Commit Graph

1519 Commits

Author SHA1 Message Date
Louis Dureuil
c621a250a7
Score for graph based ranking rules
Count phrases in matchingWords and maxMatchingWords
2023-06-22 12:39:14 +02:00
Louis Dureuil
8939e85f60
Add rank_to_score for graph based ranking rules 2023-06-22 12:39:14 +02:00
Louis Dureuil
fa41d2489e
Score for sort 2023-06-22 12:39:14 +02:00
Louis Dureuil
59c5b992c2
Score for geosort 2023-06-22 12:39:14 +02:00
Louis Dureuil
2ea8194c18
Score for exact_attributes 2023-06-22 12:39:14 +02:00
Louis Dureuil
421df64602
RankingRuleOutput now contains a Score 2023-06-22 12:39:14 +02:00
Louis Dureuil
c0fca6f884
Add score_details 2023-06-22 12:39:14 +02:00
Louis Dureuil
f050634b1e
add virtual conditions to fid and position to always have the max cost 2023-06-20 10:07:18 +02:00
Louis Dureuil
becf1f066a
Change how the cost of removing words is computed 2023-06-20 09:45:43 +02:00
Louis Dureuil
701d299369
Remove out-of-date comment 2023-06-20 09:45:42 +02:00
Louis Dureuil
a20e4d447c
Position now takes into account the distance to the position of the word in the query
it used to be based on the distance to the position 0
2023-06-20 09:45:42 +02:00
Louis Dureuil
af57c3c577
Proximity costs 0 for documents that are perfectly matching 2023-06-20 09:45:42 +02:00
Louis Dureuil
0c40ef6911
Fix sort id 2023-06-20 09:45:42 +02:00
meili-bors[bot]
45636d315c
Merge #3670
3670: Fix addition deletion bug r=irevoire a=irevoire

The first commit of this PR is a revert of https://github.com/meilisearch/meilisearch/pull/3667. It re-enable the auto-batching of addition and deletion of tasks. No new changes have been introduced outside of `milli`. So all the changes you see on the autobatcher have actually already been reviewed.

It fixes https://github.com/meilisearch/meilisearch/issues/3440.

### What was happening?

The issue was that the `external_documents_ids` generated in the `transform` were used in a very strange way that wasn’t compatible with the deletion of documents.
Instead of doing a clear merge between the external document IDs of the DB and the one returned by the transform + writing it on disk, we were doing some weird tricks with the soft-deleted to avoid writing the fst on disk as much as possible.
The new algorithm may be a bit slower but is way more straightforward and doesn’t change depending on if the soft deletion was used or not. Here is a list of the changes introduced:
1. We now do a clear distinction between the `new_external_documents_ids` coming from the transform and only held on RAM and the `external_documents_ids` coming from the DB.
2. The `new_external_documents_ids` (coming out of the transform) are now represented as an `fst`. We don't need to struggle with the hard, soft distinction + the soft_deleted => That's easier to understand
3. When indexing documents, we merge the `external_documents_ids` coming from the DB and the `new_external_documents_ids` coming from the transform.

### Other things introduced in this  PR

Since we constantly have to write small, very specialized fuzzers for this kind of bug, we decided to push the one used to reproduce this bug.
It's not perfect, but it's easy to improve in the future.
It'll also run for as long as possible on every merge on the main branch.

Co-authored-by: Tamo <tamo@meilisearch.com>
Co-authored-by: Loïc Lecrenier <loic.lecrenier@icloud.com>
2023-06-19 09:09:30 +00:00
meili-bors[bot]
cb9d78fc7f
Merge #3835
3835: Add more documentation to graph-based ranking rule algorithms + comment cleanup r=Kerollmops a=loiclec

In addition to documenting the `cheapest_path.rs` file, this PR cleans up a few outdated comments as well as some TODOs. These TODOs have been moved to https://github.com/meilisearch/meilisearch/issues/3776



Co-authored-by: Loïc Lecrenier <loic.lecrenier@icloud.com>
2023-06-15 15:30:24 +00:00
Louis Dureuil
e0c4682758
Fix tests 2023-06-14 13:30:52 +02:00
Louis Dureuil
d9b4b39922
Add trailing pipe to the snapshots so it doesn't end with trailing whitespace 2023-06-14 13:30:52 +02:00
Loïc Lecrenier
2da86b31a6 Remove comments and add documentation 2023-06-14 12:39:42 +02:00
Louis Dureuil
a2a3b8c973
Fix offset difference between query and indexing for hard separators 2023-06-08 12:07:12 +02:00
Louis Dureuil
9f37b61666
DB BREAKING: raise limit of word count from 10 to 30. 2023-06-08 12:07:12 +02:00
Louis Dureuil
c15c076da9
DB BREAKING: Count the number of words in field_id_word_count_docids 2023-06-08 12:07:11 +02:00
Loïc Lecrenier
8628a0c856 Remove docid_word_positions_db + fix deletion bug
That would happen when a word was deleted from all exact attributes
but not all regular attributes.
2023-06-07 10:52:50 +02:00
Clémentine U. - curqui
f3e2f79290
Merge branch 'main' into tmp-release-v1.2.0 2023-06-05 18:36:28 +02:00
Kerollmops
da04edff8c
Better use deserialize_unchecked_from to reduce the deserialization time 2023-05-30 14:58:30 +02:00
Louis Dureuil
1dfc4038ab
Add test that fails before PR and passes now 2023-05-29 11:58:26 +02:00
Louis Dureuil
73198179f1
Consistently use wrapping add to avoid overflow in debug when query starts with a separator 2023-05-29 11:54:12 +02:00
meili-bors[bot]
2e49d6aec1
Merge #3768
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>
2023-05-23 13:28:08 +00:00
Louis Dureuil
51043f78f0
Remove trailing whitespace 2023-05-23 15:27:25 +02:00
Louis Dureuil
a490a11325
Add explanatory comment on the way we're recomputing costs 2023-05-23 15:24:24 +02:00
Tamo
602ad98cb8 improve the way we handle the fsts 2023-05-22 11:15:14 +02:00
Tamo
7f619ff0e4 get rids of the now unused soft_deletion_used parameter 2023-05-22 10:33:49 +02:00
Tamo
4391cba6ca
fix the addition + deletion bug 2023-05-17 18:28:57 +02:00
meili-bors[bot]
101f5a20d2
Merge #3757
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>
2023-05-17 11:43:59 +00:00
Loïc Lecrenier
ec8f685d84 Fix bug in cheapest path algorithm 2023-05-16 17:01:30 +02:00
Loïc Lecrenier
5758268866 Don't compute split_words for phrases 2023-05-16 17:01:18 +02:00
Loïc Lecrenier
3e19702de6 Update snapshot tests 2023-05-16 12:22:46 +02:00
meili-bors[bot]
1e762d151f
Merge #3755
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>
2023-05-16 10:10:58 +00:00
Loïc Lecrenier
f6524a6858 Adjust costs of edges in position ranking rule
To ensure good performance
2023-05-16 11:28:56 +02:00
meili-bors[bot]
65ad8cce36
Merge #3741
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>
2023-05-16 09:03:31 +00:00
ManyTheFish
42650f82e8 Re-add final dot 2023-05-16 10:57:26 +02:00
Loïc Lecrenier
a37da36766 Implement words as a graph-based ranking rule and fix some bugs 2023-05-16 10:42:11 +02:00
Loïc Lecrenier
85d96d35a8 Highlight ngram matches as well 2023-05-16 10:39:36 +02:00
meili-bors[bot]
bf66e97b48
Merge #3749
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>
2023-05-15 14:55:51 +00:00
Kerollmops
f759ec7fad
Expose a flag to enable the MDB_WRITEMAP flag 2023-05-15 11:38:43 +02:00
ManyTheFish
4d691d071a Change double-quotes by back-ticks in sort error message 2023-05-15 11:10:36 +02:00
ManyTheFish
23d1c86825 Re-introduce the sort error message fix 2023-05-15 11:07:23 +02:00
Kerollmops
c4a40e7110
Use the writemap flag to reduce the memory usage 2023-05-15 10:15:33 +02:00
Loïc Lecrenier
4d352a21ac Compute split words derivations of terms that don't accept typos 2023-05-10 13:31:19 +02:00
Loïc Lecrenier
3625389057 Highlight ngram matches as well 2023-05-08 15:35:41 +02:00
meili-bors[bot]
eace6df91b
Merge #3726
3726: Fix prefix highlighting r=loiclec a=ManyTheFish

The prefix queries were not properly highlighted, this PR now highlights only the start of a word when it matched with a prefix

Co-authored-by: ManyTheFish <many@meilisearch.com>
Co-authored-by: Loïc Lecrenier <loic.lecrenier@me.com>
2023-05-08 07:46:46 +00:00
Loïc Lecrenier
83ab8cf4e5 Remove dbg!(..) expression in highlighter tests 2023-05-08 09:45:23 +02:00
ManyTheFish
cd2573fcc3 Fix prefix highlighting 2023-05-04 16:53:50 +02:00
meili-bors[bot]
9f7981df28
Merge #3687
3687: Allow to disable specialized tokenizations (again) r=Kerollmops a=jirutka

In PR #2773, I added the `chinese`, `hebrew`, `japanese` and `thai` feature flags to allow melisearch to be built without huge specialed tokenizations that took up 90% of the melisearch binary size. Unfortunately, due to some recent changes, this doesn't work anymore. The problem lies in excessive use of the `default` feature flag, which infects the dependency graph.

Instead of adding `default-features = false` here and there, it's easier and more future-proof to not declare `default` in `milli` and `meilisearch-types`. I've renamed it to `all-tokenizers`, which also makes it a bit clearer what it's about.


Co-authored-by: Jakub Jirutka <jakub@jirutka.cz>
2023-05-04 14:48:01 +00:00
Jakub Jirutka
e615fa5ec6 Fix unused_imports warning in milli when japanese is not enabled 2023-05-04 15:46:11 +02:00
Jakub Jirutka
13f1277637 Allow to disable specialized tokenizations (again)
In PR #2773, I added the `chinese`, `hebrew`, `japanese` and `thai`
feature flags to allow melisearch to be built without huge specialed
tokenizations that took up 90% of the melisearch binary size.
Unfortunately, due to some recent changes, this doesn't work anymore.
The problem lies in excessive use of the `default` feature flag, which
infects the dependency graph.

Instead of adding `default-features = false` here and there, it's easier
and more future-proof to not declare `default` in `milli` and
`meilisearch-types`. I've renamed it to `all-tokenizers`, which also
makes it a bit clearer what it's about.
2023-05-04 15:45:40 +02:00
Louis Dureuil
a35d3fc708
Add Index::iter_documents 2023-05-04 15:31:54 +02:00
Louis Dureuil
732c52093d
Processing time without autobatching implementation 2023-05-03 17:41:48 +02:00
Louis Dureuil
f8f190cd40
Update exactness tests following charabia camelCase tokenization 2023-05-03 14:45:09 +02:00
Louis Dureuil
3a408e8287
Increase map size for tests following charabia camelCase tokenization 2023-05-03 14:44:48 +02:00
Louis Dureuil
d3e5b10e23
fix nb of dbs 2023-05-03 14:11:20 +02:00
Louis Dureuil
1aaf24ccbf
Cargo fmt 2023-05-03 12:21:58 +02:00
Louis Dureuil
90bc230820
Merge remote-tracking branch 'origin/main' into search-refactor
Conflicts | resolution
----------|-----------
Cargo.lock | added mimalloc
Cargo.toml |  took origin/main version
milli/src/search/criteria/exactness.rs | deleted after checking it was only clippy changes
milli/src/search/query_tree.rs | deleted after checking it was only clippy changes
2023-05-03 12:19:06 +02:00
Louis Dureuil
342c4ff85d
geosort: Remove rtree unwrap 2023-05-03 09:52:16 +02:00
Tamo
c85392ce40
make the descendent geosort fast 2023-05-03 09:13:12 +02:00
Tamo
8875d24a48
deserialize the rtree only when its needed, and keep it in memory once it has been deserialized 2023-05-03 09:13:12 +02:00
Tamo
c470b67fa2
revamp the test to use execute_iterative_and_rtree_returns_the_same 2023-05-03 09:13:12 +02:00
Louis Dureuil
b60840ebff
Remove self.iterating from words 2023-05-02 18:54:23 +02:00
Louis Dureuil
fdc1763838
Use MultiOps for resolve_query_graph 2023-05-02 18:54:09 +02:00
Louis Dureuil
75819bc940
Remove too many arguments on resolve_maximally_reduced_query_graph 2023-05-02 18:53:40 +02:00
Louis Dureuil
7b8cc25625
rename located_query_terms_from_string -> located_query_terms_from_tokens 2023-05-02 18:53:01 +02:00
Loïc Lecrenier
aa63091752 Fix bug in exact_attribute 2023-05-02 10:48:32 +02:00
Loïc Lecrenier
1b514517f5 Fix bug in computation of query term at a position 2023-05-02 10:48:32 +02:00
Loïc Lecrenier
11f814821d Minor cleanup 2023-05-02 10:48:32 +02:00
Loïc Lecrenier
30fb1153cc Speed up graph based ranking rule when a lot of different costs exist 2023-05-02 09:59:42 +02:00
Loïc Lecrenier
3b2c8b9f25 Improve performance of position rr 2023-05-02 09:59:42 +02:00
Loïc Lecrenier
2a7f9adf78 Build query graph more correctly from paths
Update snapshots
2023-05-02 09:59:42 +02:00
Loïc Lecrenier
608ceea440 Fix bug in position rr 2023-05-02 09:59:42 +02:00
Loïc Lecrenier
79001b9c97 Improve performance of the cheapest path finder algorithm 2023-05-02 09:59:42 +02:00
Loïc Lecrenier
59b12fca87 Fix errors, clippy warnings, and add review comments 2023-04-29 11:48:11 +02:00
Loïc Lecrenier
48f5bb1693 Implements the geo-sort ranking rule 2023-04-29 11:02:16 +02:00
Loïc Lecrenier
93188b3c88 Fix indexing of word_prefix_fid_docids 2023-04-29 10:56:48 +02:00
Loïc Lecrenier
bc4efca611 Add more tests for the attribute ranking rule 2023-04-29 10:56:48 +02:00
bors[bot]
414b3fae89
Merge #3571
3571: Introduce two filters to select documents with `null` and empty fields r=irevoire a=Kerollmops

# Pull Request

## Related issue
This PR implements the `X IS NULL`, `X IS NOT NULL`, `X IS EMPTY`, `X IS NOT EMPTY` filters that [this comment](https://github.com/meilisearch/product/discussions/539#discussioncomment-5115884) is describing in a very detailed manner.

## What does this PR do?

### `IS NULL` and `IS NOT NULL`

This PR will be exposed as a prototype for now. Below is the copy/pasted version of a spec that defines this filter.

- `IS NULL` matches fields that `EXISTS` AND `= IS NULL`
- `IS NOT NULL` matches fields that `NOT EXISTS` OR `!= IS NULL`

1. `{"name": "A", "price": null}`
2. `{"name": "A", "price": 10}`
3. `{"name": "A"}`

`price IS NULL` would match 1
`price IS NOT NULL` or `NOT price IS NULL` would match 2,3
`price EXISTS` would match 1, 2
`price NOT EXISTS` or `NOT price EXISTS` would match 3

common query : `(price EXISTS) AND (price IS NOT NULL)` would match 2

### `IS EMPTY` and `IS NOT EMPTY`

- `IS EMPTY` matches Array `[]`, Object `{}`, or String `""` fields that `EXISTS` and are empty
- `IS NOT EMPTY` matches fields that `NOT EXISTS` OR are not empty.

1. `{"name": "A", "tags": null}`
2. `{"name": "A", "tags": [null]}`
3. `{"name": "A", "tags": []}`
4. `{"name": "A", "tags": ["hello","world"]}`
5. `{"name": "A", "tags": [""]}`
6. `{"name": "A"}`
7. `{"name": "A", "tags": {}}`
8. `{"name": "A", "tags": {"t1":"v1"}}`
9. `{"name": "A", "tags": {"t1":""}}`
10. `{"name": "A", "tags": ""}`

`tags IS EMPTY` would match 3,7,10
`tags IS NOT EMPTY` or `NOT tags IS EMPTY` would match 1,2,4,5,6,8,9
`tags IS NULL` would match 1
`tags IS NOT NULL` or `NOT tags IS NULL` would match 2,3,4,5,6,7,8,9,10
`tags EXISTS` would match 1,2,3,4,5,7,8,9,10
`tags NOT EXISTS` or `NOT tags EXISTS` would match 6

common query : `(tags EXISTS) AND (tags IS NOT NULL) AND (tags IS NOT EMPTY)` would match 2,4,5,8,9

## What should the reviewer do?

- Check that I tested the filters
- Check that I deleted the ids of the documents when deleting documents


Co-authored-by: Clément Renault <clement@meilisearch.com>
Co-authored-by: Kerollmops <clement@meilisearch.com>
2023-04-27 13:14:00 +00:00
Loïc Lecrenier
899baa0ea5 Update forgotten snapshot from previous commit 2023-04-27 13:43:04 +02:00
Loïc Lecrenier
374095d42c Add tests for stop words and fix a couple of bugs 2023-04-27 13:30:09 +02:00
Louis Dureuil
b41a6cbd7a
Check sort criteria also in placeholder search 2023-04-26 16:28:17 +02:00
Louis Dureuil
c8af572697
Add tests for exact words and exact attributes 2023-04-26 16:13:01 +02:00
Loïc Lecrenier
b448aca49c Add more tests for exactness rr 2023-04-26 11:04:18 +02:00
Loïc Lecrenier
55bad07c16 Fix bug in exact_attribute rr implementation 2023-04-26 10:40:05 +02:00
Loïc Lecrenier
3421125a55 Prevent the exactness ranking rule from removing random words
Make it strictly follow the term matching strategy
2023-04-26 09:09:19 +02:00
Clément Renault
14293f6c8f
Make rustfmt happy 2023-04-25 16:55:39 +02:00
Loïc Lecrenier
d3a94e8b25 Fix bugs and add tests to exactness ranking rule 2023-04-25 16:49:08 +02:00
Clément Renault
cfd1b2cc97
Fix the clippy warnings 2023-04-25 16:40:32 +02:00
Loïc Lecrenier
8f2e971879 Add tests for "exactness" rr, make correct universe computation 2023-04-24 16:57:34 +02:00
Loïc Lecrenier
d1fdbb63da Make all search tests pass, fix distinctAttribute bug 2023-04-24 12:12:08 +02:00
Loïc Lecrenier
84d9c731f8 Fix bug in encoding of word_position_docids and word_fid_docids 2023-04-24 09:59:30 +02:00
Loïc Lecrenier
bd9aba4d77 Add "position" part of the attribute ranking rule 2023-04-13 10:46:09 +02:00
Loïc Lecrenier
8edad8291b Add logger to attribute rr, fix a bug 2023-04-13 10:25:00 +02:00
Kerollmops
d9cebff61c Add a simple test to check that attributes are ranking correctly 2023-04-13 08:27:09 +02:00
Loïc Lecrenier
30f7bd03f6 Fix compiler warning/errors caused by previous merge 2023-04-13 08:27:09 +02:00