Commit Graph

435 Commits

Author SHA1 Message Date
Many the fish
9e1b458010
Merge branch 'main' into change-proximity-precision-settings 2023-12-18 09:08:47 +01:00
Louis Dureuil
e0cc775dc4
Various changes
- DistributionShift in Search object (to be set from model in embed?)
- Fix issue where embedder index wasn't computed at search time
- Accept as default embedder either the "default" one, or the only embedder when there is only one
2023-12-14 16:08:41 +01:00
Louis Dureuil
922a640188
WIP multi embedders
fixed template bugs
2023-12-14 16:08:41 +01:00
Louis Dureuil
abbe131084
Cosmetic change 2023-12-14 16:08:41 +01:00
Louis Dureuil
13c2c6c16b
Small commit to add hybrid search and autoembedding 2023-12-14 16:07:48 +01:00
ManyTheFish
35e1981488 Remove proximityPrecision form the experimental feature 2023-12-14 15:52:42 +01:00
Clément Renault
7e259cb0d2
Expose the --max-number-of-batched-tasks argument 2023-12-11 16:08:39 +01:00
ManyTheFish
1f4fc9c229 Make the feature experimental 2023-12-06 15:49:05 +01:00
meili-bors[bot]
6376c342c1
Merge #4223
4223: Update to heed 0.20 r=dureuill a=Kerollmops

This PR brings the v0.20-alpha.9 version of heed into Meilisearch 🎉 The main goal is to test it in a real environment to make the necessary changes if needed. We also want to merge it as soon as possible during the pre-release phase to ensure we catch bugs before the release.

Most of the calls to heed are the same as before, except:
 - The `PolyDatabase` has been replaced with a `Database<Unspecified, Unspecified>`. We replaced the `get<T, U>()` by a `remap<T, U>().get()` calls.
 - The `Database` `append(...)` method has been replaced with a `put_with_flags(PutFlags::APPEND, ...)`.
 - The `RwTxn<'e, 'p>` has been simplified into a `RwTxn<'e>`.
 - The `BytesEncode/Decode` traits return a `Result<_, BoxedError>` instead of an `Option<_>`.
 - We no longer need to wrap and unwrap the `BEU32` integer when storing/getting them from heed.

### TODO
 - [x] Create actual, simple error types instead of using strings in the codecs.

### Follow-up work
 - Move the codecs into another member crate (we depend on the uuid one in the meilitool crate).
 - Display the internal decoding error in the `SerializationError` internal error variant.

Co-authored-by: Clément Renault <clement@meilisearch.com>
2023-11-28 13:39:44 +00:00
Clément Renault
ec9b52d608
Rename copy_to_path to copy_to_file 2023-11-28 14:32:30 +01:00
Clément Renault
34c67ac389
Remove the possibility to fail fetching the env info 2023-11-28 14:31:23 +01:00
Tamo
5751f5c640 fix puffin in the index scheduler 2023-11-27 15:18:33 +01:00
Clément Renault
0dbf1a16ff
Make clippy happy 2023-11-23 14:11:38 +01:00
Clément Renault
462b4c0080
Fix the tests 2023-11-23 12:07:35 +01:00
Clément Renault
0d4482625a
Make the changes to use heed v0.20-alpha.6 2023-11-23 11:43:58 +01:00
Clément Renault
7cb7e37ba8
Merge branch 'main' into tmp-release-v1.5.0 2023-11-21 16:30:46 +01:00
meili-bors[bot]
33b7c574ea
Merge #4090
4090: Diff indexing r=ManyTheFish a=ManyTheFish

This pull request aims to reduce the indexing time by computing a difference between the data added to the index and the data removed from the index before writing in LMDB.

## Why focus on reducing the writings in LMDB?

The indexing in Meilisearch is split into 3 main phases:
1) The computing or the extraction of the data (Multi-threaded)
2) The writing of the data in LMDB (Mono-threaded)
3) The processing of the prefix databases (Mono-threaded)

see below:
![Capture d’écran 2023-09-28 à 20 01 45](https://github.com/meilisearch/meilisearch/assets/6482087/51513162-7c39-4244-978b-2c6b60c43a56)


Because the writing is mono-threaded, it represents a bottleneck in the indexing, reducing the number of writes in LMDB will reduce the pressure on the main thread and should reduce the global time spent on the indexing.

## Give Feedback

We created [a dedicated discussion](https://github.com/meilisearch/meilisearch/discussions/4196) for users to try this new feature and to give feedback on bugs or performance issues.

## Technical approach
### Part 1: merge the addition and the deletion process
This part:
a) Aims to reduce the time spent on indexing only the filterable/sortable fields of documents, for example:
  - Updating the number of "likes" or "stars" of a song or a movie
  - Updating the "stock count" or the "price" of a product

b) Aims to reduce the time spent on writing in LMDB which should reduce the global indexing time for the highly multi-threaded machines by reducing the writing bottleneck.

c) Aims to reduce the average time spent to delete documents without having to keep the soft-deleted documents implementation

- [x] Create a preprocessing function that creates the diff-based documents chuck (`OBKV<fid, OBKV<AddDel, value>>`)
  - [x] and clearly separate the faceted fields and the searchable fields in two different chunks
- Change the parameters of the input extractor by taking an `OBKV<fid, OBKV<AddDel, value>>` instead of  `OBKV<fid, value>`.
  - [x] extract_docid_word_positions
  - [x] extract_geo_points
  - [x] extract_vector_points
  - [x] extract_fid_docid_facet_values
- Adapt the searchable extractors to the new diff-chucks
  - [x] extract_fid_word_count_docids
  - [x] extract_word_pair_proximity_docids
  - [x] extract_word_position_docids
  - [x] extract_word_docids
- Adapt the facet extractors to the new diff-chucks
  - [x] extract_facet_number_docids
  - [x] extract_facet_string_docids
  - [x] extract_fid_docid_facet_values
  - [x] FacetsUpdate
- [x] Adapt the prefix database extractors ⚠️ ⚠️ 
- [x] Make the LMDB writer remove the document_ids to delete at the same time the new document_ids are added
- [x] Remove document deletion pipeline
  - [x] remove `new_documents_ids` entirely and `replaced_documents_ids`
  - [x] reuse extracted external id from transform instead of re-extracting in `TypedChunks::Documents`
  - [x] Remove deletion pipeline after autobatcher
  - [x] remove autobatcher deletion pipeline
    - [x] everything uses `IndexOperation::DocumentOperation`
    - [x] repair deletion by internal id for filter by delete
    - [x] Improve the deletion via internal ids by avoiding iterating over the whole set of external document ids.  
- [x] Remove soft-deleted documents

#### FIXME

- [x] field distribution is not correctly updated after deletion
- [x] missing documents in the tests of tokenizer_customization

### Part 2: Only compute the documents field by field
This part aims to reduce the global indexing time for any kind of partial document modification on any size of machine from the mono-threaded one to the highly multi-threaded one.

- [ ] Make the preprocessing function only send the fields that changed to the extractors
- [ ] remove the `word_docids` and `exact_word_docids` database and adapt the search (⚠️ could impact the search performances)
- [ ] replace the `word_pair_proximity_docids` database with a `word_pair_proximity_fid_docids` database and adapt the search (⚠️ could impact the search performances)
- [ ] Adapt the prefix database extractors ⚠️ ⚠️

## Technical Concerns
- The part 1 implementation could increase the indexing time for the smallest machines (with few threads) by increasing the extracting time (multi-threaded) more than the writing time (mono-threaded)
- The part 2 implementation needs to change the databases which could have a significant impact on the search performances
- The prefix databases are a bit special to process and may be a pain to adapt to the difference-based indexing

Co-authored-by: ManyTheFish <many@meilisearch.com>
Co-authored-by: Clément Renault <clement@meilisearch.com>
Co-authored-by: Louis Dureuil <louis@meilisearch.com>
2023-11-21 09:44:38 +00:00
Tamo
5b57fbab08 makes the dump cancellable 2023-11-14 11:23:13 +01:00
Louis Dureuil
a2d6dc8571
Fix typo, remove caching for the change of index 2023-11-13 10:44:36 +01:00
Louis Dureuil
492fc086f0
cargo fmt 2023-11-12 21:53:11 +01:00
Louis Dureuil
a2d0c73b41
Save the currently updating index so that the search can access it at all times 2023-11-10 10:52:03 +01:00
Louis Dureuil
f8289cd974
Use it from delete-by-filter 2023-11-09 14:23:15 +01:00
Louis Dureuil
ef6fa10f7a
Remove IndexOperation::DocumentDeletion 2023-11-06 12:16:15 +01:00
Louis Dureuil
cbaa54cafd
Fix clippy issues 2023-11-06 11:19:31 +01:00
Clément Renault
e507ef5932
Slow the logging down 2023-11-01 13:49:32 +01:00
Clément Renault
13416ccbf7
Introduce a new meilitool to help the cloud team 2023-10-30 14:30:20 +01:00
Clément Renault
dfab6293c9
Use an LMDB database to store the external documents ids 2023-10-30 11:41:23 +01:00
Louis Dureuil
652ac3052d
use new iterator in batch 2023-10-30 11:41:22 +01:00
Louis Dureuil
c534a1b687
Stop using delete documents pipeline in batch runner 2023-10-30 11:41:22 +01:00
Louis Dureuil
cf8dad1ca0
index_scheduler.features() is no longer fallible 2023-10-23 10:38:56 +02:00
bwbonanno
dd619913da Use RwLock to never persist cli state to db 2023-10-19 12:45:57 -07:00
bwbonanno
d8c649b3cd Return recoverable error if we fail to retrieve metrics state 2023-10-18 08:28:24 -07:00
bwbonanno
12fc878640 Merge remote-tracking branch 'origin/main' into enable-metrics-http 2023-10-16 13:48:01 -07:00
bwbonanno
689ec7c7ad Make the experimental route /metrics activable via HTTP 2023-10-13 22:12:54 +00:00
Clément Renault
3655d4bdca
Move the puffin file export logic into the run function 2023-10-13 13:11:30 +02:00
Clément Renault
055ca3935b
Update index-scheduler/src/batch.rs
Co-authored-by: Tamo <tamo@meilisearch.com>
2023-10-13 13:11:30 +02:00
Kerollmops
bf8fac6676
Fix the tests 2023-10-13 13:11:30 +02:00
Kerollmops
f2a9e1ebbb
Improve the debugging experience in the puffin reports 2023-10-13 13:11:30 +02:00
Kerollmops
513e61e9a3
Remove the experimental CLI flag 2023-10-13 13:11:29 +02:00
Kerollmops
90a626bf80
Use the runtime feature to enable puffin report exporting 2023-10-13 13:11:29 +02:00
Kerollmops
0d4acf2daa
Fix the metrics product URL 2023-10-13 13:11:29 +02:00
Kerollmops
58db8d85ec
Add the exportPuffinReports option to the runtime features route 2023-10-13 13:11:29 +02:00
Clément Renault
656dadabea
Expose an experimental flag to write the puffin reports to disk 2023-10-13 13:11:09 +02:00
Tamo
34fac115d5 fix clippy 2023-09-11 17:15:57 +02:00
Tamo
9258e5b5bf Fix the stats of the documents deletion by filter
The issue was that the operation « DocumentDeletionByFilter » was not
declared as an index operation. That means the indexes stats were not
reprocessed after the application of the operation.
2023-09-11 14:04:10 +02:00
meili-bors[bot]
e4e49e63d0
Merge #3993
3993: Bringing back changes from v1.3.1 to `main` r=irevoire a=curquiza



Co-authored-by: irevoire <irevoire@users.noreply.github.com>
Co-authored-by: meili-bors[bot] <89034592+meili-bors[bot]@users.noreply.github.com>
Co-authored-by: Tamo <tamo@meilisearch.com>
Co-authored-by: ManyTheFish <many@meilisearch.com>
2023-08-10 14:30:02 +00:00
Tamo
fe819a9d80 fix the get stats method
It was not taking into account the processing tasks at all
2023-08-08 13:21:15 +02:00
ManyTheFish
b45c36cd71 Merge branch 'main' into tmp-release-v1.3.0 2023-08-01 15:05:17 +02:00
Kerollmops
eef95de30e
First iteration on exposing puffin profiling 2023-07-18 17:38:13 +02:00
Clément Renault
22762808ab
Fix the tests 2023-07-06 12:13:29 +02:00