Commit Graph

1620 Commits

Author SHA1 Message Date
bors[bot]
dde1da1c0e
Merge #382
382: Refactor attribute criterion r=Kerollmops a=ManyTheFish

### Re-implement set based algorithm for attribute criterion
#### Levels
Instead of doing level iteration and digging in the interesting level, we only iterate over the lowest level.

#### crossword iteration VS minimal position iteration
Instead of crossing word position in order to iterate strictly over the position that gives the best rank in good order; we iterate word by word starting with the word that increases the rank the little as possible.
This new method is a bit less precise but way simpler.

### Simplify word-level-position database
We don't use levels anymore in the attribute criterion, and so we removed the level complexity of the database making a word-position-docids database.

### Benchmarks on search on big datasets

#### songs main VS refactor-attribute-criterion
```diff
  group                                                   search_songsmain_31c18f09               search_songsrefactor-attribute-criterion_1bd15d84
  -----                                                   -------------------------               -------------------------------------------------
- smol-songs.csv: basic filter: <=/Notstandskomitee       1.00     84.8±0.58µs        ? ?/sec     1.09     92.2±8.98µs        ? ?/sec
+ smol-songs.csv: basic filter: TO/Notstandskomitee       1.18     98.0±6.30µs        ? ?/sec     1.00     83.2±0.97µs        ? ?/sec
+ smol-songs.csv: basic with quote/"david" "bowie"        114.68    76.0±0.20ms        ? ?/sec    1.00    662.5±5.03µs        ? ?/sec
- smol-songs.csv: basic with quote/"john"                 1.00    197.4±1.06µs        ? ?/sec     1.05    208.1±1.53µs        ? ?/sec
+ smol-songs.csv: basic with quote/"michael" "jackson"    2.75      2.0±0.01ms        ? ?/sec     1.00    738.9±3.91µs        ? ?/sec
+ smol-songs.csv: basic without quote/david bowie         297.42  1499.3±0.86ms        ? ?/sec    1.00      5.0±0.02ms        ? ?/sec
+ smol-songs.csv: basic without quote/michael jackson     2.55      8.9±0.02ms        ? ?/sec     1.00      3.5±0.01ms        ? ?/sec
+ smol-songs.csv: big filter/john                         1.08    473.6±2.25µs        ? ?/sec     1.00    438.1±2.59µs        ? ?/sec
- smol-songs.csv: prefix search/a                         1.00    446.9±1.81µs        ? ?/sec     1.79    800.5±4.45µs        ? ?/sec
- smol-songs.csv: prefix search/b                         1.00    398.5±2.74µs        ? ?/sec     1.81    723.1±5.46µs        ? ?/sec
- smol-songs.csv: prefix search/i                         1.00    486.3±1.99µs        ? ?/sec     1.69    823.6±9.42µs        ? ?/sec
- smol-songs.csv: prefix search/s                         1.00    229.6±3.29µs        ? ?/sec     2.59    594.4±2.22µs        ? ?/sec
- smol-songs.csv: prefix search/x                         1.00    150.2±0.76µs        ? ?/sec     1.11    166.0±0.87µs        ? ?/sec
```

On songs, the new algorithm gives a big improvement on slow queries, and is slower on one char prefix search (fast queries <1ms).

#### wiki main VS refactor-attribute-criterion
```diff
  group                                                           search_wikimain_31c18f09               search_wikirefactor-attribute-criterion_1bd15d84
  -----                                                           ------------------------               ------------------------------------------------
- smol-wiki-articles.csv: basic with quote/"rock" "and" "roll"    1.00      3.2±0.01ms        ? ?/sec    1.15      3.7±0.01ms        ? ?/sec
- smol-wiki-articles.csv: basic without quote/film                1.00    351.5±2.47µs        ? ?/sec    1.13    396.8±1.63µs        ? ?/sec
+ smol-wiki-articles.csv: basic without quote/rock and roll       1.10      9.4±0.02ms        ? ?/sec    1.00      8.6±0.04ms        ? ?/sec
- smol-wiki-articles.csv: basic without quote/spain               1.00    446.0±3.23µs        ? ?/sec    1.11    496.6±7.75µs        ? ?/sec
- smol-wiki-articles.csv: prefix search/c                         1.00    115.6±0.61µs        ? ?/sec    2.22    256.7±1.24µs        ? ?/sec
- smol-wiki-articles.csv: prefix search/g                         1.00    189.7±2.03µs        ? ?/sec    1.57    297.0±1.35µs        ? ?/sec
- smol-wiki-articles.csv: prefix search/j                         1.00    209.2±1.11µs        ? ?/sec    1.40    293.0±2.09µs        ? ?/sec
- smol-wiki-articles.csv: prefix search/q                         1.00     79.0±0.44µs        ? ?/sec    1.10     87.2±0.69µs        ? ?/sec
- smol-wiki-articles.csv: prefix search/t                         1.00    270.1±1.15µs        ? ?/sec    1.55    419.9±5.16µs        ? ?/sec
- smol-wiki-articles.csv: prefix search/x                         1.00    244.9±1.33µs        ? ?/sec    1.07    260.9±1.95µs        ? ?/sec
- smol-wiki-articles.csv: words/Abraham machin                    1.00      8.1±0.03ms        ? ?/sec    1.17      9.4±0.02ms        ? ?/sec
- smol-wiki-articles.csv: words/Idaho Bellevue pizza              1.00     19.3±0.07ms        ? ?/sec    1.07     20.6±0.05ms        ? ?/sec
```
On wiki we have some regressions `+17%` and `+15%` on request `>1ms`.

Co-authored-by: many <maxime@meilisearch.com>
2021-10-06 09:19:33 +00:00
many
085bc6440c
Apply PR comments 2021-10-06 11:12:26 +02:00
many
1bd15d849b
Reduce candidates threshold 2021-10-05 18:52:14 +02:00
many
ea4bd29d14
Apply PR comments 2021-10-05 17:35:07 +02:00
many
5ed75de0db
Update infos crate 2021-10-05 13:56:12 +02:00
many
3296bb243c
Simplify word level position DB into a word position DB 2021-10-05 12:15:02 +02:00
many
75d341d928
Re-implement set based algorithm for attribute criterion 2021-10-05 12:14:50 +02:00
bors[bot]
31c18f0953
Merge #381
381: Update version for the next release (v0.17.1) r=irevoire a=curquiza



Co-authored-by: Clémentine Urquizar <clementine@meilisearch.com>
2021-10-03 02:12:43 +00:00
Clémentine Urquizar
05d8a33a28
Update version for the next release (v0.17.1) 2021-10-02 16:21:31 +02:00
bors[bot]
c9092c72bf
Merge #380
380: Reserved keyword error message r=Kerollmops a=irevoire

And I missed _another_ reserved keyword error message in the filter :(

Co-authored-by: Tamo <tamo@meilisearch.com>
2021-10-01 07:13:31 +00:00
Tamo
d9eba9d145
improve and test the sort error message 2021-09-30 14:38:27 +02:00
Tamo
0ee67bb7d1
improve the reserved keyword error message for the filters 2021-09-30 14:38:27 +02:00
bors[bot]
22551d0941
Merge #379
379: Revert "Change chunk size to 4MiB to fit more the end user usage" r=curquiza a=ManyTheFish

Reverts meilisearch/milli#370

Co-authored-by: Many <legendre.maxime.isn@gmail.com>
2021-09-29 13:20:53 +00:00
Many
26b5dad042
Revert "Change chunk size to 4MiB to fit more the end user usage" 2021-09-29 15:08:39 +02:00
bors[bot]
6a057a3bd0
Merge #378
378: Hotfix meilisearch#1707 r=Kerollmops a=ManyTheFish

This PR contains an ugly quick fix of [meilisearch#1707](https://github.com/meilisearch/MeiliSearch/issues/1707).

- remove comparison reverse on rank. Enhancing relevancy and performances
- iterate over level 0 only. Enhancing performances.

A better fix is in development.

Co-authored-by: many <maxime@meilisearch.com>
Co-authored-by: Many <legendre.maxime.isn@gmail.com>
2021-09-29 12:57:31 +00:00
Many
2e49230ca2
Update milli/src/search/criteria/attribute.rs
Co-authored-by: Clément Renault <clement@meilisearch.com>
2021-09-29 14:49:45 +02:00
Many
7ad0214089
Update milli/src/search/criteria/attribute.rs
Co-authored-by: Clément Renault <clement@meilisearch.com>
2021-09-29 14:49:41 +02:00
many
1df5b8712b
Hotfix meilisearch#1707 2021-09-29 14:41:56 +02:00
bors[bot]
bfedbc1b6d
Merge #374
374: Enhance CSV document parsing r=Kerollmops a=ManyTheFish

Benchmarks on `search_songs` were crashing because of the CSV parsing.

Co-authored-by: many <maxime@meilisearch.com>
2021-09-29 08:55:54 +00:00
bors[bot]
68c758a533
Merge #376
376: Stop casting integer docids to string r=Kerollmops a=irevoire

When a docid is an integer, we stop casting it to a string, and thus we don't add `"` around it.

Co-authored-by: Tamo <tamo@meilisearch.com>
2021-09-29 08:32:48 +00:00
many
d2427f18e5
Enhance CSV document parsing 2021-09-29 10:25:33 +02:00
bors[bot]
00f94b1ffd
Merge #377
377: Update version for the next release (v0.17.0) r=Kerollmops a=curquiza



Co-authored-by: Clémentine Urquizar <clementine@meilisearch.com>
2021-09-28 20:43:33 +00:00
Clémentine Urquizar
0e8665bf18
Update version for the next release (v0.17.0) 2021-09-28 19:38:12 +02:00
Tamo
f65153ad64
stop casting integer docids to string 2021-09-28 18:35:54 +02:00
bors[bot]
adddf3f179
Merge #375
375: Fixes #365 r=Kerollmops a=vishnugt



Co-authored-by: Vishnu Ganesan <vganesan@microsoft.com>
Co-authored-by: Vishnu Gt <vishnugt@hotmail.com>
2021-09-28 14:42:48 +00:00
Vishnu Gt
785c1372f2
Change "settings" to "setting"
Co-authored-by: Clément Renault <renault.cle@gmail.com>
2021-09-28 20:11:32 +05:30
Vishnu Ganesan
3580b2d803 Fixes #365 2021-09-28 19:30:23 +05:30
bors[bot]
3a12f5887e
Merge #373
373: Improve error message for bad sort syntax with geosearch r=Kerollmops a=irevoire

`@Kerollmops` This should be the last PR for the geosearch and error handling, sorry for doing it in so many steps 😬 

Co-authored-by: Tamo <tamo@meilisearch.com>
2021-09-28 12:39:32 +00:00
Tamo
a80dcfd4a3
improve error message for bad sort syntax with geosearch 2021-09-28 14:32:24 +02:00
bors[bot]
b2a332599e
Merge #372
372: Fix Meilisearch 1714 r=Kerollmops a=ManyTheFish

The bug comes from the typo tolerance, to know how many typos are accepted we were counting bytes instead of characters in a word.
On Chinese Script characters, we were allowing  2 typos on 3 characters words.
We are now counting the number of char instead of counting bytes to assign the typo tolerance.

Related to [Meilisearch#1714](https://github.com/meilisearch/MeiliSearch/issues/1714)

Co-authored-by: many <maxime@meilisearch.com>
2021-09-28 11:59:45 +00:00
many
8046ae4bd5
Count the number of char instead of counting bytes to assign the typo tolerance 2021-09-28 12:10:43 +02:00
many
1988416295
Add failing test related to Meilisearch#1714 2021-09-28 12:05:11 +02:00
bors[bot]
3b479948c6
Merge #371
371: Provide a sort error handler r=Kerollmops a=irevoire

This PR simplify the error handling of asc-desc rules for Meilisearch or any other wrapper by providing directly in milli a new error type called `SortError` that can be generated from an `AscDescError` and that can be automatically converted to a `UserError`.

Basically now, wherever you are in the code as a user or in milli you can parse an `AscDesc` syntax and depending on the context, cast it either as a `SortError` or a `CriterionError` in one line with improved error messages.

Co-authored-by: Tamo <tamo@meilisearch.com>
2021-09-28 09:28:32 +00:00
Tamo
cc732fe95e
update http-ui to use the sort-error 2021-09-28 11:15:24 +02:00
Tamo
c7cb816ae1
simplify the error handling of the sort syntax for meilisearch 2021-09-27 19:07:22 +02:00
bors[bot]
4c09f6838f
Merge #370
370: Change chunk size to 4MiB to fit more the end user usage r=ManyTheFish a=ManyTheFish

We made several indexing tests using different sizes of datasets (5 datasets from 9MiB to 100MiB) on several typologies of VMs (`XS: 1GiB RAM, 1 VCPU`, `S: 2GiB RAM, 2 VCPU`, `M: 4GiB RAM, 3 VCPU`, `L: 8GiB RAM, 4 VCPU`).
The result of these tests shows that the `4MiB` chunk size seems to be the best size compared to other chunk sizes (`2Mib`, `4MiB`, `8Mib`, `16Mib`,  `32Mib`, `64Mib`, `128Mib`).

below is the average time per chunk size:

![Capture d’écran 2021-09-27 à 14 27 50](https://user-images.githubusercontent.com/6482087/134909368-ef0bc45e-68d5-49d1-aaf9-91113b7c410f.png)

<details>
<summary>Detailled data</summary>
<br>

![Capture d’écran 2021-09-27 à 14 39 48](https://user-images.githubusercontent.com/6482087/134909952-a36b1457-bbbd-4a6c-bbe5-519e4b926b5a.png)
</br>
</details> 


Co-authored-by: many <maxime@meilisearch.com>
2021-09-27 12:57:52 +00:00
many
b188063869
Change chunk size to 4MiB to fit more the end user usage 2021-09-27 14:26:21 +02:00
bors[bot]
0f8320bdc2
Merge #369
369: Add test checking the bug reported in meilisearch issue 1716 r=Kerollmops a=ManyTheFish

The bug is not present in the newer milli version.

Related to [Meilisearch#1716](https://github.com/meilisearch/MeiliSearch/issues/1716)

Co-authored-by: many <maxime@meilisearch.com>
2021-09-23 14:27:34 +00:00
many
551df0cb77
Add test checking the bug reported in meilisearch issue 1716 2021-09-23 15:55:39 +02:00
bors[bot]
87dd441a3a
Merge #367
367: Update version for the next release (v0.16.0) r=Kerollmops a=curquiza



Co-authored-by: Clémentine Urquizar <clementine@meilisearch.com>
2021-09-22 15:20:20 +00:00
Clémentine Urquizar
1eacab2169
Update version for the next release (v0.15.1) 2021-09-22 17:18:54 +02:00
bors[bot]
b806097141
Merge #366
366: Geosearch error handling r=Kerollmops a=irevoire

Rewrite most of geosearch error handling and another batch of tests on the criterion parsing.

Co-authored-by: Tamo <tamo@meilisearch.com>
Co-authored-by: Irevoire <tamo@meilisearch.com>
2021-09-22 15:08:11 +00:00
Irevoire
218f0a6661
Apply suggestions from code review
Co-authored-by: Clément Renault <clement@meilisearch.com>
2021-09-22 17:00:27 +02:00
Tamo
47ee93b0bd
return an error when _geoPoint is used but _geo is not sortable 2021-09-22 16:37:41 +02:00
Tamo
1e5e3d57e2
auto convert AscDescError into CriterionError 2021-09-22 16:37:41 +02:00
Tamo
023446ecf3
create a smaller and easier to maintain CriterionError type 2021-09-22 16:37:41 +02:00
Tamo
86e272856a
create an asc_desc error type that is never supposed to be returned to the end user 2021-09-22 16:37:41 +02:00
Tamo
257e621d40
create an asc_desc module 2021-09-22 16:37:41 +02:00
Tamo
113a061bee
fix the error handling on the criterion side 2021-09-22 15:09:07 +02:00
bors[bot]
ad3befaaf5
Merge #364
364: Fix all the benchmarks  r=Kerollmops a=irevoire

#324 broke all benchmarks.
I fixed everything and noticed that `cargo check --all` was insufficient to check the bench in multiple workspaces, so I also updated the CI to use `cargo check --workspace --all-targets`.

Co-authored-by: Tamo <tamo@meilisearch.com>
2021-09-22 12:40:34 +00:00