Snorlax

A light-weight RESTful client built on top of Guzzle that gives you full control of your API's resources. Its based on method definitions and parameters for your URLs. See the usage below.

Basic Usage

<?php

use Snorlax\RestResource;
use Snorlax\RestClient;

class PokemonResource extends RestResource
{
    public function getBaseUri()
    {
        // You don't want a raw value like this, use an environment variable :)
        return 'http://localhost/api/pokemons';
    }

    public function getActions()
    {
        return [
            'all' => [
                'method' => 'GET',
                'path' => '/',
            ],
            'get' => [
                'method' => 'GET',
                'path' => '/{0}.json',
            ],
            'create' => [
                'method' => 'POST',
                'path' => '/',
            ],
        ];
    }
}

$client = new RestClient([
    'resources' => [
        'pokemons' => PokemonResource::class,
    ],
]);

// GET http://localhost/api/pokemons?sort=id:asc
$response = $client->pokemons->all([
    'query' => [
        'sort' => 'id:asc',
    ],
]);

// GET http://localhost/api/pokemons/143.json?fields=id,name
$response = $client->pokemons->get(143, [
    'query' => [
        'fields' => 'id,name',
    ],
]);

// POST http://localhost/api/pokemons
$response = $client->pokemons->create([
    'body' => [
        'name' => 'Bulbasaur',
    ],
]);

As you can see, each action on your resource is defined an array with two keys, method, defining the HTTP method for the request and path, which defines the path from the base URI returned by the getBaseUri method. You can mock URLs using environment variables as you wish.

Snorlax assume your API returns JSON, so it already returns an StdClass object with the response, decoded by json_decode. If you want to get the raw object returned by Guzzle, use $client->resource->getLastResponse().

Amending the response

As noted above, Snorlax returns an StdClass object, however Resources may overwrite the ->parse() method to manipulate the returned response. This is useful when an API returns a nested set of data such as {'pokemon': {'name':'Mew'}} and you only want the actual data (in this case pokemon). In this example we could use

public function parse($method, $response)
{
    return $response->pokemon;
}

This would return the actual pokemon object. Another scario is that you may want to return a Laravel Collection (Illuminate\Support\Collection) of objects, you could simply do

public function parse($method, $response)
{
    return collect($response->pokemon);
}

The $method argument is the name of the method which was called to perform the request, such as 'all', or 'get'. This is useful to manipulate different response, such as

public function parse($method, $response)
{
    switch ($method) {
        case 'all':
            return collect($response->pokemon);
            break;
        case 'get':
            return $response->pokemon;
            break;
    }
}

Another usage could be to cast certain fields are data types. In this example, we'll cast any fields called created_at or updated_at to Carbon isntances

public function parse($action, $response)
{
    $date_fields = [
        'created_at',
        'updated_at',
    ];

    $response = $response->pokemon;

    foreach ($date_fields as $date_field) {
        if (property_exists($response, $date_field)) {
            $response->{$date_field} = Carbon::parse($response->{$date_field});
        }
    }

    return $response;
}

Sending parameters and headers

As said before, Snorlax is built on top of Guzzle, so it works basically the same way on passing headers, query strings and request bodies.

<?php

$pokemons = $client->pokemons->all([
    'query' => [
        'sort' => 'name',
        'offset' => 0,
        'limit' => 150,
    ],
    'headers' => [
        'X-Foo' => 'Bar',
    ],
]);

$pokemons = $client->pokemons->create([
    'body' => [
        'name' => 'Ivysaur',
        'attacks' => [
            'Tackle',
            'Leer',
        ],
    ],
]);

Changing client options

If you want to set default headers for every request you send to the default guzzle client, just use the headers options on your params config key, just like the Guzzle docs.

<?php

$client = new Snorlax\RestClient([
    'client' => [
        'params' => [
            'headers' => [
                'X-Foo' => 'Bar',
            ],
            'defaults' => [
                'debug' => true,
            ],
            'cache' => true,
        ],
    ],
]);

Setting a base URI

If all your resources are under the same base URI, you can pass it on the constructor instead of declaring on the resource class.

<?php

$client = new Snorlax\RestClient([
    'client' => [
        'params' => [
            'base_uri' => 'http://localhost/api',
        ],
    ],
]);

Using a custom client

If you don't want to use Guzzle's default client (or want to mock one), Snorlax accepts any class that implements GuzzleHttp\ClientInterface, so just pass your custom client in the constructor. Can be an instance or a callable.

<?php

class MyOwnClient implements GuzzleHttp\ClientInterface
{
    private $config;

    public function __construct(array $params)
    {
        $this->config = $params;
    }
}

// Using a callable to instantiate a new client everytime
$client = new Snorlax\RestClient([
    'client' => [
        'custom' => function(array $params) {
            return new MyOwnClient($params);
        },
        'params' => [
            'param1' => 'value',
        ],
    ],
]);

$client = new Snorlax\RestClient([
    'client' => [
        'custom' => new MyOwnClient([
            'param1' => 1,
        ]),
    ],
]);

Using a custom cache strategy & driver

If you don't want to use the VolatileRuntime cache driver or even want to change cache strategy, Snorlax allow you to inject any of the stragies and storages provided by kevinrob/guzzle-cache-middleware.

<?php

$cacheDriver = \Illuminate\Support\Facades\Cache::store('redis');
$cacheStorage = new \Kevinrob\GuzzleCache\Storage\LaravelCacheStorage($cacheDriver);
$cacheStrategy = new \Kevinrob\GuzzleCache\Strategy\PrivateCacheStrategy($cacheStorage);

$client = new Snorlax\RestClient([
    'client' => [
        'cacheStrategy' => $cacheStrategy,
        'params' => [
            'cache' => true,
        ],
    ],
]);

Using parallel requests

The parallel pool is made on the top of Guzzle, because of that, you should be using Guzzle.

<?php
use Snorlax\RestClient;
use Snorlax\RestPool;
use Snorlax\RestResource;

class PostResource extends RestResource
{
    public function getBaseUri()
    {
        return 'https://jsonplaceholder.typicode.com/posts';
    }

    public function getActions()
    {
        return [
            'all' => [
                'method' => 'GET',
                'path' => '/',
            ],
            'show' => [
                'method' => 'GET',
                'path' => '/{0}',
            ],
        ];
    }
}

$client = new RestClient([
    'resources' => [
        'posts' => PostResource::class,
    ],
]);

$pool = new RestPool();

$pool->addResource('postsByUserOne', $client, 'posts.all', [
    'query' => [
        'userId' => '1',
        'cache' => rand(11111, 99999), // bypass cache
    ],
]);

$pool->addResource('postsByUserTwo', $client, 'posts.show', [
    'parameters' => [
        2
    ],
    'query' => [
        'userId' => '2',
        'cache' => rand(11111, 99999), // bypass cache
    ],
]);

$requests = $pool->send();

The result of $requests is a StdClass with postsByUserOne and postsByUserTwo as properties.

Reconnections

Is possible to reconnect when Guzzle throws a GuzzleHttp\Exception\ConnectException. By default, Snorlax tries to reconnect 3 times.

If you want, you can change the number of tries with the retries parameter.

<?php

$pokemons = $client->pokemons->all([
    'retries' => 3,
    'query' => [
        'limit' => 150,
    ],
]);

Contributing

Please see the CONTRIBUTING.md file for guidelines.

License

Please see the LICENSE file for License.