Lightning Fast, Ultra Relevant, and Typo-Tolerant Search Engine

MeiliSearch

Last update: Dec 31, 2022

Related tags

Text search rust search-engine database highlighting synonyms instant typo-tolerance site-search entreprise-search

Overview

MeiliSearch

Website | Roadmap | Blog | LinkedIn | Twitter | Documentation | FAQ

⚡ Lightning Fast, Ultra Relevant, and Typo-Tolerant Search Engine 🔍

MeiliSearch is a powerful, fast, open-source, easy to use and deploy search engine. Both searching and indexing are highly customizable. Features such as typo-tolerance, filters, and synonyms are provided out-of-the-box. For more information about features go to our documentation.

✨ Features

Search as-you-type experience (answers < 50 milliseconds)
Full-text search
Typo tolerant (understands typos and miss-spelling)
Faceted search and filters
Supports Kanji characters
Supports Synonym
Easy to install, deploy, and maintain
Whole documents are returned
Highly customizable
RESTful API

Getting started

Deploy the Server

Homebrew (Mac OS)

brew update && brew install meilisearch
meilisearch

Docker

docker run -p 7700:7700 -v $(pwd)/data.ms:/data.ms getmeili/meilisearch

Try MeiliSearch in our Sandbox

Create a MeiliSearch instance in MeiliSearch Sandbox. This instance is free, and will be active for 72 hours.

Run on Digital Ocean

Deploy on Platform.sh

APT (Debian & Ubuntu)

echo "deb [trusted=yes] https://apt.fury.io/meilisearch/ /" > /etc/apt/sources.list.d/fury.list
apt update && apt install meilisearch-http
meilisearch

Download the binary (Linux & Mac OS)

curl -L https://install.meilisearch.com | sh
./meilisearch

Compile and run it from sources

If you have the latest stable Rust toolchain installed on your local system, clone the repository and change it to your working directory.

git clone https://github.com/meilisearch/MeiliSearch.git
cd MeiliSearch
cargo run --release

Create an Index and Upload Some Documents

Let's create an index! If you need a sample dataset, use this movie database. You can also find it in the datasets/ directory.

curl -L 'https://bit.ly/2PAcw9l' -o movies.json

MeiliSearch can serve multiple indexes, with different kinds of documents. It is required to create an index before sending documents to it.

curl -i -X POST 'http://127.0.0.1:7700/indexes' --data '{ "name": "Movies", "uid": "movies" }'

Now that the server knows about your brand new index, you're ready to send it some data.

curl -i -X POST 'http://127.0.0.1:7700/indexes/movies/documents' \
  --header 'content-type: application/json' \
  --data-binary @movies.json

Search for Documents

In command line

The search engine is now aware of your documents and can serve those via an HTTP server.

The jq command-line tool can greatly help you read the server responses.

curl 'http://127.0.0.1:7700/indexes/movies/search?q=botman+robin&limit=2' | jq

{
  "hits": [
    {
      "id": "415",
      "title": "Batman & Robin",
      "poster": "https://image.tmdb.org/t/p/w1280/79AYCcxw3kSKbhGpx1LiqaCAbwo.jpg",
      "overview": "Along with crime-fighting partner Robin and new recruit Batgirl...",
      "release_date": "1997-06-20",
    },
    {
      "id": "411736",
      "title": "Batman: Return of the Caped Crusaders",
      "poster": "https://image.tmdb.org/t/p/w1280/GW3IyMW5Xgl0cgCN8wu96IlNpD.jpg",
      "overview": "Adam West and Burt Ward returns to their iconic roles of Batman and Robin...",
      "release_date": "2016-10-08",
    }
  ],
  "offset": 0,
  "limit": 2,
  "processingTimeMs": 1,
  "query": "botman robin"
}

Use the Web Interface

We also deliver an out-of-the-box web interface in which you can test MeiliSearch interactively.

You can access the web interface in your web browser at the root of the server. The default URL is http://127.0.0.1:7700. All you need to do is open your web browser and enter MeiliSearch’s address to visit it. This will lead you to a web page with a search bar that will allow you to search in the selected index.

| See the gif above

Documentation

Now that your MeiliSearch server is up and running, you can learn more about how to tune your search engine in the documentation.

Contributing

Hey! We're glad you're thinking about contributing to MeiliSearch! If you think something is missing or could be improved, please open issues and pull requests. If you'd like to help this project grow, we'd love to have you! To start contributing, checking issues tagged as "good-first-issue" is a good start!

Telemetry

MeiliSearch collects anonymous data regarding general usage. This helps us better understand developers' usage of MeiliSearch features.

To see what information we're retrieving, please see the complete list on the dedicated issue.

We also use Sentry to make us crash and error reports. If you want to know more about what Sentry collects, please visit their privacy policy website.

This program is optional, you can disable these analytics by using the MEILI_NO_ANALYTICS env variable.

💌 Contact

Feel free to contact us with any questions you may have:

🆕 Join our GitHub Discussions forum
Join our Slack community.
By opening an issue.

MeiliSearch is developed by Meili, a young company. To know more about us, you can read our blog. Any suggestion or feedback is highly appreciated. Thank you for your support!

Comments

Using Docker, Meilisearch v0.24 does not write data in /data.ms anymore
Describe the bug Following the docs in Getting Started / Installation / Download and launch / Docker, it says that Meilisearch write data in /data.ms. When new user try it with Docker using persistence volume, they will use -v $(pwd)/data.ms:/data.ms. But looking inside the folder, they will find no data.

To Reproduce Steps to reproduce the behavior:

Go to https://docs.meilisearch.com/learn/getting_started/installation.html#download-and-launch

Click on 'Docker' tab

Copy/Paste the command to start using meilisearch

Check your $(pwd)/data.ms folder on your host machine

No data is persist

Expected behavior Data appears in the right folder of the host machine

MeiliSearch version: v0.24

Additional context When changing user https://github.com/meilisearch/MeiliSearch/issues/1757 the workspace dir has also changed. It is no more /data.ms but /home/meili/data.ms

We could update the documentation or remove the WORDIR ${HOME} and let it be in /data.ms
bug impacts docs v0.25.0
opened by sovanna 46

The attribute does not work attributesToHighlight

Hello everyone! The word highlighting function in meilisearch does not work. Please help me solve this issue.

 public function __invoke (Request $request){
        
        $rubric  = Category::select('id', 'title')->where('type', 0)->get()->sortBy('title');
        $query = $request['query'];
        $post = Post::search($query,  function($meilisearch, $query, $options) use ($request){
            $date = strtotime(now());
            
            if(empty($request->rubric))
            {
                $options['filters'] = "status = 1";
            }
            else
            {
                $options['filters'] = "status = 1 AND (category_id = $request->rubric)";
                $options['attributesToHighlight'] = ['title', 'excerpt'];
            }
           
            return $meilisearch->search($query,$options );
        } )
        ->paginate(10);

        $modelCollection =  $post->load('category');
      
        return view('search.index', compact('post', 'request', 'rubric'));
    }

needs more info support

opened by Sliper959595 42

Meilisearch slow on updates, backlogged/behind by 30-50min
Describe the bug Not sure if meilisearch has been benchmarked by large scale apps as i have tried to find evidence of other users and most implementations talk about the size of the data and not the frequency of updates. I work on an app doing 40k+ DAUs and while search query perf is good. Updating index on single events like 'a user starring/liking a post' is not good. The index gets backlogged and is behind by 30-50min. I have tried many ways to optimize:

use sqs queue

give meilisearch its own cpu-optimize vm... its now only doing 20% load on the vm, so its not even using the resources effectively

minimize attributes, we have less than 10

minimize payload size, less than 1mb

To Reproduce Will be tough to repro as you need to be running meili under high load from many clients.

Expected behavior I would be fine if it was 5-10min behind in events, but almost 1 hour behind is too much to cause user confusion on the frontend.

Desktop (please complete the following information):

ubuntu vm, cpu optimized, 8vcpu, 32gb ram

bug performance
opened by aysark 42
The facets attributes are not added in displayedAttributes and searchableAttributes
I create an index indexUID with no primary key

Then I add the genre attribute as attributesForFaceting (via this route POST {{url}}/indexes/indexUID/settings/attributes-for-faceting) and this addition is well processed. Here are the settings right now:

{ "rankingRules": [ "typo", "words", "proximity", "attribute", "wordsPosition", "exactness" ], "distinctAttribute": null, "searchableAttributes": [], "displayedAttributes": [], "stopWords": [], "synonyms": {}, "acceptNewFields": true, "attributesForFaceting": [ "genre" ] }

Then, I add this document:

[ { "id": 2, "title": "Pride and Prejudice", "author": "Jane Austin", "genre": "romance" } ]

And when I use the route GET {{url}}/indexes/indexUID/documents, I don't have the genre attributes:

[ { "id": 2, "title": "Pride and Prejudice", "author": "Jane Austin" } ]

Indeed, here are the settings now:

{ "rankingRules": [ "typo", "words", "proximity", "attribute", "wordsPosition", "exactness" ], "distinctAttribute": null, "searchableAttributes": [ "id", "title", "author" ], "displayedAttributes": [ "title", "id", "author" ], "stopWords": [], "synonyms": {}, "acceptNewFields": true, "attributesForFaceting": [ "genre" ] }

Bug: the attributesForFaceting are not added to searchableAttribtues and displayedAttributes.

FYI, if I add the document before changing the attributesForFaceting settings, the genre attribute is present in searchableAttributes and displayedAttributes.

So, the problem seems to come from the order of steps execution.

Version: on master.

[ ] Fix in the code

[ ] Add test in the "classic" order = create index -> add document -> change facets settings -> check searchable and displayed attributes

[ ] Add test in the "buggy" order = create index -> change facets settings -> add document -> check searchable and displayed attributes

bug
opened by curquiza 35
Tracking issue: SortBy at query time
The SortBy feature is something that will help e-commerce websites or any website that needs to display documents in an order that depends on the user front-end settings, therefore at query time. This can be useful when you want the user to be able to choose if he want to see documents by ascending or descending price for example.

For that we will add two settings:

The sortableAttributes field to the settings, this way the engine is aware of the fields that it needs to index like numbers, etc...

The sortBy field to the search query parameters, it contains the name of one of the sortable attribute you defined in the settings and will be used as the last sorting criterion.

How Algolia handles that

Algolia uses replicas to handle this kind of custom search, for each of the custom sortable attributes you want to have, the tools of Algolia creates replicas of your main index and changes the settings.

When a search query with custom sorting rules is performed, the front-end tools will request the right index. This index is named based on the rules you changed e.g. "movies_asc(price)_desc(released_at)".

Notice that you pay for all the replicas you have and all documents that are in any index count, it means if your plan allows you to have 10k documents but you have one replica, therefore you can only have 5k documents.
feature request
opened by Kerollmops 30
Unclear error when there is not enough space left
Describe the bug When trying to index documents in a kubernetes volume that does not have enough space left, Meilisearch returns the following error:

{ “details”: { “indexedDocuments”: 0, “receivedDocuments”: 46206 }, “duration”: “PT512.665928885S”, “enqueuedAt”: “2022-03-17T14:58:24.465745303Z”, “error”: { “code”: “internal”, “link”: “https://docs.meilisearch.com/errors#internal”, “message”: “I/O error (os error 5)“, “type”: “internal” }, “finishedAt”: “2022-03-17T15:06:57.168374834Z”, “indexUid”: “bgg”, “startedAt”: “2022-03-17T14:58:24.502445949Z”, “status”: “failed”, “type”: “documentAddition”, “uid”: 3 }

This error might also happen outside of a kubernetes environment, not tested.

To Reproduce Steps to reproduce the behavior:

Create a Meilisearch instance in a Kubernetes cluster and use a persistent volume to store the data.ms

Index documents until the volume is full

Expected behavior A clear error that is documented should be returned

Meilisearch version: v0.26.0

EDIT from @curquiza

How?

When Meilisearch does not have enough place on the machine, you get the following error":

{ "message": "I/O error (os error 5).", "code": "internal", "type": "internal", "link": "https://docs.meilisearch.com/errors#internal" }

We want to replace

the code (not type) internal by no_space_left_on_device

the link #internal by #no_space_left_on_device

⚠️ This is what the specification already mentioned and it looks like Meilisearch does not follow it yet: https://github.com/meilisearch/specifications/blob/main/text/0061-error-format-and-definitions.md#no_space_left_on_device

Impacted teams

Since this error is already in the spec (and so in the docs) and supposed to exist, no ping to do. https://docs.meilisearch.com/reference/errors/error_codes.html#no-space-left-on-device
bug milli impacts cloud error handler
opened by nicolasvienot 29
Heed error: Function not implemented (os error 38)
Describe the bug Running the container fails due to 'heed error'. Changing version doesn't seem to make a difference. Perhaps related to underlying Rust function?

To Reproduce Steps to reproduce the behaviour:

docker-compose.yml contains...

search: image: getmeili/meilisearch:v0.18.1 restart: always volumes: - ./data.ms:/data.ms ports: - 7700:7700

docker-compose up In container log you will see... Error: heed error; Function not implemented (os error 38) Have tested with image 0.13.0, 0.16.0, 0.17.0, and 0.18.1. Using version 0.12.0 you will see a different error.

Expected behavior A clear and concise description of what you expected to happen.

Desktop (please complete the following information): Docker Desktop, also reproducible on Google Cloud, Cloud Compute, VM Instance.
bug
opened by chaskett 29
Attributes known by MeiliSearch but invisible for the user
Related to this comment in this discussion.

There are some situations where MeiliSearch knows information that the user cannot guess.

Example:

I add a document with title and author attributes.

I update the settings to only have title as displayed and searchable attributes -> now, I cannot see any author in my settings and in my documents either.

I add another document with title and author

The author field is not considered as a new field and is not added to the searchable and displayed attributes (we can expect this addition because the acceptNewFields boolean is set to true by default in my settings).

But here, for MeiliSearch, author is not a new field.

In your comment @Kerollmops you say:

We must not save an attribute that will never be used (not in the settings) and consider it a new attribute if we see new documents containing it (because we must not be aware of it) #761 (comment).

I understand here you want to remove the author field from MeiliSearch if it's not used in the settings or in the documents. But what about the information (= the value) in the document for the author field? The user might have miss-manipulated the settings and wants the author field to come back right after. We cannot erase their data. (Hope I'm clear)

I don't know what is the best behavior to adopt. Maybe having an entrypoint in the API to know what MeiliSearch knows...? I mean, all the attributes it already knows and to which you can access/manipulate with your settings...?
opened by curquiza 28
Ability to pass a configuration file to Meilisearch
Discussed with @gmourier and following this information: https://github.com/meilisearch/product/discussions/85#discussioncomment-2864656

Related to the following spec: https://github.com/meilisearch/specifications/pull/185

We want Meilisearch to accept a configuration file that would be an alternative to the command-line options you can set when launching Meilisearch, for instance

meilisearch --master-key=XXX

Requirements

Format

The file format should be TOML.

Priority order

Configuration file < Env variables < command-line options

It means the options in the configuration file can be overwritten by the environment variable set (if they exist) and that can be themself overwritten by the command-line options (if they exist)

Example

With this config file

master_key = "thisisasuper1337masterkey@@@" db_path = "./filepath1" http_address = "127.0.0.1:7701" no_anayltics = true

and with the following variables and command-line options:

export MEILI_DB_PATH=./filepath2 export MEILI_HTTP_ADDR=127.0.0.1:7702 meilisearch --master-key=abcd1234 --db-path=./filepath3

The final values kept by Meilisearch will be

No analytics thanks to no_analytics = true set in the config file

The db path is filepath3 because --db-path=./filepath3 has been set as command line options. ./filepath1 and ./filepath2 are ignored.

The HTTP address is 127.0.0.1:7702 because set as an environment variable. The value in the config file (127.0.0.1:7701) is ignored.

The master key is abcd1234 set in the command-line option. The value in the config file is ignored ( thisisasuper1337masterkey@@@ ).

Config file content

~~The config file will have sections gathering the same kind of variable.~~

Example:

master_key = "thisisasuper1337masterkey@@@" no_analytics = true [dump] path = "dumps" [snapshot] enabled = true [ssl] require_auth = true

See comments below, sections are not something we want.

What we would rather: a default config file containing

fake sections separated by comments

a comment header redirecting to the documentation for more information

@meilisearch/docs-team this file could be easily downloaded

curl https://raw.githubusercontent.com/meilisearch/meilisearch/main/config.toml > config.toml

Config file command line

The default config file name is config.toml and will be retrieved by Meilisearch by default in the current working directory

You can pass the config file as command-line option

meilisearch --config-file-path="./config.toml"

So the default value of config-file-path is ./config.toml.

This means by default, Meilisearch will try to read ./config.toml.

If no ./config.toml file is found and no --config-file-path option is passed, Meilisearch does not throw any error. If --config-file-path is passed and is associated to an existing file (or impossible to open), Meilisearch throw an error

Display when launching Meilisearch

888b d888 d8b 888 d8b 888 8888b d8888 Y8P 888 Y8P 888 88888b.d88888 888 888 888Y88888P888 .d88b. 888 888 888 .d8888b .d88b. 8888b. 888d888 .d8888b 88888b. 888 Y888P 888 d8P Y8b 888 888 888 88K d8P Y8b "88b 888P" d88P" 888 "88b 888 Y8P 888 88888888 888 888 888 "Y8888b. 88888888 .d888888 888 888 888 888 888 " 888 Y8b. 888 888 888 X88 Y8b. 888 888 888 Y88b. 888 888 888 888 "Y8888 888 888 888 88888P' "Y8888 "Y888888 888 "Y8888P 888 888 Config file path: "./config.toml" Database path: "./data.ms"

If the users do not pass any config file and no ./config.toml exists, we must display none instead of "./config.toml"

Errors (bonus)

We need to parse and display clear errors message for the configuration scheme. e.g. https://docs.rs/miette/latest/miette/

Implementations

We could probably start by using the toml crate and declare a struct with a simple derive on it. We could start the job in the options.rs file or maybe create a new one specifically for this config file.

If you have any questions regarding the feature or the implementation, please let us know about this issue! Every feedback or questions are welcome!

TODO

[x] Merge the first iteration of this issue (https://github.com/meilisearch/meilisearch/pull/2745)

[x] Improve the default config file (assigned: @curquiza) #2928

[x] Fix the following sub-issues

[x] https://github.com/meilisearch/meilisearch/issues/2800

[x] https://github.com/meilisearch/meilisearch/issues/2801

[X] Create the specification -> https://github.com/meilisearch/specifications/pull/185

For Meili people

A spec will be open once we have implemented it. Some changes can still be applied, we need to see the final implementation to create the spec. Since we let this issue (so far) to the community, we don't know yet in which release this new feature will be available (at minimum v0.29.0, not before).

Ping @meilisearch/cloud-team @meilisearch/docs-team who can be concerned by this issue.
enhancement good first issue impacts docs impacts integrations impacts cloud v0.30.0
opened by curquiza 27

Migrating 40M records from Postgres to Meilisearch takes too long

I am trying to migrate 40M Postgres records (user comments) to Melisearch. Each time, I am querying 10K records from Postgres and writing to Meilisearch. The migration script has been completed. But it has been more than 3 days, only 4.2 Million records only migrated. The count is increasing very slowly.

I also see the following error messages in the error log:

[2021-02-18T12:48:44Z INFO  ureq::unit] sending request POST https://api.amplitude.com/httpapi
[2021-02-18T13:14:58Z ERROR meilisearch_core::database] commit nested transaction failed: Input/output error (os error 5)
[2021-02-18T13:43:50Z ERROR meilisearch_core::database] commit nested transaction failed: Input/output error (os error 5)
[2021-02-18T13:48:45Z INFO  ureq::unit] sending request POST https://api.amplitude.com/httpapi
[2021-02-18T14:13:03Z ERROR meilisearch_core::database] commit nested transaction failed: Input/output error (os error 5)
[2021-02-18T14:42:05Z ERROR meilisearch_core::database] commit nested transaction failed: Input/output error (os error 5)
[2021-02-18T14:48:45Z INFO  ureq::unit] sending request POST https://api.amplitude.com/httpapi
[2021-02-18T15:12:06Z ERROR meilisearch_core::database] commit nested transaction failed: Input/output error (os error 5)
[2021-02-18T15:41:40Z ERROR meilisearch_core::database] commit nested transaction failed: Input/output error (os error 5)
[2021-02-18T15:48:46Z INFO  ureq::unit] sending request POST https://api.amplitude.com/httpapi

Also, I tried to get updates using the following curl

curl 'http://localhost:7700/indexes/comments/updates' | jq

It is running forever, not getting any result from the server.

What is commit nested transaction failed and How to fix it?
Is there any best practice to write large amounts data to meilisearch?
Is there any limitation in the amount of data to be handled by meilisearch?
Is there an option to disable amplitude data?

bug performance

opened by aaqibjavith 26

Unable to add documents and endpoints timeout/unresponsive
Describe the bug Since running v0.21.0 I've been noticing that after a short while, there are no more documents being added. This is because the document endpoint simply times out (HttpTimeout of 100 seconds is hit on our backend). Also trying to call the /stats endpoint for an index does not respond.

However, other endpoints like /indexes/{index_uid} and /indexes/{index_uid}/search respond fine.

To Reproduce Steps to reproduce the behavior:

Not found.

Expected behavior That documents can be added and /stats returns the index statistics.

Screenshots Timeout from our backend:

Example of curl from inside the container on localhost. First showing the index exists/returns. Second with short 10 second timeout.

MeiliSearch version: v0.21.0rc3

Additional context Running on kubernetes, using the meilisearch chart. Deletion of the volume/fresh run doesn't fix anything (have to do this on upgrades anyway)
bug needs more info performance
opened by carlreid 25
Meilisearch Version 30.4 Undefined array key "facets"
I am new to Meilisearch been on this for days,

I have followed the documentation

These are the steps

POSTMAN "filterableAttributes":["colour", "size", "fragrance", "theme", "crystal"] TO ADD THEM

PRODUCT MODEL public function toSearchableArray() {

return array_merge([ 'id' => $this->id, 'title' => $this->title, 'slug' => $this->slug, 'price' => $this->price, 'category_ids' => $this->categories->pluck('id')->toArray() ], $this->variations->groupBy('type') ->mapWithKeys(fn ($variation, $key) => [ $key => $variation->pluck('title') ]) ->toArray()

);

Next PRODUCT BROWESER $searchquery = Product::search('', function ($meilisearch, string $query, array $options) {

$options['facets'] = ['colour', 'theme', 'size', 'fragrance']; return $meilisearch->search($query, $options); })->raw();

$search = $this->category->products->find(collect($searchquery['hits'])->pluck('id'));

return view('livewire.product-browser', [ 'currencies' => $currencies, 'filters' => $search['facets'], 'products' => $products ]);

C:\wamp64\www\www.leeweaver.uk\vendor\laravel\framework\src\Illuminate\Collections\Collection .php : 1701 public function offsetGet($key): mixed

{ return $this->items[$key]; }

What am I doing wrong please?.

Lee
opened by leeweaver2222 0
Bump tokio from 1.21.2 to 1.23.1
Bumps tokio from 1.21.2 to 1.23.1.

Release notes

Sourced from tokio's releases.

Tokio v1.23.1

This release forward ports changes from 1.18.4.

Fixed

net: fix Windows named pipe server builder to maintain option when toggling pipe mode (#5336).

#5336: tokio-rs/tokio#5336

Tokio v1.23.0

Fixed

net: fix Windows named pipe connect (#5208)

io: support vectored writes for ChildStdin (#5216)

io: fix async fn ready() false positive for OS-specific events (#5231)

Changed

runtime: yield_now defers task until after driver poll (#5223)

runtime: reduce amount of codegen needed per spawned task (#5213)

windows: replace winapi dependency with windows-sys (#5204)

#5208: tokio-rs/tokio#5208 #5216: tokio-rs/tokio#5216 #5213: tokio-rs/tokio#5213 #5204: tokio-rs/tokio#5204 #5223: tokio-rs/tokio#5223 #5231: tokio-rs/tokio#5231

Tokio v1.22.0

Added

runtime: add Handle::runtime_flavor (#5138)

sync: add Mutex::blocking_lock_owned (#5130)

sync: add Semaphore::MAX_PERMITS (#5144)

sync: add merge() to semaphore permits (#4948)

sync: add mpsc::WeakUnboundedSender (#5189)

Added (unstable)

process: add Command::process_group (#5114)

runtime: export metrics about the blocking thread pool (#5161)

task: add task::id() and task::try_id() (#5171)

Fixed

macros: don't take ownership of futures in macros (#5087)

runtime: fix Stacked Borrows violation in LocalOwnedTasks (#5099)

runtime: mitigate ABA with 32-bit queue indices when possible (#5042)

task: wake local tasks to the local queue when woken by the same thread (#5095)

time: panic in release mode when mark_pending called illegally (#5093)

runtime: fix typo in expect message (#5169)

... (truncated)

Commits

1a997ff chore: prepare Tokio v1.23.1 release

a8fe333 Merge branch 'tokio-1.20.x' into tokio-1.23.x

ba81945 chore: prepare Tokio 1.20.3 release

763bdc9 ci: run WASI tasks using latest Rust

9f98535 Merge remote-tracking branch 'origin/tokio-1.18.x' into fix-named-pipes-1.20

9241c3e chore: prepare Tokio v1.18.4 release

699573d net: fix named pipes server configuration builder

3ce5a26 chore: prepare Tokio v1.23 release (#5270)

644cb82 rt: fix *_closed false positives (#5231)

a1316cd io: impl std::io::BufRead on SyncIoBridge\<T> (#5265)

Additional commits viewable in compare view

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

@dependabot rebase will rebase this PR

@dependabot recreate will recreate this PR, overwriting any edits that have been made to it

@dependabot merge will merge this PR after your CI passes on it

@dependabot squash and merge will squash and merge this PR after your CI passes on it

@dependabot cancel merge will cancel a previously requested merge and block automerging

@dependabot reopen will reopen this PR if it is closed

@dependabot close will close this PR and stop Dependabot recreating it. You can achieve the same result by closing it manually

@dependabot ignore this major version will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)

@dependabot ignore this minor version will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)

@dependabot ignore this dependency will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself) You can disable automated security fix PRs for this repo from the Security Alerts page.

dependencies rust
opened by dependabot[bot] 0
Update dependencies

This PR bumps most of the dependencies to the next compatible patch version. It doesn't bump the incompatible ones and neither enum-iterator due to vergen forcing us to keep the old version.

opened by Kerollmops 1
Primary Key inference simplification: Address error message change
The changes requested here need to be addressed:

https://github.com/meilisearch/meilisearch/pull/3269#discussion_r1061527697

https://github.com/meilisearch/meilisearch/pull/3269#discussion_r1061540382

This requires:

[ ] A new milli PR to change the error message

[ ] A new meilisearch PR to integrate the changes and update the tests

[ ] An update to the related spec PR: https://github.com/meilisearch/specifications/pull/208

milli
opened by dureuill 0
Meilisearch use way more memory than it should
Describe the bug As you can see in the following image, meilisearch constantly goes over 100% of RAM usage, then it gets killed, tries to index data again and gets killed again etc...

Context

The searches running at the same time are probably not the issue (the machine isn't used a lot)

Meilisearch was able to guess the 70% of the RAM available correctly; it limits itself to 2.5GiB and then tries to use 4GiB and die

There is a lot of tasks processing at the same time. 45522, to be exact; maybe this is linked to the issue, maybe not.

The geosearch is not used

The issue appeared to two persons who were sending pretty small batches;

1000 documents per batch with ~1M documents in total

256 documents per batch with ~2M documents in total

The specs of the first person are;

4 GB Memory / 2 Intel vCPUs / 25 GB Disk + 90 GB / NYC3 - MeiliSearch 0.30.0 on Debian 10 (buster)

MAYBE, a first fast fix could be to provide a way to limit the number of tasks the autobatching can batch together. imo that's not the best option, and maybe it won't even solve the issue, but we never know. If we do that, I'm 100% toward an unstable cli flag because I don't want to stabilize this option.

The real fix should be to find where it comes from in milli and stops using too much memory, but that could take a long time.
bug milli
opened by irevoire 3
Create release changelogs for v1.1.0

This issue is the entry point regarding the release changelog we must prepare before the first RC of v1.1.0.

The changelog will be available as a PR in the engine-team repository. Once the PR is ready to be reviewed, the link will be shared in this issue.

📚 @meilisearch/docs-team we need you to review the PR, if possible, before the first RC of v1.1.0. If lacking of time, the most important is to review it by the date of the official release. The release date is available on the v1.1.0 Milestone.

Once the official release of v1.1.0 is out, the PR will be merged and this issue closed.
documentation impacts docs

opened by meili-bot 0

Releases(v0.30.5)

v0.30.5(Dec 21, 2022)
🚨 Please prefer using v0.30.5 instead of any v0.30.4 and before.

⚠️ If you encountered the following bug, have already tried to download this version and still have the bug, it means your database is corrupted. The only way you can recover your data is to send your documents again in a fresh index. Using a dump will not work, either. We are sorry for the inconvenience.

Bug fixes 🐞

Fix bug regarding the duplicated documents when searching, see this issue (#3267 and meilisearch/milli#754) @loiclec

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.84 MB)
meilisearch-linux-amd64(111.17 MB)
meilisearch-macos-amd64(104.35 MB)
meilisearch-macos-apple-silicon(108.41 MB)
meilisearch-windows-amd64.exe(103.21 MB)
meilisearch.deb(18.16 MB)
v0.30.4(Dec 15, 2022)
🚨 Replace v0.30.3. Please prefer using v0.30.4 instead if you use pagination.

Bug fixes 🐞

Fix pagination issue, see this issue (#3258 and meilisearch/milli#743) @ManyTheFish

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.86 MB)
meilisearch-linux-amd64(111.18 MB)
meilisearch-macos-amd64(104.38 MB)
meilisearch-macos-apple-silicon(108.41 MB)
meilisearch-windows-amd64.exe(103.23 MB)
meilisearch.deb(18.17 MB)
v0.30.3(Dec 14, 2022)
🚨 Replace v0.30.2. Please prefer using v0.30.3 instead.

⚠️ If you encountered one of this bug (#2945), have already tried to download this version and still have the bug, it means your database is corrupted. The only way you can recover your data is to send your documents again in a fresh index. Using a dump will not work, either. We are sorry for the inconvenience.

Bug fixes 🐞

Fix the pagination bugs, see issues #3200 and #3228 (meilisearch/milli#737 #3248) @ManyTheFish

Fix missing internal key bug see issue #2945 (meilisearch/milli#734 #3248) @loiclec

Fix CLI snapshot and dumps options (#3237) @irevoire

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.70 MB)
meilisearch-linux-amd64(111.06 MB)
meilisearch-macos-amd64(104.35 MB)
meilisearch-macos-apple-silicon(108.38 MB)
meilisearch-windows-amd64.exe(103.22 MB)
meilisearch.deb(18.20 MB)
v0.30.2(Dec 8, 2022)
🚨 Replace v0.30.1. Please prefer using v0.30.2 instead of v0.30.0 or v0.30.1 if possible

Bug fixes 🐞

Fix bugs regarding filters (meilisearch/milli#727 and #3217) @loiclec

Badly formatted instance-uid (#3213) @irevoire

Fix pagination with distinct attribute (meilisearch/milli#729 and #3217) @ManyTheFish

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.71 MB)
meilisearch-linux-amd64(111.07 MB)
meilisearch-macos-amd64(104.35 MB)
meilisearch-macos-apple-silicon(108.38 MB)
meilisearch-windows-amd64.exe(103.22 MB)
meilisearch.deb(18.20 MB)
v0.30.1(Dec 6, 2022)
🚨 Replace v0.30.0, please use v0.30.1 instead of v0.30.0 if possible

⚠️ If you encountered one of these bugs, have already tried to download this version and still have the bug, it means your database is corrupted. The only way you can recover your data is to send your documents again in a fresh index. Using a dump will not work, either. We are sorry for the inconvenience.

Bug fixes 🐞

Fix duplicated documents at search time (https://github.com/meilisearch/milli/pull/723 and #3202) @loiclec

Fix search with filter (https://github.com/meilisearch/milli/pull/712 and #3202) @loiclec

Prevent crash when indexing some documents (https://github.com/meilisearch/milli/pull/712 and #3202) @loiclec

Filtering results with _geoRadius(lat,lng,0) now returns results (https://github.com/meilisearch/milli/pull/722 and #3202) @amab8901

Use config sizes in Gb again (https://github.com/meilisearch/meilisearch/pull/3160) @Kerollmops

Thanks again to @amab8901 for his contribution to our release bug fixes! Also, a big thanks to every people who helped us debug, especially @0x15f,@Fronix, @gbuckingham89, @mmachatschek, @roparz, and @thijndehaas ❤️
Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.68 MB)
meilisearch-linux-amd64(111.06 MB)
meilisearch-macos-amd64(104.34 MB)
meilisearch-macos-apple-silicon(108.37 MB)
meilisearch-windows-amd64.exe(103.22 MB)
meilisearch.deb(18.19 MB)
v0.29.3(Dec 5, 2022)
🚨 Replace v0.29.2 which did not contain all the expected bug fixes.

⚠️ Like for v0.29.2, if you encountered one of these bugs, have already tried to download this version and still have the bug, it means your database is corrupted. The only way you can recover your data is to send your documents again in a fresh index. Using a dump will not work, either. We are sorry for the inconvenience

Bug fixes 🐞

Prevent duplicate documents in the search route (#3047 and meilisearch/milli#698) @Kerollmops

Fix missing key in the documents database (#3047 and meilisearch/milli#698) @Kerollmops

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.30 MB)
meilisearch-linux-amd64(110.34 MB)
meilisearch-macos-amd64(103.70 MB)
meilisearch-windows-amd64.exe(102.42 MB)
meilisearch.deb(17.97 MB)
v0.30.0(Nov 28, 2022)
This release introduces index deployment with zero downtime and adds task management features. It also brings the capability to navigate search results by page selection.

Thanks to Hacktoberfest, tons of updates and bug fixes have been done by the fantastic Meilisearch community. You were more present than ever, and words cannot do justice to just how grateful we are for your support. A big thank you from the whole Meilisearch team! ❤️

🧰 Most integrations (SDKs, clients, libraries) will be compatible with this version within four hours of the release. Sometimes this can take up to 48 hours, depending on the issues we encounter during the release.

Here is the exhaustive list of integrations not immediately compatible with v0.30.0:

meilisearch-dart and meilisearch-swift: though you can still use both libraries with v0.30.0, the new available features are not yet available. Let us know if you really need it—and contributions are always welcome!

meilisearch-java: Our team is still working on refactoring the code; this is reaching the end, but the SDK cannot be used with v0.30.0 yet. Follow the repository to know when it will be compatible.

New features and improvements 🔥

Improve search result navigation: exhaustive number of search results

When paginating search results with offset and limit, Meilisearch only returns an estimate of the total number of results. Since estimatedTotalHits can change, creating reliable pagination interfaces with numbered page selectors is challenging.

v0.30 introduces two new search parameters, page and hitsPerPage. Queries with these parameters return an exhaustive number of totalHits and totalPages, which you can then use to create UI elements such as numbered page selectors.

The following example fetches the second page of results for a given query:

curl \ -X POST 'http://localhost:7700/indexes/movies/search' \ -H 'Content-Type: application/json' \ --data-binary '{ "q": "shifu", "page": "2", "hitsPerPage": "10" }'

{ "hits": [ // … 10 hits ], // … "page": 2, "hitsPerPage": 10, "totalHits": 2100, "totalPages": 210 }

Done in #2601 by @ManyTheFish.

Cancel processing and enqueued tasks

Meilisearch now allows you to cancel enqueued or processing tasks through a new API route, /tasks/cancel. This can be useful when you need to interrupt a task that is taking too much time to be completed or enqueued by mistake.

To cancel tasks, filter them by specifying one of the following query parameters: uids, statuses, types, or indexUids.

Tasks can be filtered by date fields:

beforeEnqueuedAt / afterEnqueuedAt

beforeStartedAt / afterStartedAt

For example, to cancel tasks by uid:

curl \ -X POST 'http://localhost:7700/tasks/cancel?uids=1,2'

{ "taskUid": 1, "indexUid": "null", "status": "enqueued", "type": "taskCancelation", "enqueuedAt": "2021-08-12T10:00:00.000000Z" }

To cancel all enqueued tasks across all indexes in an instance:

curl \ -X POST 'http://localhost:7700/tasks/cancel?statuses=enqueued'

Read more about cancelling tasks.

Done in #2763 by @irevoire, @Kerollmops and @loiclec.

Delete the task history

This release introduces a new DELETE endpoint to the /tasks route. You can use it to remove finished tasks (succeeded, failed or canceled) from Meilisearch's task history. This can be helpful in reducing the amount of occupied disk space.

To delete finished tasks, filter them by specifying one of the following query parameters: uids, statuses, types, indexUids, or canceledBy.

Tasks can be filtered by date fields:

beforeEnqueuedAt / afterEnqueuedAt

beforeStartedAt / afterStartedAt

beforeFinishedAt / afterFinishedAt

For example, to delete tasks by uid:

curl \ -X DELETE 'http://localhost:7700/tasks?uids=1,2'

{ "taskUid": 3, "indexUid": null, "status": "enqueued", "type": "taskDeletion", "enqueuedAt": "2021-08-12T10:00:00.000000Z" }

To delete all finished tasks in an instance:

curl \ -X DELETE 'http://localhost:7700/tasks?statuses=failed,succeeded,canceled'

Read more about deleting finished tasks.

Done in #2763 by @irevoire, @Kerollmops, and @loiclec.

Deploy indexes with zero downtime

Deploy new indexes version with zero downtime to the search clients. Meilisearch now provides a seamless way to deploy multiple indexes atomically with the new swap indexes API.

To swap indexes, use the new /swap-indexes route:

curl \ -X POST 'http://localhost:7700/swap-indexes'\ -H 'Content-Type: application/json' \ --data-binary '[ { "indexes": ["indexA", "indexB"] } ]'

{ "taskUid": 3, "indexUid": null, "status": "enqueued", "type": "indexSwap", "enqueuedAt": "2021-08-12T10:00:00.000000Z" }

Read more about swapping indexes.

Done in #2763 by @irevoire, @Kerollmops, and @loiclec.

Configuration file support

You can now use a configuration file to set instance options.

By default, Meilisearch will search the working directory for a file named ./config.toml.

You can use the --config-file-path option if you want to keep your configuration file in another directory:

meilisearch --config-file-path="./my-config.toml"

Example of configuration file:

env = "production" master_key = "MY_MASTER_KEY" schedule_snapshot = true

You can download a default configuration file from our repository:

curl https://raw.githubusercontent.com/meilisearch/meilisearch/main/config.toml > config.toml

Read more about customizing Meilisearch with a configuration file.

Done in #2745, #2928, #2804, #2841, and #2961 by @mlemesle, @choznerol, @arriven, @LunarMarathon and @curquiza,

Other improvements

The fresh new filters are also availabe for the GET /tasks: uids, canceledBy, beforeEnqueuedAt, afterEnqueuedAt, beforeStartedAt, afterStartedAt, beforeFinishedAt, afterFinishedAt.

Full support for compressed API requests (Gzip, Brotli, Zlib) (#2876) @mou

Significant indexing speed improvements (#2763, meilisearch/milli#639, meilisearch/milli#619) @loiclec

Change default bind address from 127.0.0.1 to localhost, making the default address available in IPv4 and IPv6 (#2861) @Fall1ngStar

Provide Apple Silicon binaries (#2837) @jeertmans

Add missing environment variables for dump and snapshot features (#2738) @gmourier

Increase max concurrent readers from 126 to 1024 (#2830) @arriven

Reduce the size of the Meilisearch binary: allow excluding specialized tokenizations when building binaries (#2773) @jirutka

Introduce the snapshotCreation task type: snapshot tasks are added to the task list, for more transparency (#2763) @Kerollmops

Reduce the size taken by Meilisearch on the disk space (#2763, meilisearch/milli#639) @loiclec

canceledBy and error fields are now always part of the task object (#2763) @irevoire, @Kerollmops, and @loiclec

Behavior changes ⚠️

Rename receivedDocumentIds into providedIds in the task object when performing a delete-batch action (#2826) @Ugzuzg

Return a new error when using /keys without any master key set: missing_master_key instead of invalid_api_key (#2922) @vishalsodani

Some names of filter parameters for /tasks and some error codes have been renamed (#3023) @Kerollmops @loiclec

indexUid query parameter is renamed indexUids

type query parameter is renamed types

status query parameter is renamed statuses

invalid_task_type error is renamed invalid_task_types_filter

invalid_task_status error is renamed invalid_task_statuses_filter

Fixes 🐞

Improve error message when adding documents: replace a meaningless serde message (#2819) @onyxcherry

Fix dumpCreation tasks bug (#2890) @washbin

Don't panic when the error length is slightly over 100 (#2727) @onyxcherry

Fix phrase search bug (#2763, meilisearch/milli#639 and meilisearch/milli#647) @loiclec

Fix filtering bug (meilisearch/milli#619) @loiclec

Fix facetDistribution when setting maxFacetPerValue if only a few documents are returned (meilisearch/milli#619) @loiclec

Correct variant returned for invalid_api_key_indexes error when creating an API key (#3032) @ManyTheFish

Prevent phrase search containing stop words never retrieves any documents (#3036 and meilisearch/milli#664) @Samyak2 @Kerollmops

Add missing IN/NOT IN operators to the invalid_filter error message (#3036 and meilisearch/milli#676) @Pranav-yadav @Kerollmops

Filtering by "inf", "infinity", or "NaN" as numbers will yield an internal error (#3036 and meilisearch/milli#689) @dureuill @Kerollmops

Prevent duplicate documents in the search route (#3036 and meilisearch/milli#690) @Kerollmops

Re-introduce missing key in the documents database (#3036 and meilisearch/milli#690) @Kerollmops

Don't remove DB if unreadable (#3077) @dureuill

Fix MDB_BAD_VALSIZE error (#3084 meilisearch/milli#696) @Kerollmops @loiclec

Prevent OS error 22 when creating indexes (#3084 meilisearch/milli#699) @Kerollmops @dureuill

Display the dumpUid as null until we create it (#3122) @Kerollmops

Misc

GitHub CIs and tests

Add dry run for publishing binaries: check if the compilation works (#2726 and #2733) @curquiza

Add CI workflow to update the Meilisearch version in Cargo.toml files (#2741 and #2744) @curquiza

Add CI manifest to automate some steps when closing/creating a Milestone (#2739, #2852, #2853) @curquiza

Improve Docker CI for cloud team (#2790, #2896) @curquiza

Update checkout v2 to v3 in CI manifests and use a unique GitHub PAT (#2740) @curquiza

Uncomment cache steps in Github CI (#2868) @AM1TB

Skip dashboard test if the mini-dashboard feature is disabled (#2814) @jirutka

Use pre-compiled binaries for faster CI (meilisearch/milli#685) @azzamsa

Refactorize the whole test suite (#3085) @irevoire

Dependencies

Upgrade dependencies (#2847) @loiclec

Upgrade clap to 4.0 (#2851) @choznerol

Upgrade to alpine 3.16 in Dockerfile (#2827) @nwnt

Documentation

Update internal CLI documentation (#2839) @jeertmans

Improve issue template to avoid support questions in Meilisearch issues (#2772) @curquiza

Update Hacktoberfest section in CONTRIBUTING.md (#2817, #2794, #2793) @meili-bot @curquiza @Luna-meili

Fix typos (#2789) @kianmeng

Fix broken link in CONTRIBUTING.md (#2845) @AnirudhDaya

Delete v1.rs since it is not included in project (#2834) @Himanshu664

Download-latest script: refactoring (#2913) @nfsec

Make Clippy happy (#2831) @Kerollmops

Improve debuggability by naming spawned threads (#3061) @dureuill

❤️ Thanks again to our external contributors:

Meilisearch: @AM1TB, @AnirudhDaya, @arriven, @choznerol, @Fall1ngStar, @Himanshu664, @jeertmans, @jirutka, @kianmeng, @LunarMarathon, @mlemesle, @mou, @nfsec, @nwnt, @onyxcherry, @Ugzuzg, @vishalsodani and @washbin.

Milli: @anirudhRowjee, @azzamsa, @ehiggs, @jeertmans, @msvaljek, @Pranav-yadav, @Samyak2, @vincent-herlemont, and @vishalsodani.

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.70 MB)
meilisearch-linux-amd64(111.04 MB)
meilisearch-macos-amd64(104.33 MB)
meilisearch-macos-apple-silicon(108.37 MB)
meilisearch-windows-amd64.exe(103.20 MB)
meilisearch.deb(18.20 MB)
v0.30.0rc5(Nov 24, 2022)
Bug fixes

Re-introduced missing_master_key error (https://github.com/meilisearch/meilisearch/pull/3136) @ManyTheFish

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.70 MB)
meilisearch-linux-amd64(111.04 MB)
meilisearch-macos-amd64(104.33 MB)
meilisearch-macos-apple-silicon(108.37 MB)
meilisearch-windows-amd64.exe(103.20 MB)
v0.30.0rc4(Nov 23, 2022)
Bug fixes

Fix linking error when building binaries for aarch64 (#3106) @Kerollmops

Always display deletedDocuments in the IndexDeletion details (#3110) @Kerollmops

Display the dumpUid as null until we create it (#3122) @Kerollmops

Misc

Refactorize the whole test suite (#3085) @irevoire

Update the analytics on the ranking rules (#3114) @irevoire

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.67 MB)
meilisearch-linux-amd64(111.03 MB)
meilisearch-macos-amd64(104.31 MB)
meilisearch-macos-apple-silicon(108.37 MB)
meilisearch-windows-amd64.exe(103.19 MB)
v0.30.0rc3(Nov 21, 2022)
Fix total memory computation (#3096) @dureuill

Source code(tar.gz)
Source code(zip)
meilisearch-linux-amd64(111.03 MB)
meilisearch-macos-amd64(104.31 MB)
meilisearch-macos-apple-silicon(108.36 MB)
meilisearch-windows-amd64.exe(103.19 MB)
v0.30.0rc2(Nov 17, 2022)
Changes

Rename matchedDocuments into providedIds (#3081) @Kerollmops

Rename originalFilters into originalFilter on the cancelation and deletion routes (#3080) @Kerollmops

Add missing analytics (#3056 #3059 #3060 #3063 #3065) @irevoire

Reminder:

We collect telemetry data for only two reasons: so that we can improve our product, and so that we can continue working on this project full-time. In order to create a better product, we need reliable quantitative information. The data we collect helps us fix bugs, evaluate the success of features, and better understand our users' needs. We also need to prove that people are actually using Meilisearch. Usage metrics help us justify our existence to investors so that we can keep this project alive.

Bug fixes 🐞

Don't remove DB if unreadable (#3077) @dureuill

Fix: MDB_BAD_VALSIZE: Unsupported size of key/DB name/data, or wrong DUPFIXED size (#3084 meilisearch/milli#696) @Kerollmops @loiclec

Thread panicked in incremental facets (#3084 meilisearch/milli#696) @Kerollmops @loiclec

Thread Panic from grenad - must be greater than (#3084 meilisearch/milli#697) @Kerollmops @loiclec

Can't create index (OS error 22) (#3084 meilisearch/milli#699) @Kerollmops @dureuill

Fixes regarding the newly introduced features

missing new query filters in cancel-tasks route (#3049) @irevoire

error code on cancel tasks not updated with new filters naming (#3049) @irevoire

task future details fields are not displayed at null when the task status is enqueued (#3067) @ManyTheFish

tasks future details fields are not displayed to 0 when the task reaches a finished state canceled (#3067) @ManyTheFish

Task field originalFilters should contain the first ? char extracted from the path (#3068) @Kerollmops

Misc

Improve debuggability by naming spawned threads (#3061) @dureuill

Source code(tar.gz)
Source code(zip)
meilisearch-linux-amd64(111.03 MB)
meilisearch-macos-amd64(104.31 MB)
meilisearch-macos-apple-silicon(108.36 MB)
meilisearch-windows-amd64.exe(103.19 MB)
v0.29.2(Nov 15, 2022)
🚨 Replaced by v0.29.3

Bug fixes 🐞

Prevent duplicate documents in the search route (#3047 and meilisearch/milli#698) @Kerollmops

Fix missing key in the documents database (#3047 and meilisearch/milli#698) @Kerollmops

⚠️ If you encountered one of these bugs, have already tried to download this version and still have the bug, it means your database is corrupted. The only way you can recover your data is to send your documents again in a fresh index. Using a dump will not work, either. We are sorry for the inconvenience.
Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.30 MB)
meilisearch-linux-amd64(110.34 MB)
meilisearch-macos-amd64(103.70 MB)
meilisearch-windows-amd64.exe(102.42 MB)
meilisearch.deb(17.98 MB)
v0.30.0rc1(Nov 10, 2022)
⚠️ Breaking changes

Some names of filter parameters for /tasks and some error codes have been renamed (#3023) @Kerollmops @loiclec

indexUid query parameter is renamed indexUids

type query parameter is renamed types

status query parameter is renamed statuses

invalid_task_type error is renamed invalid_task_types_filter

invalid_task_status error is renamed invalid_task_statuses_filter

🐞 Bug fixes

Wrong variant returned for invalid_api_key_indexes error when creating an API key (#3032) @ManyTheFish

Phrase search containing stop words never retrieves any documents (#3036 and meilisearch/milli#664) @Samyak2 @Kerollmops

Missing IN / NOT IN operator in the invalid_filter error message (#3036 and meilisearch/milli#676) @Pranav-yadav @Kerollmops

Filtering by the value of "inf" or "infinity" yields an internal error (#3036 and meilisearch/milli#689) @dureuill @Kerollmops

Duplicate documents in the search route (#3036 and meilisearch/milli#690) @Kerollmops

Missing key in the documents database (#3036 and meilisearch/milli#690) @Kerollmops

Fixes regarding the newly introduced features

Fix error code duplicate_index_found error on the index swap route (#3019) @loiclec

In the task object originalQuery has been renamed originalFilters for cancel/deletion tasks

canceledBy filter not implemented on GET /tasks route (#3023) @Kerollmops @loiclec

invalid_task_type is never raised on GET /tasks (#3023) @Kerollmops @loiclec

invalid_task_status is never raised on GET /tasks (#3023) @Kerollmops @loiclec

Store the started_at for a task that is canceled when processing (#3022) @Kerollmops

Thanks to the following persons who have helped us investigate and fix these bugs ❤️ @Arsapol @brentd @ManyTheFish @mason-stewart @qdequele @thijndehaas
Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.72 MB)
meilisearch-linux-amd64(110.96 MB)
meilisearch-macos-amd64(104.26 MB)
meilisearch-macos-apple-silicon(108.30 MB)
meilisearch-windows-amd64.exe(103.09 MB)
v0.30.0rc0(Oct 31, 2022)
⚠️ Since this is a release candidate (RC), we do NOT recommend using it in a production environment. Something not working as expected? We welcome bug reports and feedback about new features.

This release introduces index deployment with zero downtime and adds task management features. It also brings the capability to navigate search results by page selection.

Thanks to Hacktoberfest, tons of updates and bug fixes have been done by the fantastic Meilisearch community. You were more present than ever, and words cannot do justice tojust how grateful we are for your support. A big thank you from the whole Meilisearch team! ❤️

📖 Meilisearch's official documentation will be ready for the official v0.30.0 release. Meanwhile, work-in-progress documentation is available here.

New features and improvements 🔥

Ease search results navigation with pagination

You can now use page and hitsPerPage to fetch the total number of pages and results for a query. These new search parameters can be used in the place of offset and limit when creating interfaces that require knowing the exhaustive number of search results.

The following example fetches the second page of results for a given query:

curl \ -X POST 'http://localhost:7700/indexes/movies/search' \ -H 'Content-Type: application/json' \ --data-binary '{ "q": "shifu", "page": "2", "hitsPerPage": "10" }'

{ "hits": [ // … 10 hits ], // … "page": 2, "hitsPerPage": 10, "totalHits": 2100, "totalPages": 210 }

Done in #2601 by @ManyTheFish.

Cancel processing and enqueued tasks

Meilisearch now allows you to cancel enqueued or processing tasks through a new API route, /tasks/cancel. This can be useful when you need to interrupt a task that is taking too much time to be completed or enqueued by mistake.

To cancel tasks, filters them by specifying one of the following query parameters: uid, status, type, or indexUid.

Tasks can be filtered by date fields:

beforeEnqueuedAt / afterEnqueuedAt

beforeStartedAt / afterStartedAt

beforeFinishedAt / afterFinishedAt

For example, to cancel tasks by uid:

curl \ -X POST 'http://localhost:7700/tasks/cancel?uid=1,2'

{ "taskUid": 1, "indexUid": "movies", "status": "enqueued", "type": "taskCancelation", "enqueuedAt": "2021-08-12T10:00:00.000000Z" }

To cancel all enqueued tasks across all indexes in an instance:

curl \ -X POST 'http://localhost:7700/tasks/cancel?status=enqueued'

Done in #2763 by @irevoire, @Kerollmops and @loiclec.

Delete the task history

This release introduces a new DELETE endpoint to the /tasks route. You can use it to remove finished tasks (succeeded, failed or canceled) from Meilisearch's task history. This can be helpful in reducing the amount of occupied disk space.

To delete finished tasks, filter them by specifying one of the following query parameters: uid, status, type, or indexUid.

Tasks can be filtered by date fields:

beforeEnqueuedAt / afterEnqueuedAt

beforeStartedAt / afterStartedAt

beforeFinishedAt / afterFinishedAt

For example, to delete tasks by uid:

curl \ -X DELETE 'http://localhost:7700/tasks?uid=1,2'

{ "taskUid": 3, "indexUid": null, "status": "enqueued", "type": "taskDeletion", "enqueuedAt": "2021-08-12T10:00:00.000000Z" }

To delete all finished tasks in an instance:

curl \ -X DELETE 'http://localhost:7700/tasks?status=failed,succeeded,canceled'

Done in #2763 by @irevoire, @Kerollmops, and @loiclec.

Deploy indexes with zero downtime

Deploy new indexes version with zero downtime to the search clients. Meilisearch now provides a seamless way to deploy multiple indexes atomically with the new swap indexes API.

To swap indexes, use the new /swap-indexes route:

curl \ -X POST 'http://localhost:7700/swap-indexes'\ -H 'Content-Type: application/json' \ --data-binary '[ { "indexes": ["indexA", "indexB"] } ]'

{ "taskUid": 3, "indexUid": null, "status": "enqueued", "type": "indexSwap", "enqueuedAt": "2021-08-12T10:00:00.000000Z" }

Done in #2763 by @irevoire, @Kerollmops, and @loiclec.

Configuration file support

You can now use a configuration file to set instance options.

By default, Meilisearch will search the working directory for a file named ./config.toml.

You can use the --config-file-path option if you want to keep your configuration file in another directory:

meilisearch --config-file-path="./my-config.toml"

Example of configuration file:

env = "production" master_key = "MY_MASTER_KEY" schedule_snapshot = true

You can download a default configuration file from our repository:

curl https://raw.githubusercontent.com/meilisearch/meilisearch/main/config.toml > config.toml

Done in #2745, #2928, #2804, #2841, and #2961 by @mlemesle, @choznerol, @arriven, @LunarMarathon and @curquiza,

Other improvements

The fresh new filters are also availabe for the GET /tasks: canceledBy, beforeEnqueuedAt, afterEnqueuedAt, beforeStartedAt, afterStartedAt, beforeFinishedAt, afterFinishedAt.

Full support for compressed API requests (Gzip, Brotli, Zlib) (#2876) @mou

Significant indexing speed improvements (#2763, meilisearch/milli#639, meilisearch/milli#619) @loiclec

Provide Apple Silicon binaries (#2837) @jeertmans

Add missing environment variables for dump and snapshot features (#2738) @gmourier

Increase max concurrent readers from 126 to 1024 (#2830) @arriven

Reduce the size of the Meilisearch binary: allow excluding specialized tokenizations when building binaries (#2773) @jirutka

Introduce the snapshotCreation task type: snapshot tasks are added to the task list, for more transparency (#2763) @Kerollmops

Reduce the size taken by Meilisearch on the disk space (#2763, meilisearch/milli#639) @loiclec

Behavior changes ⚠️

Rename receivedDocumentIds into matchedDocuments in the task object when performing a delete-batch action (#2826) @Ugzuzg

Change default bind address from 127.0.0.1 to localhost, making the default address available in IPv4 and IPv6 (#2861) @Fall1ngStar

Return a new error when using /keys without any master key set: missing_master_key instead of invalid_api_key (#2922) @vishalsodani

Fixes 🐞

Improve error message when adding documents: replace a meaningless serde message (#2819) @onyxcherry

Fix dumpCreation tasks bug (#2890) @washbin

Don't panic when the error length is slightly over 100 (#2727) @onyxcherry

Fix phrase search bug (meilisearch/milli#639 and meilisearch/milli#647) @loiclec

Fix filtering bug (meilisearch/milli#619) @loiclec

Fix facetDistribution when setting maxFacetPerValue if only a few documents are returned (meilisearch/milli#619) @loiclec

Misc

GitHub CIs and tests

Add dry run for publishing binaries: check if the compilation works (#2726 and #2733) @curquiza

Add CI workflow to update the Meilisearch version in Cargo.toml files (#2741 and #2744) @curquiza

Add CI manifest to automate some steps when closing/creating a Milestone (#2739, #2852, #2853) @curquiza

Improve Docker CI for cloud team (#2790, #2896) @curquiza

Update checkout v2 to v3 in CI manifests and use a unique GitHub PAT (#2740) @curquiza

Uncomment cache steps in Github CI (#2868) @AM1TB

Skip dashboard test if the mini-dashboard feature is disabled (#2814) @jirutka

Dependencies

Upgrade dependencies (#2847) @loiclec

Upgrade clap to 4.0 (#2851) @choznerol

Upgrade to alpine 3.16 in Dockerfile (#2827) @nwnt

Documentation

Update internal CLI documentation (#2839) @jeertmans

Improve issue template to avoid support questions in Meilisearch issues (#2772) @curquiza

Update Hacktoberfest section in CONTRIBUTING.md (#2817, #2794, #2793) @meili-bot @curquiza @Luna-meili

Fix typos (#2789) @kianmeng

Fix broken link in CONTRIBUTING.md (#2845) @AnirudhDaya

Delete v1.rs since it is not included in project (#2834) @Himanshu664

Download-latest script: refactoring (#2913) @nfsec

Make Clippy happy (#2831) @Kerollmops

❤️ Thanks again to our external contributors:

Meilisearch: @AM1TB, @AnirudhDaya, @arriven, @choznerol, @Fall1ngStar, @Himanshu664, @jeertmans, @jirutka, @kianmeng, @LunarMarathon, @mlemesle, @mou, @nfsec, @nwnt, @onyxcherry, @Ugzuzg, @vishalsodani and @washbin.

Milli: @anirudhRowjee, @ehiggs, @jeertmans, @msvaljek, @vincent-herlemont, and @vishalsodani.

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.74 MB)
meilisearch-linux-amd64(111.04 MB)
meilisearch-macos-amd64(104.55 MB)
meilisearch-macos-apple-silicon(107.69 MB)
meilisearch-windows-amd64.exe(103.04 MB)
v0.29.1(Oct 10, 2022)
⚠️ Use this release instead of v0.29.0 if you have compatibility issues with glibc, see this issue.

Bug fixes 🐞

Make Meilisearch compatible with libc 2.27 again (so Debian 10): use Ubuntu 18.04 for all CI tasks that previously used Ubuntu 20.04 (https://github.com/meilisearch/meilisearch/pull/2862) @loiclec

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.26 MB)
meilisearch-linux-amd64(110.41 MB)
meilisearch-macos-amd64(103.99 MB)
meilisearch-windows-amd64.exe(102.40 MB)
meilisearch.deb(18.04 MB)
v0.29.0(Oct 3, 2022)
⚠️ If you have a compatibility issue with libc (for example, with Debian 10) use v0.29.1 instead. See this issue.

This release includes a much-requested feature: an option to require all query terms to be present in search results. The absence of this feature prevented some users from upgrading to newer Meilisearch versions. Additionally, v0.29 contains several indexing speed improvements.

🧰 Most integrations (SDKs, clients, libraries) will be compatible with this version within four hours of the release. Sometimes this can take up to 24 hours, depending on the issues we encounter during the release.

Here is the exhaustive list of integrations not compatible with v0.29.0:

meilisearch-dart and meilisearch-swift: though you can use still use both libraries with v0.29.0, the matchingStrategy feature is not yet available. Let us know if you really need it—and contributions are always welcome!

meilisearch-java: Our team is still working on refactoring the code, and the SDK cannot be used with v0.29.0. Follow the repository to know when it will be compatible.

New features and improvements 🔥

Customize matching strategy

It is now possible to control how Meilisearch matches documents during search by choosing to return only documents containing all query terms.

How to activate it?

curl \ -X POST 'http://localhost:7700/indexes/movies/search' \ -H 'Content-Type: application/json' \ --data-binary '{ "q": "american ninja", "matchingStrategy": "all" }'

By setting "matchingStrategy": "all" In the above example, we ensure Meilisearch will only return documents containing both "american" AND "ninja".

Meilisearch's default matching strategy is last. It will still return documents containing all the query words, followed by documents that don't contain all the query words.

Check out the documentation for more information on matchingStrategy.

Made by @ManyTheFish in #2578 and meilisearch/milli#598.

Thai language support

Meilisearch now offers optimized support for Thai 🇹🇭

Want optimized support for your language? Consider contributing to the Meilisearch tokenizer by following the CONTRIBUTING.md file and integrating your intended normalizer/segmenter.

The core team is always happy to help our contributors. Do drop us a line if you need us to clarify anything!

Made by our dear contributor @aFluffyHotdog and by @ManyTheFish in meilisearch/charabia#114, meilisearch/charabia#115, meilisearch/milli##611, and #2636.

Performance improvements for indexing

We improved Meilisearch's indexing speed, especially for indexes using filterableAttributes and when updating and deleting documents.

For more information on the main improvements, check out the following issues: meilisearch/product#513 and meilisearch/product#520.

Made by @loiclec and @irevoire in #2689, #2690, meilisearch/milli#607, meilisearch/milli#606, meilisearch/milli#587, meilisearch/milli#557,meilisearch/milli#579, and meilisearch/milli#590.

Filtering: introduce the EXISTS and IN operators

You can now use the EXISTS and IN operators with the filter search parameter during search.

With the following set of documents:

{ "id": 1, "product": "T-shirt", "price": 20, "color": "yellow" } { "id": 2, "product": "T-shirt", "color": "red" } { "id": 3, "product": "T-shirt", "price": 30 } { "id": 4, "product": "T-shirt", "price": 40 }

The filter price EXISTS will select documents 1, 3, and 4.

The filter price IN [30, 20] will select documents 1 and 3.

These operators can be combined with the existing NOT (!=) operator. Read more about filtering in Meilisearch's documentation.

Made by @loiclec in #2636, meilisearch/milli#596, and meilisearch/milli#556.

Key management

When managing keys, the * helper can now be used to express a set of actions in the actions array.

For example, instead of specifying indexes.create, indexes.get, indexes.update, indexes.delete; you can now use indexes.*.

Made by our dear contributors @Mcdostone and @phdavis1027 in #2080 and #2579.

Behavior changes ⚠️

Auto-batching

This previously experimental feature is now enabled by default.

If you enabled auto-batching in previous versions using --enable-auto-batching, remove this flag when running Meilisearch v0.29.0. All other auto-batching settings (--debounce-duration-sec, --max-batch-size, and --max-documents-per-batch) have been removed.

If you did not enable auto-batching before, the transition should be seamless and increase indexing speed 🏎️

If you experience any issues during indexing, try using --disable-auto-batching flag (or the environment variable MEILI_DISABLE_AUTO_BATCHING).

Made by @Kerollmops in meilisearch/milli#561, #2523, and #2677.

!= (NOT) filter operator

This behavior change is consistent with the EXISTS and IN operators. The NOT filter now also matches documents missing the specified attributes. More explanation with examples here.

Made by @loiclec in #2636, and meilisearch/milli#596.

Fixes 🐞

Clamp max size of databases to a multiple of system page size (#2662) @ManyTheFish

Prevent whitespace trimming during document id validation (#2636 and meilisearch/milli#599) @ManyTheFish

Support https in print_launch_resume (#2664) @evpeople

Filters: reserved keywords must wrapped in quotation marks when used field names/values (#2636 and meilisearch/milli#596) @loiclec

Fix filtering selection (#2636 and meilisearch/milli#568) @GraDKh

Misc

Completely rework the README (#2504) @CaroFG, @curquiza, @guimachiavelli, and @gmourier

Add analytics to the /stats routes (#2674) @irevoire

Update link to Cloud beta form (#2625) @davelarkan

Internal feature: integrate /metrics route (#2657) @pavo-tusker

Fix typo (#2508, #2517, #2556) @ryanrussell

Use nightly for cargo fmt in CI (#2519) @curquiza

Fix content of dump/assets for testing (#2544) @loiclec

Add Dependabot for GHA (#2561) @curquiza

Update Docker credentials (#2539) @curquiza

Add new CI: cron GHA to create an issue "Update dependencies" (#2585) @VasiliySoldatkin

Add a dedicated section for Language Support in the issue template (#2592) @ManyTheFish

❤️ Thanks again to our external contributors:

Meilisearch: @evpeople, @Mcdostone, @pavo-tusker, @phdavis1027, @ryanrussell, and @VasiliySoldatkin.

Milli: @GraDKh.

Charabia: @aFluffyHotdog, @matthias-wright, and @mosuka.

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.26 MB)
meilisearch-linux-amd64(110.94 MB)
meilisearch-macos-amd64(103.99 MB)
meilisearch-windows-amd64.exe(102.40 MB)
meilisearch.deb(18.02 MB)
v0.29.0rc4(Sep 14, 2022)
🐞Bug fixes

Fix indexing issues when too many FDs were opened: now throw an error instead of faking a succeeded indexing @Kerollmops

Update dependencies to remove potential security issues #2768 @curquiza

Big thanks to @gmourier for the bug report! ❤️
Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.30 MB)
meilisearch-linux-amd64(110.93 MB)
meilisearch-macos-amd64(104.02 MB)
meilisearch-windows-amd64.exe(102.39 MB)
v0.29.0rc3(Sep 8, 2022)
Improvements

Update mini-dashboard version to v0.2.2 #2755 @mdubus

Update ubuntu version from 18.04 to 20.04 in GHA #2758 @curquiza

Bug fixes

Fix API key bugs (#2730 and #2731 issues) #2734 @ManyTheFish

Fix indexing bug introduced in v0.29.0 with huge words in documents #2759 and meilisearch/milli#628 @Kerollmops

Special thanks to @Northn and @bidoubiwa for the bug reports 🙏
Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.28 MB)
meilisearch-linux-amd64(110.89 MB)
meilisearch-macos-amd64(104.00 MB)
meilisearch-windows-amd64.exe(102.36 MB)
v0.29.0rc2(Sep 1, 2022)
Bug fixes 🐞

Make the document addition done log to appear once indexing is over (#2724) @evpeople

Fix the unexpected response in a simple search (#2732) @ManyTheFish

CSV document indexing is now working properly when the id is a number (#2732) @Kerollmops

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.28 MB)
meilisearch-linux-amd64(110.35 MB)
meilisearch-macos-amd64(103.98 MB)
meilisearch-windows-amd64.exe(102.35 MB)
v0.29.0rc1(Aug 30, 2022)
🐞 Bug fixes

Make the aarch64 available: disable LTO (#2717) @loiclec

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(114.33 MB)
meilisearch-linux-amd64(110.38 MB)
meilisearch-macos-amd64(103.88 MB)
meilisearch-windows-amd64.exe(102.30 MB)
v0.29.0rc0(Aug 29, 2022)
⚠️ Since this is a release candidate (RC), we do NOT recommend using this version in production. This is only for test purposes. Please, feel free to report bugs and give feedback about the new feature introductions.

In this release, we integrated a demanded feature: requiring all query terms to be present in search results. Some users have been stuck in the pre-0.21 versions because of this. We also improved the indexing speed.

📖 the official documentation is not updated yet, but will be ready for the official v0.29.0 release. The work-in-progress documentation is available here.

New features and improvements 🔥

Customize matching strategy

It's now possible to control how Meilisearch matches documents during search by choosing to return only documents containing all the query words.

How to activate it?

curl \ -X POST 'http://localhost:7700/indexes/movies/search' \ -H 'Content-Type: application/json' \ --data-binary '{ "q": "american ninja", "matchingStrategy": "all" }'

Here Meilisearch will only return documents containing "american" AND "ninja".

By default, Meilisearch will still return documents containing all the query words, followed by documents that don't contain all the query words.

Check out the documentation.

Made by @ManyTheFish in #2578 and meilisearch/milli#598.

Thai language support

Thai is now supported 🇹🇭

If your language is missing in Meilisearch, please consider contributing to the Meilisearch tokenizer by following the CONTRIBUTING.md file and integrating your intended normalizer/segmenter.

The whole core team is available to help you contribute. Feel free to ask for any clarifications!

Made by our dear contributor @aFluffyHotdog and by @ManyTheFish in meilisearch/charabia#114, meilisearch/charabia#115, meilisearch/milli##611, and #2636.

Performance improvements for indexing

We improved the indexing speed, especially for filterableAttributes and document update and deletion.

For more information on the main improvements, check out the following issues: meilisearch/product#513 and meilisearch/product#520.

Made by @loiclec and @irevoire in #2689, #2690, meilisearch/milli#607, meilisearch/milli#606, meilisearch/milli#587, meilisearch/milli#557,meilisearch/milli#579, and meilisearch/milli#590.

Filtering: introduce the EXISTS and IN operators

You can now use the EXISTS and IN operators with the filter search parameter during search.

With the following set of documents:

{ "id": 1, "product": "T-shirt", "price": 20, "color": "yellow" } { "id": 2, "product": "T-shirt", "color": "red" } { "id": 3, "product": "T-shirt", "price": 30 } { "id": 4, "product": "T-shirt", "price": 40 }

The filter price EXISTS will select documents 1, 3, and 4.

The filter price IN [30, 20] will select documents 1 and 3.

These operators can be combined with the existing NOT (!=) operator. More about filtering.

Made by @loiclec in #2636, meilisearch/milli#596, and meilisearch/milli#556.

Key management

When managing keys, the * helper can now be used to express a set of actions in the actions array.

For example, instead of specifying indexes.create, indexes.get, indexes.update, indexes.delete; you can now use indexes.*.

Made by our dear contributors @Mcdostone and @phdavis1027 in #2080 and #2579.

Behavior changes ⚠️

Auto-batching

This previously experimental feature is now activated by default.

For people who activated it in previous versions using --enable-auto-batching, you need to remove this flag when running Meilisearch. All the previous settings (--debounce-duration-sec, --max-batch-size and --max-documents-per-batch) have been removed.

For people who hadn't activated it, it does not change anything but increases the indexing speed 🏎️

If there are any issues during indexing, the --disable-auto-batching flag (or the environment variable MEILI_DISABLE_AUTO_BATCHING) is available.

Made by @Kerollmops in meilisearch/milli#561, #2523, and #2677.

!= (NOT) filter operator

This behavior change is consistent with the EXISTS and IN operators. The NOT filter now also matches documents that don't have the related attributes. More explanation with examples here.

Made by @loiclec in #2636, and meilisearch/milli#596.

Fixes 🐞

Clamp databases max size to a multiple of system page size (#2662) @ManyTheFish

Prevent removing whitespace trimming during document id validation (#2636 and meilisearch/milli#599) @ManyTheFish

Support https in print_launch_resume (#2664) @evpeople

Filters can not accept unquoted reserved keywords as field names/values (#2636 and meilisearch/milli#596) @loiclec

Fix filtering selection (#2636 and meilisearch/milli#568) @GraDKh

Misc

Completely rework the README (#2504) @CaroFG, @curquiza, @guimachiavelli, and @gmourier

Add analytics to the /stats routes (#2674) @irevoire

Update link to Cloud beta form (#2625) @davelarkan

Internal feature only: integrate /metrics route (#2657) @pavo-tusker

Fix typo (#2508, #2517, #2556) @ryanrussell

Use nightly for cargo fmt in CI (#2519) @curquiza

Fix content of dump/assets for testing (#2544) @loiclec

Add Dependabot for GHA (#2561) @curquiza

Update Docker credentials (#2539) @curquiza

Add new CI: cron GHA to create an issue "Update dependencies" (#2585) @VasiliySoldatkin

Add a dedicated section for Language Support in the issue template (#2592) @ManyTheFish

❤️ Thanks again to our external contributors:

Meilisearch: @evpeople, @Mcdostone, @pavo-tusker, @phdavis1027, @ryanrussell, and @VasiliySoldatkin.

Milli: @GraDKh.

Charabia: @aFluffyHotdog, @matthias-wright, and @mosuka.

Source code(tar.gz)
Source code(zip)
meilisearch-linux-amd64(109.53 MB)
meilisearch-macos-amd64(104.19 MB)
meilisearch-windows-amd64.exe(103.16 MB)
v0.28.1(Jul 21, 2022)
⚠️ If possible, use v0.28.1 instead of v0.28.0 to include the following fixes:

Update mini-dashboard to v0.2.1 (#2631) @mdubus

Fix: all the words are now highlighted in _formatted when using attributesToCrop + attributesToHighlight (https://github.com/meilisearch/milli/pull/592 and #2633) @ManyTheFish

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(124.85 MB)
meilisearch-linux-amd64(120.38 MB)
meilisearch-macos-amd64(106.65 MB)
meilisearch-windows-amd64.exe(103.13 MB)
meilisearch.deb(17.99 MB)
v0.28.0(Jul 11, 2022)
⚠️ If possible, use v0.28.1 instead of this version (v0.28.0) to include the highlight fix available in v0.28.1

The main goal of this new release is to stabilize the Meilisearch API 🚀

This release introduces a lot of breaking changes (especially regarding names), and we are sorry for this. The purpose is to guarantee what we care about the most—providing the best search engine experience—by stabilizing the Meilisearch API. There should not be any breaking changes to the API in the future.

New features 🚀

Smart crop

The smart crop is an improvement of the Meilisearch crop.

The cropping algorithm tries to find the window(s) with the highest density of query terms within the cropLength limit. From these, it picks the window that contains the most query terms in order.

Done by @ManyTheFish (#2408)

API key management: deterministic keys

It is now possible to create a deterministic key value by specifying a uid field at creation. uid value must be a UUID v4. If omitted, Meilisearch generates it.

The key field is generated by hashing the master key and the uid:

echo -n $HYPHENATED_UUID | openssl dgst -sha256 -hmac $MASTER_KEY

As a result, key values are deterministic between instances sharing the same configuration. Since the key field depends on the master key, it is not propagated to dumps and snapshots. If a malicious user ever gets access to your dumps or snapshots, they will not have access to your instance's API keys.

Adds a name field: specify a human-readable name to make API key retrieval more convenient.

Introduces new actions to manage API keys (keys.get, keys.create, keys.update, keys.delete).

It is no longer possible to update an API key's actions, indexes, or expiresAt after creation.

Done by @ManyTheFish in #2438, #2451, #2475, and #2474

Hebrew language support

New language support: Hebrew, using the Niqqud library.

Done by the amazing @benny-n in meilisearch/charabia#101.

New tokenizer contribution experience: add your language to Meilisearch

The contribution experience in the tokenizer has been significantly improved.

If your language is missing in Meilisearch, please consider contributing to the Meilisearch tokenizer by following the CONTRIBUTING.md file and integrating your intended normalizer/segmenter.

The whole core team is available to help you contribute. Feel free to ask for any clarifications you need!

Done by @ManyTheFish, @loiclec @gmourier, and @curquiza.

Customizable limits

Limit of facet values

When searching using the facets search parameter, Meilisearch will now return a maximum of 100 different facet values per faceted attribute.

This default limit is customizable using the new faceting setting, accessible via the /indexes/{uid}/settings/faceting route.

To increase this limit to 3000:

curl \ -X PATCH 'http://localhost:7700/indexes/books/settings/faceting' \ -H 'Content-Type: application/json' \ --data-binary '{ "maxValuesPerFacet": 3000 }'

You can also get the value using the HTTP GET verb, or reset the value using the DELETE verb.

Done by @Kerollmops in meilisearch/milli#550 and #2494

Limit of documents returned by search

When making a search request, Meilisearch returns a maximum of 1000 documents: the documents deemed most relevant to the query.

This default limit is now customizable using the pagination setting, accessible via the /indexes/{uid}/settings/pagination route.

To increase this limit to 2000:

curl \ -X PATCH 'http://localhost:7700/indexes/books/settings/pagination' \ -H 'Content-Type: application/json' \ --data-binary '{ "maxTotalHits": 2000 }'

You can also get the value using the HTTP GET verb, or reset the value using the DELETE verb.

Done by @Kerollmops in meilisearch/milli#550 and #2494

Breaking changes and improvement ⚠️

For most of the following changes, if you use one of our SDKs, these changes will be transparent for you by only updating to the latest version of the SDK.

Changes in /indexes/{uid}/search endpoint

nbHits is renamed estimatedTotalHits. Some users were confused by the old name and used it for their pagination, which we discourage. Please check out this fresh new guide to learn how to paginate with Meilisearch without using nbHits.

The facetsDistribution search parameter is renamed facets

The facetsDistribution field in the response body is renamed facetDistribution

The matches search parameter is renamed showMatchesPosition

The _matchesInfo field in the response body is renamed _matchesPosition

exhaustiveNbHits and exhaustiveFacetsCount are removed. Meilisearch does not return exhaustive hit or facet counts at this time.

Done by @irevoire in #2406.

Changes in /indexes endpoints

When getting the index list (GET /indexes), the response is now paginated. That means:

You can use the offset and limit parameters to browse indexes.

The indexes are returned in the results field in the response body.

limit, offset, and total now appear in the response body.

The endpoint to update an index is now PATCH /indexes/{indexUid} instead of PUT

The already deprecated name field has now been removed entirely.

Done by @irevoire and @Kerollmops in #2423, #2452, and #2454.

Changes in indexes/{uid}/documents endpoints

When getting the documents list (GET /indexes/{uid}/documents), pagination has been improved. That means:

You can still use the offset and limit parameters to browse documents.

The documents are returned in the results field in the response body.

limit, offset, and total now appear in the response body.

When getting documents via the GET /indexes/{uid}/documents route, the attributesToRetrieve parameter has been renamed fields

The fields parameter is also available for the GET /indexes/{uid}/documents/{document_uid} endpoint.

The displayedAttributes setting does not impact GET indexes/{uid}/documents anymore.

Done by @irevoire in #2372, #2433, and #2454.

Changes in /tasks and /indexes/{uid}/tasks endpoints

You can now filter on the GET /tasks endpoint:

Filtering is possible on the type, status, and indexUid fields.

Use the , character to specify several values for a filter, e.g. ?status=enqueued,processing. This is equivalent to an OR operation.

Use the & character to apply multiple filters, e.g. ?status=enqueued&type=indexCreation. This is equivalent to an AND operation.

Add keyset pagination on the GET /tasks endpoint, allowing users to browse multiple sets of task items.

Use the limit and from query parameters to browse tasks.

limit, from, and next now appear in the response body.

GET /indexes/:indexUid/tasks has been removed. It can be replaced by GET /tasks?indexUid=:indexUid

GET /indexes/:indexUid/tasks/:taskUid has been removed.

uid has been renamed taskUid in the response body returned by every asynchronous task (e.g. index creation, document addition, etc...)

Some task properties have been renamed

documentPartial-> documentAdditionOrUpdate

documentAddition-> documentAdditionOrUpdate

clearAll -> documentDeletion

Done by Kerollmops in #2399 and #2445.

Changes regarding dumps

Dump creation is now an asynchronous task.

Dump tasks are placed at the end of the task queue, but processed in priority.

In other words, a dump will run as soon as the current task is finished processing.

The dump's uid will reflect when it was enqueued in relation to the other tasks, not when it was processed.

POST /dumps now returns 202 like any other asynchronous operation (e.g. index creation, document addition...)

The GET /dumps/:dump_uid/status endpoint has been removed. It can be replaced by GET /tasks?type=dumpCreation.

The dump_already_processing error has been removed since dumps can now be enqueued.

Done by @MarinPostma in #2410

Changes in /keys endpoint

Tenant tokens: apiKeyPrefix has been renamed apiKeyUid.

It is no longer possible to update actions, indexes, or expiresAt after key creation. Only the name and description fields are updatable.

When getting the keys list (GET /keys), the response is now paginated. That means:

You can use the offset and limit parameters to browse keys.

limit, offset, and total now appear in the response body.

Due to changes to the keys resource, dumps are not fully compatible with API keys. As a result, keys imported from an old version of Meilisearch will have their key and uid fields regenerated. In other words, you will need to update your keys. Other key fields such as actions and indexes will not be affected.

The dumps.get action has been removed.

Done by @ManyTheFish in #2438, #2451, #2475, and #2474)

Changes in /indexes/{uid}/settings endpoints

The following changes only concern the HTTP verb associated with a given method.

The endpoint to update the global settings is now PATCH /indexes/{indexUid}/settings instead of POST

The endpoint to update the typo tolerance settings is now PATCH /indexes/{indexUid}/settings/typo-tolerance instead of POST

The endpoint to update the displayed attributes settings is now PUT /indexes/{indexUid}/settings/displayed-attributes instead of POST

The endpoint to update the distinct attribute settings is now PUT /indexes/{indexUid}/settings/distinct-attribute instead of POST

The endpoint to update the filterable attribute settings is now PUT /indexes/{indexUid}/settings/filterable-attributes instead of POST

The endpoint to update the ranking rules settings is now PUT /indexes/{indexUid}/settings/ranking-rules instead of POST

The endpoint to update the searchable attributes settings is now PUT /indexes/{indexUid}/settings/searchable-attributes instead of POST

The endpoint to update the sortable attributes settings is now PUT /indexes/{indexUid}/settings/sortable-attributes instead of POST

The endpoint to update the stop words settings is now PUT /indexes/{indexUid}/settings/stop-words instead of POST

The endpoint to update the synonyms settings is now PUT /indexes/{indexUid}/settings/synonyms instead of POST

Done by @Kerollmops in #2452

Other improvements 🔨

Geosearch: Support string type for _geo, lat, and lng fields (#2408 and meilisearch/milli#523) @irevoire

Improve ranking rules error message (meilisearch/milli#536, #2468) @matthias-wright

Improve docker CI: push vX.Y tag (without patch) to Docker Hub (#2529) @janithpet

Remove the connection keep-alive timeout (#2471) @Thearas

Improve index uid validation upon API key creation (#2414) @pierre-l

Improve estimatedTotalHits behavior with distinct attribute (#2546) @Kerollmops

Fixes 🐞

Fix filtering on documents with double and/or single quote characters (#2500 and meilisearch/milli#552) @irevoire

Misc

Telemetry: handle the X-Meilisearch-Client header (#2374) @irevoire

Telemetry: send the analytics to telemetry.meilisearch.com instead of segment (#2429) @irevoire

Create custom error types for TaskType, TaskStatus, and IndexUid (#2460) @walterbm

Improve the Docker CIs (#2477 and #2491) @choznerol and @curquiza

Some internal changes (#2447, #2464, #2453, #2455) @MarinPostma @Kerollmops

Update dependencies (#2450) @Kerollmops

Add dump tests in CI (#2357) @irevoire

Deny warnings in CI (#2339) @MarinPostma

Update mini-dashboard for v0.28.0 (#2587) @mdubus @bidoubiwa

Change Nelson path (#2347) @curquiza

Re-integrate Clippy in the CI (#2424) @irevoire

Fix typo in README (#2448) @ryanrussell

Fix typo in code base comments (#2459, and #2508) @ryanrussell

Update url of movies.json in README (#2422) @0x0x1

Fix command in README (#2434) @0x0x1

❤️ Thanks again to our external contributors:

Meilisearch: @0x0x1, @choznerol, @janithpet, @pierre-l, @ryanrussell, @Thearas, and @walterbm

Milli: @matthias-wright

Charabia: @benny-n

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(124.91 MB)
meilisearch-linux-amd64(120.50 MB)
meilisearch-macos-amd64(106.76 MB)
meilisearch-windows-amd64.exe(103.26 MB)
meilisearch.deb(18.00 MB)
v0.28.0rc4(Jul 7, 2022)
Improvements

Add missing analytics (#2591) @Kerollmops

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(124.33 MB)
meilisearch-linux-amd64(120.53 MB)
meilisearch-macos-amd64(106.78 MB)
meilisearch-windows-amd64.exe(103.25 MB)
v0.28.0rc3(Jul 6, 2022)
Improvements

Format API keys in hexa instead of base64 (#2584) @ManyTheFish

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(124.33 MB)
meilisearch-linux-amd64(120.53 MB)
meilisearch-macos-amd64(106.78 MB)
meilisearch-windows-amd64.exe(103.25 MB)
v0.28.0rc2(Jul 6, 2022)
Fixes

Update mini-dashboard for v0.28.0 #2587 @mdubus @bidoubiwa

Misc

Add more tests on the formatted route #2557 @irevoire

Fix Clippy #2586 @Kerollmops

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(124.31 MB)
meilisearch-linux-amd64(120.52 MB)
meilisearch-macos-amd64(106.81 MB)
meilisearch-windows-amd64.exe(103.24 MB)
v0.28.0rc1(Jun 27, 2022)
Fixes / Improvements

* not working on GET /search route for attributesToHighlight/attributesToCrop (#2527) @irevoire

Cannot access pagination and faceting settings via the settings sub routes (#2524) @Kerollmops

Smart crop add an extra space in _formatted field (#2546) @ManyTheFish and @Kerollmops

Rename settings limitedTo into maxTotalHits (#2546) @Kerollmops

Wrong nbHits when using distinct attribute (#2546) @Kerollmops

Misc

Check the version in Cargo.toml before publishing (#2530) @curquiza

Publish Docker image tag without the patch version -> v0.28 docker tag will be available! (#2507 and #2529) @janithpet @curquiza

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(124.46 MB)
meilisearch-linux-amd64(120.74 MB)
meilisearch-macos-amd64(106.93 MB)
meilisearch-windows-amd64.exe(103.38 MB)
v0.28.0rc0(Jun 13, 2022)
The main goal of this new release is to stabilize the Meilisearch API 🚀

This release introduces a lot of breaking changes (especially regarding names), and we are sorry for this. The purpose is to guarantee what we care about the most—providing the best search engine experience—by stabilizing the Meilisearch API. There should not be any breaking changes to the API in the future.

⚠️ Since this is a release candidate (RC), the official documentation is not updated yet, but will be ready for the official v0.28.0 release. The work-in-progress documentation is available here.

New features 🚀

Smart crop

The smart crop is an improvement of the Meilisearch crop.

The cropping algorithm tries to find the window(s) with the highest density of query terms within the cropLength limit. From these, it picks the window that contains the most query terms in order.

Done by @ManyTheFish (#2408)

API key management: deterministic keys

It is now possible to create a deterministic key value by specifying a uid field at creation. uid value must be a UUID v4. If omitted, Meilisearch generates it.

Adds a name field: specify a human-readable name to make API key retrieval more convenient.

Introduces new actions to manage API keys (keys.get, keys.create, keys.update, keys.delete).

It is no longer possible to update an API key's actions, indexes, or expiresAt after creation.

Done by @ManyTheFish in #2438, #2451, #2475, and #2474)

Hebrew language support

New language support: Hebrew, using the Niqqud library.

Done by the amazing @benny-n in meilisearch/charabia#101.

New tokenizer contribution experience: add your language to Meilisearch

The contribution experience in the tokenizer has been significantly improved.

If your language is missing in Meilisearch, please consider contributing to the Meilisearch tokenizer by following the CONTRIBUTING.md file and integrating your intended normalizer/segmenter.

The whole core team is available to help you contribute. Feel free to ask for any clarifications you need!

Done by @ManyTheFish, @loiclec @gmourier, and @curquiza.

Customizable limits

Limit of facet values

When searching using the facets search parameter, Meilisearch will now return a maximum of 100 different facet values per faceted attribute.

This default limit is customizable using the new faceting setting, accessible via the /indexes/{uid}/settings/faceting route.

To increase this limit to 3000:

curl \ -X PATCH 'http://localhost:7700/indexes/books/settings/faceting' \ -H 'Content-Type: application/json' \ --data-binary '{ "maxValuesPerFacet": 3000 }'

You can also get the value using the HTTP GET verb, or reset the value using the DELETE verb.

Done by @Kerollmops in meilisearch/milli#550 and #2494

Limit of documents returned by search

When making a search request, Meilisearch returns a maximum of 1000 documents: the documents deemed most relevant to the query.

This default limit is now customizable using the pagination setting, accessible via the /indexes/{uid}/settings/pagination route.

To increase this limit to 2000:

curl \ -X PATCH 'http://localhost:7700/indexes/books/settings/pagination' \ -H 'Content-Type: application/json' \ --data-binary '{ "limitedTo": 2000 }'

You can also get the value using the HTTP GET verb, or reset the value using the DELETE verb.

Done by @Kerollmops in meilisearch/milli#550 and #2494

Breaking changes and improvement ⚠️

Changes in /indexes/{uid}/search endpoint

nbHits is renamed estimatedTotalHits. Some users were confused by the old name and used it for their pagination, which we discourage.

The facetsDistribution search parameter is renamed facets

The facetsDistribution field in the response body is renamed facetDistribution

The matches search parameter is renamed showMatchesPosition

The _matchesInfo field in the response body is renamed _matchesPosition

exhaustiveNbHits and exhaustiveFacetsCount are removed. Meilisearch does not return exhaustive hit or facet counts at this time.

Done by @irevoire in #2406.

Changes in /indexes endpoints

When getting the index list (GET /indexes), the response is now paginated. That means:

You can use the offset and limit parameters to browse indexes.

The indexes are returned in the results field in the response body.

limit, offset, and total now appear in the response body.

The endpoint to update an index is now PATCH /indexes/{indexUid} instead of PUT

The already deprecated name field has now been removed entirely.

Done by @irevoire and @Kerollmops in #2423, #2452, and #2454.

Changes in indexes/{uid}/documents endpoints

When getting the documents list (GET /indexes/{uid}/documents), pagination has been improved. That means:

You can still use the offset and limit parameters to browse documents.

The documents are returned in the results field in the response body.

limit, offset, and total now appear in the response body.

When getting documents via the GET /indexes/{uid}/documents route, the attributesToRetrieve parameter has been renamed fields

The fields parameter is also available for the GET /indexes/{uid}/documents/{document_uid} endpoint.

The displayedAttributes setting does not impact GET indexes/{uid}/documents anymore.

Done by @irevoire in #2372, #2433, and #2454.

Changes in /tasks and /indexes/{uid}/tasks endpoints

You can now filter on the GET /tasks endpoint:

Filtering is possible on the type, status, and indexUid fields.

Use the , character to specify several values for a filter, e.g. ?status=enqueued,processing. This is equivalent to an OR operation.

Use the & character to apply multiple filters, e.g. ?status=enqueued&type=indexCreation. This is equivalent to an AND operation.

Add keyset pagination on the GET /tasks endpoint, allowing users to browse multiple sets of task items.

Use the limit and from query parameters to browse tasks.

limit, from, and next now appear in the response body.

GET /indexes/:indexUid/tasks has been removed. It can be replaced by GET /tasks?indexUid=:indexUid

GET /indexes/:indexUid/tasks/:taskUid has been removed.

uid has been renamed taskUid in the response body returned by every asynchronous task (e.g. index creation, document addition, etc...)

Some task properties have been renamed

documentPartial-> documentAdditionOrUpdate

documentAddition-> documentAdditionOrUpdate

clearAll -> documentDeletion

Done by Kerollmops in #2399 and #2445.

Changes regarding dumps

Dump creation is now an asynchronous task.

Dump tasks are placed at the end of the task queue, but processed in priority.

In other words, a dump will run as soon as the current task is finished processing.

The dump's uid will reflect when it was enqueued in relation to the other tasks, not when it was processed.

POST /dumps now returns 202 like any other asynchronous operation (e.g. index creation, document addition...

The GET /dumps/:dump_uid/status endpoint has been removed. It can be replaced by GET /tasks?type=dumpCreation.

The dump_already_processing error has been removed since dumps can now be enqueued.

Done by @MarinPostma in #2410

Changes in /keys endpoint

Tenant tokens: apiKeyPrefix has been renamed apiKeyUid.

It is no longer possible to update actions, indexes, or expiresAt after key creation. Only the name and description fields are updatable.

When getting the keys list (GET /keys), the response is now paginated. That means:

You can use the offset and limit parameters to browse keys.

limit, offset, and total now appear in the response body.

Due to changes to the keys resource, dumps are not fully compatible with API keys. As a result, keys imported from an old version of Meilisearch will have their key and uid fields regenerated. In other words, you will need to update your keys. Other key fields such as actions and indexes will not be affected.

The dumps.get action has been removed.

Done by @ManyTheFish in #2438, #2451, #2475, and #2474)

Changes in /indexes/{uid}/settings endpoints

The following changes only concern the HTTP verb associated with a given method.

The endpoint to update the global settings is now PATCH /indexes/{indexUid}/settings instead of POST

The endpoint to update the typo tolerance settings is now PATCH /indexes/{indexUid}/settings/typo-tolerance instead of POST

The endpoint to update the displayed attributes settings is now PUT /indexes/{indexUid}/settings/displayed-attributes instead of POST

The endpoint to update the distinct attribute settings is now PUT /indexes/{indexUid}/settings/distinct-attribute instead of POST

The endpoint to update the filterable attribute settings is now PUT /indexes/{indexUid}/settings/filterable-attributes instead of POST

The endpoint to update the ranking rules settings is now PUT /indexes/{indexUid}/settings/ranking-rules instead of POST

The endpoint to update the searchable attributes settings is now PUT /indexes/{indexUid}/settings/searchable-attributes instead of POST

The endpoint to update the sortable attributes settings is now PUT /indexes/{indexUid}/settings/sortable-attributes instead of POST

The endpoint to update the stop words settings is now PUT /indexes/{indexUid}/settings/stop-words instead of POST

The endpoint to update the synonyms settings is now PUT /indexes/{indexUid}/settings/synonyms instead of POST

Done by @Kerollmops in #2452

Other improvements 🔨

Geosearch: Support string type for _geo, lat, and lng fields (#2408 and meilisearch/milli#523) @irevoire

Improve ranking rules error message (meilisearch/milli#536, #2468) @matthias-wright

Remove the connection keep-alive timeout (#2471) @Thearas

Improve index uid validation upon API key creation (#2414) @pierre-l

Fixes 🐞

Fix filtering on documents with double and/or single quote characters (#2500 and meilisearch/milli#552) @irevoire

Misc

Telemetry: handle the X-Meilisearch-Client header (#2374) @irevoire

Telemetry: send the analytics to telemetry.meilisearch.com instead of segment (#2429) @irevoire

Create custom error types for TaskType, TaskStatus, and IndexUid (#2460) @walterbm

Improve the Docker CIs (#2477 and #2491) @choznerol and @curquiza

Some internal changes (#2447, #2464, #2453, #2455) @MarinPostma @Kerollmops

Update dependencies (#2450) @Kerollmops

Add dump tests in CI (#2357) @irevoire

Deny warnings in CI (#2339) @MarinPostma

Change Nelson path (#2347) @curquiza

Re-integrate Clippy in the CI (#2424) @irevoire

Fix typo in README (#2448) @ryanrussell

Fix typo in code base comments (#2459, and #2508) @ryanrussell

Update url of movies.json in README (#2422) @0x0x1

Fix command in README (#2434) @0x0x1

❤️ Thanks again to our external contributors:

Meilisearch: @0x0x1, @choznerol, @pierre-l, @ryanrussell, @Thearas, and @walterbm

Milli: @matthias-wright

Charabia: @benny-n

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(125.04 MB)
meilisearch-linux-amd64(120.64 MB)
meilisearch-macos-amd64(106.86 MB)
meilisearch-windows-amd64.exe(103.28 MB)
v0.27.2(Jun 7, 2022)
⚠️ Use v0.27.2 instead of v0.27.1 if you want to use the experimental auto batching feature

Fixes

Fix wrong internal ids assignments that lead to bugs when using autobatching #2470 @irevoire @Kerollmops

Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(123.29 MB)
meilisearch-linux-amd64(123.50 MB)
meilisearch-macos-amd64(109.72 MB)
meilisearch-windows-amd64.exe(106.18 MB)
meilisearch.deb(18.34 MB)
v0.27.1(May 17, 2022)
⚠️ Use v0.27.1 instead of v0.27.0 if possible

Fix dumps bug during import (#2396) @MarinPostma

Fix geosearch bug when using mixed datasets (#2397 and https://github.com/meilisearch/milli/pull/529) @irevoire

Fix nested bug when using searchable attributes (#2397 and https://github.com/meilisearch/milli/pull/528) @irevoire

Thank you @leopku, @martinmaillard and @Thearas for the bug reports ❤️
Source code(tar.gz)
Source code(zip)
meilisearch-linux-aarch64(120.64 MB)
meilisearch-linux-amd64(120.84 MB)
meilisearch-macos-amd64(108.17 MB)
meilisearch-windows-amd64.exe(105.99 MB)
meilisearch.deb(18.24 MB)

Lightning Fast, Ultra Relevant, and Typo-Tolerant Search Engine

Related tags

Overview

MeiliSearch

Website | Roadmap | Blog | LinkedIn | Twitter | Documentation | FAQ

✨ Features

Getting started

Deploy the Server

Homebrew (Mac OS)

Docker

Try MeiliSearch in our Sandbox

Run on Digital Ocean

Deploy on Platform.sh

APT (Debian & Ubuntu)

Download the binary (Linux & Mac OS)

Compile and run it from sources

Create an Index and Upload Some Documents

Search for Documents

In command line

Use the Web Interface

Documentation

Contributing

Telemetry

💌 Contact

Comments

How Algolia handles that

How?

Impacted teams

Requirements

Format

Priority order

Config file content

Config file command line

Display when launching Meilisearch

Errors (bonus)

Implementations

TODO

For Meili people

Tokio v1.23.1

Fixed

Tokio v1.23.0

Fixed

Changed

Tokio v1.22.0

Added

Added (unstable)

Fixed

Releases(v0.30.5)

v0.30.5(Dec 21, 2022)

Bug fixes 🐞

v0.30.4(Dec 15, 2022)

Bug fixes 🐞

v0.30.3(Dec 14, 2022)

Bug fixes 🐞

v0.30.2(Dec 8, 2022)

Bug fixes 🐞

v0.30.1(Dec 6, 2022)

Bug fixes 🐞

v0.29.3(Dec 5, 2022)

Bug fixes 🐞

v0.30.0(Nov 28, 2022)

New features and improvements 🔥

Improve search result navigation: exhaustive number of search results

Cancel processing and enqueued tasks

Delete the task history

Deploy indexes with zero downtime

Configuration file support

Other improvements

Behavior changes ⚠️

Fixes 🐞

Misc

v0.30.0rc5(Nov 24, 2022)

Bug fixes

v0.30.0rc4(Nov 23, 2022)

Bug fixes

Misc

v0.30.0rc3(Nov 21, 2022)

v0.30.0rc2(Nov 17, 2022)

Changes

Bug fixes 🐞

Filtering: introduce the `EXISTS` and `IN` operators

`!=` (`NOT`) filter operator

Filtering: introduce the `EXISTS` and `IN` operators

`!=` (`NOT`) filter operator