Commit Graph

1529 Commits

Author SHA1 Message Date
bors[bot]
51cf44d6fd
Merge #456
456: Remove useless grenad merging r=Kerollmops a=Kerollmops

This PR must be merged after #454.

This PR removes the part of code that was merging all of the grenad Readers merging that we don't need as the indexer should have merged them and, therefore, we should only have one final grenad Reader. We reduce the amount of CPU usage and memory pressure we were doing uselessly.

`@ManyTheFish` are you sure I can skip merging the `word_docids` database?

Here is the benchmark comparison with the previously merged PR #454:
```
group                                              indexing_reintroduce-appending-sorted-values_c05e42a8    indexing_remove-useless-grenad-merging_d5b8b5a2
-----                                              -----------------------------------------------------    -----------------------------------------------
indexing/Indexing movies with default settings     1.06      16.6±1.04s        ? ?/sec                      1.00      15.7±0.93s        ? ?/sec
indexing/Indexing songs with default settings      1.16      60.1±7.07s        ? ?/sec                      1.00      51.7±5.98s        ? ?/sec
indexing/Indexing songs without faceted numbers    1.06      55.4±6.14s        ? ?/sec                      1.00      52.2±4.13s        ? ?/sec
```

And the comparison with multi-batch indexing before #436, we can see that we gain time for benchmarks that index datasets in multiple batches but there is _so much_ variance that it's not clear.

```
group                                                             indexing_benchmark-multi-batch-indexing-before-speed-up_45f52620    indexing_remove-useless-grenad-merging_d5b8b5a2
-----                                                             ----------------------------------------------------------------    -----------------------------------------------
indexing/Indexing geo_point                                       1.07       6.6±0.08s        ? ?/sec                                 1.00       6.2±0.11s        ? ?/sec
indexing/Indexing songs in three batches with default settings    1.12      57.7±2.14s        ? ?/sec                                 1.00      51.5±3.80s        ? ?/sec
indexing/Indexing songs with default settings                     1.00      47.5±2.52s        ? ?/sec                                 1.09      51.7±5.98s        ? ?/sec
indexing/Indexing songs without any facets                        1.00      43.5±1.43s        ? ?/sec                                 1.12      48.8±3.73s        ? ?/sec
indexing/Indexing songs without faceted numbers                   1.00      47.1±2.23s        ? ?/sec                                 1.11      52.2±4.13s        ? ?/sec
indexing/Indexing wiki                                            1.00    917.3±30.01s        ? ?/sec                                 1.09    998.7±38.92s        ? ?/sec
indexing/Indexing wiki in three batches                           1.09   1091.2±32.73s        ? ?/sec                                 1.00    996.5±15.70s        ? ?/sec
```

What do you think `@irevoire?` Should we change the benchmarks to make them do more runs?

Co-authored-by: Kerollmops <clement@meilisearch.com>
2022-03-01 16:48:08 +00:00
Kerollmops
d5b8b5a2f8
Replace the ugly unwraps by clean if let Somes 2022-02-28 16:31:33 +01:00
Kerollmops
8d26f3040c
Remove a useless grenad file merging 2022-02-28 16:31:33 +01:00
bors[bot]
21898ffc60
Merge #454
454: Reintroduce appending sorted entries when possible r=Kerollmops a=Kerollmops

This PR modifies the `sorter_into_lmdb_database` function to append values into the database instead of get-put-merging them, it should improve the indexation speed for when the database is empty.

```txt
group                                             indexing_main_25123af3                 indexing_reintroduce-appending-sorted-values_c05e42a8
-----                                             ----------------------                 -----------------------------------------------------
indexing/Indexing movies with default settings    1.07      17.8±0.99s        ? ?/sec    1.00      16.6±1.04s        ? ?/sec
indexing/Indexing songs with default settings     1.00      57.0±6.01s        ? ?/sec    1.05      60.1±7.07s        ? ?/sec
indexing/Indexing songs without any facets        1.10      51.8±5.36s        ? ?/sec    1.00      47.3±3.30s        ? ?/sec
```

Co-authored-by: Clément Renault <clement@meilisearch.com>
2022-02-28 14:55:37 +00:00
Clément Renault
04b1bbf932
Reintroduce appending sorted entries when possible 2022-02-24 14:50:45 +01:00
bors[bot]
382be56d36
Merge #453
453: Benchmark multi batch indexing r=Kerollmops a=Kerollmops

Hey `@irevoire,` could you please add the new benchmarks into influx?

Co-authored-by: Kerollmops <clement@meilisearch.com>
Co-authored-by: Clément Renault <clement@meilisearch.com>
2022-02-24 12:33:13 +00:00
Clément Renault
acfc96525c
Apply GitHub suggestions 2022-02-23 16:20:29 +01:00
Clément Renault
a820aa11e6
Add a new movies benchmark to test multi batch indexing 2022-02-23 16:20:29 +01:00
Kerollmops
8d2e3e4aba
Add a new wiki benchmark to test multi batch indexing 2022-02-23 16:20:29 +01:00
Kerollmops
ab5247dc64
Add a new songs benchmark to test multi batch indexing 2022-02-23 16:20:28 +01:00
bors[bot]
acd9535588
Merge #455
455: Raise the GitHub CI timeout limit to 72h r=irevoire a=Kerollmops



Co-authored-by: Kerollmops <clement@meilisearch.com>
2022-02-23 14:33:31 +00:00
Kerollmops
19bfb2649b
Raise the GitHub CI timeout limit to 72h 2022-02-23 15:27:51 +01:00
bors[bot]
25123af3b8
Merge #436
436: Speed up the word prefix databases computation time r=Kerollmops a=Kerollmops

This PR depends on the fixes done in #431 and must be merged after it.

In this PR we will bring the `WordPrefixPairProximityDocids`, `WordPrefixDocids` and, `WordPrefixPositionDocids` update structures to a new era, a better era, where computing the word prefix pair proximities costs much fewer CPU cycles, an era where this update structure can use the, previously computed, set of new word docids from the newly indexed batch of documents.

---

The `WordPrefixPairProximityDocids` is an update structure, which means that it is an object that we feed with some parameters and which modifies the LMDB database of an index when asked for. This structure specifically computes the list of word prefix pair proximities, which correspond to a list of pairs of words associated with a proximity (the distance between both words) where the second word is not a word but a prefix e.g. `s`, `se`, `a`. This word prefix pair proximity is associated with the list of documents ids which contains the pair of words and prefix at the given proximity.

The origin of the performances issue that this struct brings is related to the fact that it starts its job from the beginning, it clears the LMDB database before rewriting everything from scratch, using the other LMDB databases to achieve that. I hope you understand that this is absolutely not an optimized way of doing things.

Co-authored-by: Clément Renault <clement@meilisearch.com>
Co-authored-by: Kerollmops <clement@meilisearch.com>
2022-02-16 15:41:14 +00:00
Clément Renault
ff8d7a810d
Change the behavior of the as_cloneable_grenad by taking a ref 2022-02-16 15:40:08 +01:00
Clément Renault
f367cc2e75
Finally bump grenad to v0.4.1 2022-02-16 15:28:48 +01:00
bors[bot]
f2984f66e6
Merge #452
452: bump milli r=curquiza a=irevoire



Co-authored-by: Irevoire <tamo@meilisearch.com>
2022-02-16 13:49:14 +00:00
Irevoire
0defeb268c
bump milli 2022-02-16 13:27:41 +01:00
bors[bot]
030064da25
Merge #451
451: Update LICENSE with Meili SAS name r=Kerollmops a=curquiza

Check with thomas, we must put the real name of the company

Co-authored-by: Clémentine Urquizar - curqui <clementine@meilisearch.com>
2022-02-15 16:18:47 +00:00
Clémentine Urquizar - curqui
84035a27f5
Update LICENSE 2022-02-15 15:52:50 +01:00
bors[bot]
0885fcf973
Merge #450
450: Get rid of chrono in favor of time r=Kerollmops a=irevoire

We only use `chrono` as a wrapper around `time`, and since there has been an [open CVE on `chrono` for at least 3 months now](https://github.com/chronotope/chrono/pull/632) and the repo seems to be [struggling with maintenance](https://github.com/chronotope/chrono/pull/639), I think we should use `time` directly which is way more active and sufficient for our use case.

EDIT: Actually the CVE status has been known for more than 6 months: https://github.com/chronotope/chrono/issues/602

Co-authored-by: Irevoire <tamo@meilisearch.com>
2022-02-15 10:54:46 +00:00
Irevoire
48542ac8fd
get rid of chrono in favor of time 2022-02-15 11:41:55 +01:00
bors[bot]
ea15ad6c34
Merge #447
447: Update version for the next release (v0.22.1) r=curquiza a=curquiza



Co-authored-by: Clémentine Urquizar <clementine@meilisearch.com>
2022-02-07 17:44:09 +00:00
Clémentine Urquizar
d03b3ceb58
Update version for the next release (v0.22.1) 2022-02-07 18:39:29 +01:00
bors[bot]
5d58cb7449
Merge #442
442: fix phrase search r=curquiza a=MarinPostma

Run the exact match search on 7 words windows instead of only two. This makes false positive very very unlikely, and impossible on phrase query that are less than seven words.


Co-authored-by: ad hoc <postma.marin@protonmail.com>
2022-02-07 16:18:20 +00:00
bors[bot]
c5a996aa78
Merge #446
446: Update LICENSE r=Kerollmops a=curquiza



Co-authored-by: Clémentine Urquizar - curqui <clementine@meilisearch.com>
2022-02-07 09:47:39 +00:00
Clémentine Urquizar - curqui
1279c38ac9
Update LICENSE 2022-02-05 18:29:11 +01:00
bors[bot]
267d14c28d
Merge #445
445: allow null values in csv r=Kerollmops a=MarinPostma

This pr allows null values in csv:
- if the field is of type string, then an empty field is considered null (`,,`), anything other is turned into a string (i.e `, ,` is a single whitespace string)
- if the field is of type number, when the trimmed field is empty, we consider the value null (i.e `,,`, `, ,` are both null numbers) otherwise we try to parse the number.


Co-authored-by: ad hoc <postma.marin@protonmail.com>
2022-02-03 15:11:32 +00:00
ad hoc
bd2262ceea
allow null values in csv 2022-02-03 16:03:01 +01:00
ad hoc
13de251047
rewrite word pair distance gathering 2022-02-03 15:57:20 +01:00
bors[bot]
fda4f229bb
Merge #417
417: Change chunk size to 4MiB to fit more the end user usage r=Kerollmops a=ManyTheFish

Reverts meilisearch/milli#379

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 <many@meilisearch.com>
2022-02-02 18:30:59 +00:00
bors[bot]
2468ebb76b
Merge #444
444: Fix the parsing of ndjson requests to index more than the first line r=Kerollmops a=Kerollmops

This PR correctly uses the `BufRead` trait to read every line of the content instead of just the first one. This bug was only affecting the http-ui test crate.

Co-authored-by: Kerollmops <clement@meilisearch.com>
2022-02-02 17:59:44 +00:00
Kerollmops
9142ba9dd4
Fix the parsing of ndjson requests to index more than the first line 2022-02-02 17:55:13 +01:00
Many
d59bcea749 Revert "Revert "Change chunk size to 4MiB to fit more the end user usage"" 2022-02-02 17:01:13 +01:00
ad hoc
d852dc0d2b
fix phrase search 2022-02-01 20:21:33 +01:00
Kerollmops
fb79c32430
Compute the new, common and, deleted prefix words fst once 2022-01-27 11:00:18 +01:00
Clément Renault
51d1e64b23
Remove, now useless, the WriteMethod enum 2022-01-27 10:08:35 +01:00
Clément Renault
e9c02173cf
Rework the WordsPrefixPositionDocids update to compute a subset of the database 2022-01-27 10:08:35 +01:00
Clément Renault
dbba5fd461
Create a function to simplify the word prefix pair proximity docids compute 2022-01-27 10:08:35 +01:00
Clément Renault
e760e02737
Fix the computation of the newly added and common prefix pair proximity words 2022-01-27 10:08:35 +01:00
Clément Renault
d59e559317
Fix the computation of the newly added and common prefix words 2022-01-27 10:08:34 +01:00
Clément Renault
2ec8542105
Rework the WordPrefixDocids update to compute a subset of the database 2022-01-27 10:08:34 +01:00
Clément Renault
28692f65be
Rework the WordPrefixDocids update to compute a subset of the database 2022-01-27 10:08:34 +01:00
Clément Renault
5404bc02dd
Move the fst_stream_into_hashset method in the helper methods 2022-01-27 10:06:00 +01:00
Clément Renault
c90fa95f93
Only compute the word prefix pairs on the created word pair proximities 2022-01-27 10:06:00 +01:00
Clément Renault
822f67e9ad
Bring the newly created word pair proximity docids 2022-01-27 10:06:00 +01:00
Clément Renault
d28f18658e
Retrieve the previous version of the words prefixes FST 2022-01-27 10:05:59 +01:00
bors[bot]
38d23546a5
Merge #431
431: Fix and improve word prefix pair proximity r=ManyTheFish a=Kerollmops

This PR first fixes the algorithm we used to select and compute the word prefix pair proximity database. The previous version was skipping nearly all of the prefixes. The issue is that this fix made this method to take more time and we were trying to reduce the time spent in it.

With `@ManyTheFish` we found out that we could skip some of the work we were doing by:
 - discarding the prefixes that were shorter than a specific threshold (default: 2).
 - discarding the word prefix pairs with proximity bigger than a specific threshold (default: 4).
 - remove the unused threshold that was specifying a minimum amount of word docids to merge.

We will take more time to do some more optimization, like stop clearing and recomputing from scratch the database, we will compute the subsets of keys to create, keep and merge. This change is a little bit more complex than what this PR does.

I keep this PR as a draft as I want to further test the real gain if it is enough or not if it is valid or not. I advise reviewers to review commit by commit to see the changes bit by bit, reviewing the whole PR can be hard.

Co-authored-by: Clément Renault <clement@meilisearch.com>
2022-01-27 07:04:56 +00:00
bors[bot]
c63f945093
Merge #441
441: Changes related to the rebranding r=curquiza a=meili-bot

_This PR is auto-generated._

 - [X] Change the name `MeiliSearch` to `Meilisearch` in README.
 - [x] ⚠️ Ensure the bot did not update part you don’t want it to update, especially in the code examples in the Getting started.
 - [x] Please, ensure there is no other "MeiliSearch". For example, in the comments or in the tests name.
 - [x] Put the new logo on the README if needed -> still using the milli logo so far


Co-authored-by: meili-bot <74670311+meili-bot@users.noreply.github.com>
Co-authored-by: Clémentine Urquizar <clementine@meilisearch.com>
2022-01-26 17:07:37 +00:00
Clémentine Urquizar
0f213f2202
Replace MeiliSearch by Meilisearch 2022-01-26 17:49:55 +01:00
Clémentine Urquizar
de808a391a
Replace meilisearch by Meilisearch 2022-01-26 17:48:22 +01:00