `per_filter` | `true` if `POST /indexes/:indexUid/documents/fetch` endpoint was used with a filter in this batch, otherwise `false` | false |
| `pagination.max_limit` | Highest value given for the `limit` parameter in this batch | 60 |
| `pagination.max_offset` | Highest value given for the `offset` parameter in this batch | 1000 |
Co-authored-by: Tamo <tamo@meilisearch.com>
3759: Invalid error code when parsing filters r=dureuill a=irevoire
# Pull Request
## Related issue
Fixes https://github.com/meilisearch/meilisearch/issues/3753
## What does this PR do?
Fix the error code in case the error comes from the evaluate of the filter for the get, fetch and delete documents routes.
Co-authored-by: Tamo <tamo@meilisearch.com>
3755: Re-add final dot r=curquiza a=ManyTheFish
I removed the final dot of the error message in my last PR, this one re-adds it.
related to https://github.com/meilisearch/meilisearch/pull/3749
> Oups 😬
Co-authored-by: ManyTheFish <many@meilisearch.com>
3741: Add ngram support to the highlighter r=ManyTheFish a=loiclec
This PR fixes a bug introduced by the search refactor, where ngrams were not highlighted.
The solution was to add the ngrams to the vector of `LocatedQueryTerm` that is given to the `MatchingWords` structure.
Co-authored-by: Loïc Lecrenier <loic.lecrenier@me.com>
3749: Fix back: sort error message r=ManyTheFish a=ManyTheFish
This PR reintroduces the error message modified in https://github.com/meilisearch/milli/pull/375.
However, this added double-quotes around `sort` in the message. I don't think another message contains double-quotes, so I have added a separate commit replacing the double-quotes with back-ticks, which seems more consistent with the other error messages, this last change can be reverted easily.
## Detailed changes
#### v1.2-rc0
```
The sort ranking rule must be specified in the ranking rules settings to use the sort parameter at search time.
```
#### [Reintroduce fix (previous and expected behavior)](23d1c86825)
```
You must specify where "sort" is listed in the rankingRules setting to use the sort parameter at search time
```
#### [Replace double-quotes with back-ticks (my suggestion)](4d691d071a)
```
You must specify where `sort` is listed in the rankingRules setting to use the sort parameter at search time
```
## Related
Fixes#3722
## Reviewers
- technical review: `@irevoire`
- to validate the replacement: `@macraig`
Co-authored-by: ManyTheFish <many@meilisearch.com>
3651: Use the writemap flag to reduce the memory usage r=irevoire a=Kerollmops
This draft PR is showing some stats about the memory usage of Meilisearch when [the LMDB `MDB_WRITEMAP` flag](3947014aed/libraries/liblmdb/lmdb.h (L573-L581)) is enabled and when it is not. As you can see there is a reduction of about 50% of the memory usage pick. The dataset used was [the Wikipedia one](https://www.notion.so/meilisearch/Wikipedia-8b1486e4b17547c5bda485d2d97767a0) with the first 30 000 first CSV documents without settings. This PR depends on https://github.com/meilisearch/heed/pull/168.
I just [opened a discussion](https://github.com/meilisearch/product/discussions/652) for people to understand the tradeoffs and give their feedback.
- [x] Create an experiment flag `--experimental-reduce-indexing-memory-usage`.
- [x] Add it to the config file.
- [x] Explain the tradeoff and copy/link the LMDB documentation in the help message.
- [x] Add analytics about the experimental flag.
- [x] Document that this flag cannot be used on Windows, ~~or hide it~~.
<details>
<summary>The command I used to run the tests</summary>
#### Sign the binary to be able to use Instruments / xcrun
```sh
codesign -s - -f --entitlements ~/ent.plist target/release/meilisearch
```
where `ent.plist` contains:
```xml
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
<key>com.apple.security.get-task-allow</key>
<true/>
</dict>
</plist>
```
#### Run Meilisearch in measure-mode
```sh
xcrun xctrace record --template 'Allocations' --launch -- target/release/meilisearch --max-indexing-memory 0MiB
```
#### Send the wiki dataset available on notion.so / Public
```sh
for f in 0.csv 15000.csv; do echo sending $f; xh 'localhost:7700/indexes/wiki/documents' 'content-type:text/csv' `@$f;` done
```
#### Wait for the task to finish
```sh
watch --color xh --pretty all 'localhost:7700/tasks?statuses=processing'
```
</details>
Keep in mind that I tested that with the Instruments Apple tools on an iMac 5k 2019. More benchmarks must be done, especially on the indexation speed, as the flag is told to slow down writing into databases bigger that the amount of memory.
On the left Meilisearch is running without the flag. On the right, it is running with the flag.
<p align="center">
<img align="left" width="45%" alt="Instrument showing the memory usage of Meilisearch without the MDB_WRITEMAP flag" src="https://user-images.githubusercontent.com/3610253/234299524-7607f1df-6fc1-45d3-bd3d-4f9388002857.png">
<img align="right" width="45%" alt="Instrument showing the memory usage of Meilisearch with the MDB_WRITEMAP flag" src="https://user-images.githubusercontent.com/3610253/234299534-6cc3ae58-8bd9-426c-aa79-4c78f9e88b94.png">
</p>
Co-authored-by: Kerollmops <clement@meilisearch.com>
Co-authored-by: Clément Renault <clement@meilisearch.com>
