Stars: 816
Forks: 98
Pull Requests: 64
Issues: 133
Watchers: 14
Last Updated: 2023-08-16 12:44:05
A Laravel Fractal package for building API responses, giving you the power of Fractal with Laravel's elegancy.
License: MIT License
Languages: PHP
Laravel Responder is a package for building API responses, integrating Fractal into Laravel and Lumen. It can transform your data using transformers, create and serialize success- and error responses, handle exceptions and assist you with testing your responses.
Laravel lets you return models directly from a controller method to convert it to JSON. This is a quick way to build APIs but leaves your database columns exposed. Fractal, a popular PHP package from The PHP League, solves this by introducing transformers. However, it can be a bit cumbersome to integrate into the framework as seen below:
public function index()
{
$resource = new Collection(User::all(), new UserTransformer());
return response()->json((new Manager)->createData($resource)->toArray());
}
Not that bad, but we all get a little spoiled by Laravel's magic. Wouldn't it be better if we could refactor it to:
public function index()
{
return responder()->success(User::all())->respond();
}
The package will allow you to do this and much more. The goal has been to create a high-quality package that feels like native Laravel. A package that lets you embrace the power of Fractal, while hiding it behind beautiful abstractions. There has also been put a lot of focus and thought to the documentation. Happy exploration!
This package requires:
To get started, install the package through Composer:
composer require flugger/laravel-responder
Append the following line to the providers
key in config/app.php
to register the package:
Flugg\Responder\ResponderServiceProvider::class,
The package supports auto-discovery, so if you use Laravel 5.5 or later you may skip registering the service provider and facades as they will be registered automatically.
If you like facades, you may also append the Responder
and Transformation
facades to the aliases
key:
'Responder' => Flugg\Responder\Facades\Responder::class,
'Transformation' => Flugg\Responder\Facades\Transformation::class,
You may additionally publish the package configuration and language file using the vendor:publish
Artisan command:
php artisan vendor:publish --provider="Flugg\Responder\ResponderServiceProvider"
This will publish a responder.php
configuration file in your config
folder. It will also publish an errors.php
file inside your lang/en
folder which can be used for storing error messages.
Add the following line to app/bootstrap.php
to register the package:
$app->register(Flugg\Responder\ResponderServiceProvider::class);
You may also add the following lines to app/bootstrap.php
to register the facades:
class_alias(Flugg\Responder\Facades\Responder::class, 'Responder');
class_alias(Flugg\Responder\Facades\Transformation::class, 'Transformation');
Seeing there is no vendor:publish
command in Lumen, you will have to create your own config/responder.php
file if you want to configure the package.
This documentation assumes some knowledge of how Fractal works.
The package has a Responder
service class, which has a success
and error
method to build success- and error responses respectively. To use the service and begin creating responses, pick one of the options below:
Responder
ServiceYou may inject the Flugg\Responder\Responder
service class directly into your controller methods:
public function index(Responder $responder)
{
return $responder->success();
}
You can also use the error
method to create error responses:
return $responder->error();
responder
HelperIf you're a fan of Laravel's response
helper function, you may like the responder
helper function:
return responder()->success();
return responder()->error();
Responder
FacadeOptionally, you may use the Responder
facade to create responses:
return Responder::success();
return Responder::error();
MakesResponses
TraitLastly, the package provides a Flugg\Responder\Http\MakesResponses
trait you can use in your controllers:
return $this->success();
return $this->error();
Which option you pick is up to you, they are all equivalent, the important thing is to stay consistent. The helper function (option 2) will be used for the remaining of the documentation.
The success
and error
methods return a SuccessResponseBuilder
and ErrorResponseBuilder
respectively, which both extend an abstract ResponseBuilder
, giving them common behaviors. They will be converted to JSON when returned from a controller, but you can explicitly create an instance of Illuminate\Http\JsonResponse
with the respond
method:
return responder()->success()->respond();
return responder()->error()->respond();
The status code is set to 200
by default, but can be changed by setting the first parameter. You can also pass a list of headers as the second argument:
return responder()->success()->respond(201, ['x-foo' => true]);
return responder()->error()->respond(404, ['x-foo' => false]);
Consider always using the respond
method for consistency's sake.
Instead of converting the response to a JsonResponse
using the respond
method, you can cast the response data to a few other types, like an array:
return responder()->success()->toArray();
return responder()->error()->toArray();
You also have a toCollection
and toJson
method at your disposal.
A response decorator allows for last minute changes to the response before it's returned. The package comes with two response decorators out of the box adding a status
and success
field to the response output. The decorators
key in the configuration file defines a list of all enabled response decorators:
'decorators' => [
\Flugg\Responder\Http\Responses\Decorators\StatusCodeDecorator::class,
\Flugg\Responder\Http\Responses\Decorators\SuccessFlagDecorator::class,
],
You may disable a decorator by removing it from the list, or add your own decorator extending the abstract class Flugg\Responder\Http\Responses\Decorators\ResponseDecorator
. You can also add additional decorators per response:
return responder()->success()->decorator(ExampleDecorator::class)->respond();
return responder()->error()->decorator(ExampleDecorator::class)->respond();
The package also ships with some situational decorators disabled by default, but which can be added to the decorator list:
PrettyPrintDecorator
decorator will beautify the JSON output;\Flugg\Responder\Http\Responses\Decorators\PrettyPrintDecorator::class,
EscapeHtmlDecorator
decorator, based on the "sanitize input, escape output" concept, will escape HTML entities in all strings returned by your API. You can securely store input data "as is" (even malicious HTML tags) being sure that it will be outputted as un-harmful strings. Note that, using this decorator, printing data as text will result in the wrong representation and you must print it as HTML to retrieve the original value.\Flugg\Responder\Http\Responses\Decorators\EscapeHtmlDecorator::class,
As briefly demonstrated above, success responses are created using the success
method:
return responder()->success()->respond();
Assuming no changes have been made to the configuration, the above code would output the following JSON:
{
"status": 200,
"success": true,
"data": null
}
The success
method takes the response data as the first argument:
return responder()->success(Product::all())->respond();
It accepts the same data types as you would normally return from your controllers, however, it also supports query builder and relationship instances:
return responder()->success(Product::where('id', 1))->respond();
return responder()->success(Product::first()->shipments())->respond();
The package will run the queries and convert them to collections behind the scenes.
The response data will be transformed with Fractal if you've attached a transformer to the response. There are two ways to attach a transformer; either explicitly by setting it on the response, or implicitly by binding it to a model. Let's look at both ways in greater detail.
You can attach a transformer to the response by sending a second argument to the success
method. For instance, below we're attaching a simple closure transformer, transforming a list of products to only output their names:
return responder()->success(Product::all(), function ($product) {
return ['name' => $product->name];
})->respond();
You may also transform using a dedicated transformer class:
return responder()->success(Product::all(), ProductTransformer::class)->respond();
return responder()->success(Product::all(), new ProductTransformer)->respond();
You can read more about creating dedicated transformer classes in the Creating Transformers chapter.
If no transformer is set, the package will search the response data for an element implementing the Flugg\Responder\Contracts\Transformable
interface to resolve a transformer from. You can take use of this by implementing the Transformable
interface in your models:
class Product extends Model implements Transformable {}
You can satisfy the contract by adding a transformer
method that returns the corresponding transformer:
/**
* Get a transformer for the class.
*
* @return \Flugg\Responder\Transformers\Transformer|string|callable
*/
public function transformer()
{
return ProductTransformer::class;
}
You're not limited to returning a class name string, you can return a transformer instance or closure transformer, just like the second parameter of the success
method.
Instead of implementing the Transformable
contract for all models, an alternative approach is to bind the transformers using the bind
method on the TransformerResolver
class. You can place the code below within AppServiceProvider
or an entirely new TransformerServiceProvider
:
use Flugg\Responder\Contracts\Transformers\TransformerResolver;
public function boot()
{
$this->app->make(TransformerResolver::class)->bind([
\App\Product::class => \App\Transformers\ProductTransformer::class,
\App\Shipment::class => \App\Transformers\ShipmentTransformer::class,
]);
}
After you've bound a transformer to a model you can skip the second parameter and still transform the data:
return responder()->success(Product::all())->respond();
As you might have noticed, unlike Fractal, you don't need to worry about creating resource objects like Item
and Collection
. The package will make one for you based on the data type, however, you may wrap your data in a resource object to override this.
If the data you send into the response is a model or contains a list of models, a resource key will implicitly be resolved from the model's table name. You can overwrite this by adding a getResourceKey
method to your model:
public function getResourceKey(): string {
return 'products';
}
You can also explicitly set a resource key on a response by sending a third argument to the ´success` method:
return responder()->success(Product::all(), ProductTransformer::class, 'products')->respond();
Sending a paginator to the success
method will set pagination meta data and transform the data automatically, as well as append any query string parameters to the paginator links.
return responder()->success(Product::paginate())->respond();
Assuming there are no products and the default configuration is used, the JSON output would look like:
{
"success": true,
"status": 200,
"data": [],
"pagination": {
"total": 0,
"count": 0,
"perPage": 15,
"currentPage": 1,
"totalPages": 1,
"links": []
}
}
Instead of sending a paginator as data, you may set the data and paginator seperately, like you traditionally would with Fractal. You can manually set a paginator using the paginator
method, which expects an instance of League\Fractal\Pagination\IlluminatePaginatorAdapter
:
$paginator = Product::paginate();
$adapter = new IlluminatePaginatorAdapter($paginator);
return responder()->success($paginator->getCollection())->paginator($adapter)->respond();
You can also set cursors using the cursor
method, expecting an instance of League\Fractal\Pagination\Cursor
:
if ($request->has('cursor')) {
$products = Product::where('id', '>', request()->cursor)->take(request()->limit)->get();
} else {
$products = Product::take(request()->limit)->get();
}
$cursor = new Cursor(request()->cursor, request()->previous, $products->last()->id ?? null, Product::count());
return responder()->success($products)->cursor($cursor)->respond();
If a transformer class is attached to the response, you can include relationships using the with
method:
return responder()->success(Product::all())->with('shipments')->respond();
You can send multiple arguments and specify nested relations using dot notation:
return responder()->success(Product::all())->with('shipments', 'orders.customer')->respond();
All relationships will be automatically eager loaded, and just like you would when using with
or load
to eager load with Eloquent, you may use a callback to specify additional query constraints. Like in the example below, where we're only including related shipments that hasn't yet been shipped:
return responder()->success(Product::all())->with(['shipments' => function ($query) {
$query->whereNull('shipped_at');
}])->respond();
Relationships are loaded from a query string parameter if the load_relations_parameter
configuration key is set to a string. By default, it's set to with
, allowing you to automatically include relations from the query string:
GET /products?with=shipments,orders.customer
In your transformer classes, you may specify relations to automatically load. You may disable any of these relations using the without
method:
return responder()->success(Product::all())->without('comments')->respond();
The technique of filtering the transformed data to only return what we need is called sparse fieldsets and can be specified using the only
method:
return responder()->success(Product::all())->only('id', 'name')->respond();
When including relationships, you may also want to filter fields on related resources as well. This can be done by instead specifying an array where each key represents the resource keys for the resources being filtered
return responder()->success(Product::all())->with('shipments')->only([
'products' => ['id', 'name'],
'shipments' => ['id']
])->respond();
Fields will automatically be filtered if the filter_fields_parameter
configuration key is set to a string. It defaults to only
, allowing you to filter fields from the query string:
GET /products?only=id,name
You may automatically filter related resources by setting the parameter to a key-based array:
GET /products?with=shipments&only[products]=id,name&only[shipments]=id
You may want to attach additional meta data to your response. You can do this using the meta
method:
return responder()->success(Product::all())->meta(['count' => Product::count()])->respond();
When using the default serializer, the meta data will simply be appended to the response array:
{
"success": true,
"status": 200,
"data": [],
"count": 0
}
After the data has been transformed, it will be serialized using the specified success serializer in the configuration file, which defaults to the package's own Flugg\Responder\Serializers\SuccessSerializer
. You can overwrite this on your responses using the serializer
method:
return responder()->success()->serializer(JsonApiSerializer::class)->respond();
return responder()->success()->serializer(new JsonApiSerializer())->respond();
Above we're using Fractal's JsonApiSerializer
class. Fractal also ships with an ArraySerializer
and DataArraySerializer
class. If none of these suit your taste, feel free to create your own serializer by extending League\Fractal\Serializer\SerializerAbstract
. You can read more about it in Fractal's documentation.
A dedicated transformer class gives you a convenient location to transform data and allows you to reuse the transformer at multiple places. It also allows you to include and transform relationships. You can create a transformer using the make:transformer
Artisan command:
php artisan make:transformer ProductTransformer
The command will generate a new ProductTransformer.php
file in the app/Transformers
folder:
<?php
namespace App\Transformers;
use App\User;
use Flugg\Responder\Transformers\Transformer;
class ProductTransformer extends Transformer
{
/**
* List of available relations.
*
* @var string[]
*/
protected $relations = [];
/**
* A list of autoloaded default relations.
*
* @var array
*/
protected $load = [];
/**
* Transform the model.
*
* @param \App\Product $product
* @return array
*/
public function transform(Product $product): array
{
return [
'id' => (int) $product->id,
];
}
}
It will automatically resolve a model name from the name provided. For instance, the package will extract Product
from ProductTransformer
and assume the models live directly in the app
folder (as per Laravel's convention). If you store them somewhere else, you can use the --model
(or -m
) option to override it:
php artisan make:transformer ProductTransformer --model="App\Models\Product"
The transformer file generated above is a model transformer expecting an App\Product
model for the transform
method. However, we can create a plain transformer by applying the --plain
(or -p
) modifier:
php artisan make:transformer ProductTransformer --plain
This will remove the typehint from the transform
method and add less boilerplate code.
The $relations
and $load
properties in the transformer are the equivalent to Fractal's own $availableIncludes
and $defaultIncludes
. In addition to the slight name change, the package uses the $relations
and $load
properties to map out all available relationships for eager loading, so in contrast to Fractal, you should map the relationship to their corresponding transformer:
protected $relations = [
'shipments' => ShipmentTransformer::class,
];
You can choose to skip the mapping and just pass the strings like with Fractal, but that means the package wont be able to eager load relationships automatically.
The $relations
property specifies a list of relations available to be included. You can set a list of relations mapped to their corresponding transformers:
protected $relations = [
'shipments' => ShipmentTransformer::class,
'orders' => OrderTransformer::class,
];
The $load
property specifies a list of relations to be autoloaded every time you transform data with the transformer:
protected $load = [
'shipments' => ShipmentTransformer::class,
'orders' => OrderTransformer::class,
];
You don't have to add relations to both $relations
and $load
, all relations in $load
will be available by nature.
While Fractal requires you to to create a method in your transformer for every included relation, this package lets you skip this when transforming models, as it will automatically fetch relationships from the model. You may of course override this functionality by creating an "include" method:
/**
* Include related shipments.
*
* @param \App\Product $product
* @return mixed
*/
public function includeShipments(Product $product)
{
return $product->shipments;
}
Unlike Fractal you can just return the data directly without wrapping it in an item
or collection
method.
You should be careful with executing database calls inside the include methods as you might end up with an unexpected amount of hits to the database.
Fractal can parse query string parameters which can be used when including relations. For more information about how to format the parameters see Fractal's documentation on parameters. You may access the parameters by adding a second parameter to the "include" method:
public function includeShipments(Product $product, Collection $parameters)
{
return $product->shipments->take($parameters->get('limit'));
}
To be as decoupled from Fractal as possible the parameters (which are normally accessed using League\Fractal\ParamBag
) are accessed as Laravel collections instead.
Just as you can specify a query constraint when including a relationship with the with
method, you can also add query constraints as a "load" method on the transformer. This will automatically be applied when extracting relationships for eager loading.
/**
* Load shipments with constraints.
*
* @param \Illuminate\Database\Eloquent\Builder $query
* @return \Illuminate\Database\Eloquent\Builder
*/
public function loadShipments($query)
{
return $query->whereNull('shipped_at');
}
Note: You cannot mix "include" and "load" methods because the package doesn't eager load relationships included with an "include" method.
After a relation has been included, you can make any last second changes to it using a filter method. For instance, below we're filtering the list of related shipments to only include shipments that has not been shipped:
/**
* Filter included shipments.
*
* @param \Illuminate\Database\Eloquent\Collection $shipments
* @return \Illuminate\Support\Collection
*/
public function filterShipments($shipments)
{
return $shipments->filter(function ($shipment) {
return is_null($shipment->shipped_at);
});
}
We've looked at how to transform response data of success responses, however, there may be other places than your controllers where you want to transform data. An example is broadcasted events where you're exposing data using websockets instead of HTTP. You just want to return the transformed data, not an entire response.
It's possible to simply transform data by newing up the transformer and calling transform
:
return (new ProductTransformer)->transform(Product::first());
However, this approach might become a bit messy when building transformations with relationships:
return array_merge((new ProductTransformer)->transform($product = Product::first()), [
'shipments' => $product->shipments->map(function ($shipment) {
return (new ShipmentTransformer)->transform($shipment);
})
]);
Yuck! Imagine that with multiple nested relationships. Let's explore a simpler way to handle this.
The SuccessResponseBuilder
actually delegates all of the transformation work to a dedicated Flugg\Responder\TransformBuilder
class. We can use this class ourself to transform data. For instance, if the product and shipment transformers were bound to the models, we could replicate the code above in the following way:
public function index(TransformBuilder $transformation)
{
return $transformation->resource(Product::all())->with('shipments')->transform();
}
Instead of using the success
method on the Responder
service, we use the resource
method on the TransformBuilder
with the same method signature. We also use transform
to execute the transformation instead of respond
as we did when creating responses. In addition to the with
method, you also have access to the other transformation methods like without
, only
, meta
and serializer
.
Using toArray
on the Responder
service is almost the same as the code above, however, it will also include response decorators which might not be desired.
When using the TransformBuilder
to transform data it will still serialize the data using the configured serializer. Fractal requires the use of a serializer to transform data, but sometimes we're just interested in the raw transformed data. The package ships with a Flugg\Responder\Serializers\NoopSerializer
to solve this, a no-op serializer which leaves the transformed data untouched:
return $transformation->resource(Product::all())->serializer(NoopSerializer::class)->transform();
If you think this is looking messy, don't worry, there's a quicker way. In fact, you will probably never even need to utilize the NoopSerializer
or TransformBuilder
manually, but it helps to know how it works. The Flugg\Responder\Transformation
is a class which can be used for quickly transforming data without serializing.
Transformation
ServiceThe Transformation
class utilizes the TransformBuilder
class to build a transformation using the NoopSerializer
. You can inject the Transformation
class and call make
to obtain an instance of TransformBuilder
which gives you access to all of the chainable methods including with
, like below:
public function __construct(Transformation $transformation)
{
$transformation->make(Product::all())->with('shipments')->transform();
}
transformation
HelperYou can use the transformation
helper function to transform data without serializing:
transformation(Product::all())->with('shipments')->transform();
Transformation
FacadeYou can also use the Transformation
facade to achieve the same thing:
Transformation::make(Product::all())->with('shipments')->transform();
Model attributes are traditionally specified in snake case, however, you might prefer to use camel case for the response fields. A transformer makes for a perfect location to convert the fields, as seen from the soldOut
field in the example below:
return responder()->success(Product::all(), function ($product) {
return ['soldOut' => (bool) $product->sold_out];
})->respond();
After responding with camel case, you probably want to let people send in request data using camel cased parameters as well. The package provides a Flugg\Responder\Http\Middleware\ConvertToSnakeCase
middleware you can append to the $middleware
array in app/Http/Kernel.php
to convert all parameters to snake case automatically:
protected $middleware = [
// ...
\Flugg\Responder\Http\Middleware\ConvertToSnakeCase::class,
];
The middleware will run before request validation, so you should specify your validation rules in snake case as well.
Whenever a consumer of your API does something unexpected, you can return an error response describing the problem. As briefly shown in a previous chapter, an error response can be created using the error
method:
return responder()->error()->respond();
The error response has knowledge about an error code, a corresponding error message and optionally some error data. With the default configuration, the above code would output the following JSON:
{
"success": false,
"status": 500,
"error": {
"code": null,
"message": null
}
}
You can fill the first parameter of the error
method to set an error code:
return responder()->error('sold_out_error')->respond();
You may optionally use integers for error codes.
In addition, you may set the second parameter to an error message describing the error:
return responder()->error('sold_out_error', 'The requested product is sold out.')->respond();
You can set the error messages in a language file, which allows for returning messages in different languages. The configuration file has an error_message_files
key defining a list of language files with error messages. By default, it is set to ['errors']
, meaning it will look for an errors.php
file inside resources/lang/en
. You can use these files to map error codes to corresponding error messages:
return [
'sold_out_error' => 'The requested product is sold out.',
];
ErrorMessageResolver
Instead of using language files, you may alternatively set error messages directly on the ErrorMessageResolver
class. You can place the code below within AppServiceProvider
or an entirely new TransformerServiceProvider
:
use Flugg\Responder\ErrorMessageResolver;
public function boot()
{
$this->app->make(ErrorMessageResolver::class)->register([
'sold_out_error' => 'The requested product is sold out.',
]);
}
You may want to set additional data on the error response. Like in the example below, we're returning a list of shipments with the sold_out
error response, giving the consumer information about when a new shipment for the product might arrive.
return responder()->error('sold_out')->data(['shipments' => Shipment::all()])->respond();
The error data will be appended to the response data. Assuming we're using the default serializer and there are no shipments in the database, the code above would look like:
{
"success": false,
"status": 500,
"error": {
"code": "sold_out",
"message": "The requested product is sold out.",
"shipments": []
}
}
Similarly to success responses, error responses will be serialized using the specified error serializer in the configuration file. This defaults to the package's own Flugg\Responder\Serializers\ErrorSerializer
, but can of course be changed by using the serializer
method:
return responder()->error()->serializer(ExampleErrorSerializer::class)->respond();
return responder()->success()->serializer(new ExampleErrorSerializer())->respond();
You can create your own error serializer by implementing the Flugg\Responder\Contracts\ErrorSerializer
contract.
No matter how much we try to avoid them, exceptions do happen. Responding to the exceptions in an elegant manner will improve the user experience of your API. The package can enhance your exception handler to automatically turn exceptions in to error responses. If you want to take use of this, you can either use the package's exception handler or include a trait as described in further details below.
Handler
ClassTo use the package's exception handler you need to replace the default import in app/Exceptions/Handler.php
:
use Illuminate\Foundation\Exceptions\Handler as ExceptionHandler;
With the package's handler class:
use Flugg\Responder\Exceptions\Handler as ExceptionHandler;
This will not work with Lumen as its exception handler is incompatible with Laravel's. Look instead at the second option below.
ConvertsExceptions
TraitThe package's exception handler uses the Flugg\Responder\Exceptions\ConvertsExceptions
trait to load of most of its work. Instead of replacing the exception handler, you can use the trait in your own handler class. To replicate the behavior of the exception handler, you would also have to add the following code to the render
method:
public function render($request, Exception $exception)
{
$this->convertDefaultException($exception);
if ($exception instanceof HttpException) {
return $this->renderResponse($exception);
}
return parent::render($request, $exception);
}
If you only want to return JSON error responses on requests actually asking for JSON, you may wrap the code above in a wantsJson
check as seen below:
if ($request->wantsJson()) {
$this->convertDefaultException($exception);
if ($exception instanceof HttpException) {
return $this->renderResponse($exception);
}
}
Once you've implemented one of the above options, the package will convert some of Laravel's exceptions to an exception extending Flugg\Responder\Exceptions\Http\HttpException
. It will then convert these to an error response. The table below shows which Laravel exceptions are converted and what they are converted to. All the exceptions on the right is under the Flugg\Responder\Exceptions\Http
namespace and extends Flugg\Responder\Exceptions\Http\HttpException
. All exceptions extending the HttpException
class will be automatically converted to an error response.
Caught Exceptions | Converted To |
---|---|
Illuminate\Auth\AuthenticationException |
UnauthenticatedException |
Illuminate\Auth\Access\AuthorizationException |
UnauthorizedException |
Symfony\Component\HttpKernel\Exception\NotFoundHttpException |
PageNotFoundException |
Illuminate\Database\Eloquent\ModelNotFoundException |
PageNotFoundException |
Illuminate\Database\Eloquent\RelationNotFoundException |
RelationNotFoundException |
Illuminate\Validation\ValidationException |
ValidationFailedException |
You can disable the conversions of some of the exceptions above using the $dontConvert
property:
/**
* A list of default exception types that should not be converted.
*
* @var array
*/
protected $dontConvert = [
ModelNotFoundException::class,
];
If you're using the trait option, you can disable all the default conversions by removing the call to convertDefaultException
in the render
method.
In addition to letting the package convert Laravel exceptions, you can also convert your own exceptions using the convert
method in the render
method:
$this->convert($exception, [
InvalidValueException => PageNotFoundException,
]);
You can optionally give it a closure that throws the new exception, if you want to give it constructor parameters:
$this->convert($exception, [
MaintenanceModeException => function ($exception) {
throw new ServerDownException($exception->retryAfter);
},
]);
An exception class is a convenient place to store information about an error. The package provides an abstract exception class Flugg\Responder\Exceptions\Http\HttpException
, which has knowledge about status code, an error code and an error message. Continuing on our product example from above, we could create our own HttpException
class:
<?php
namespace App\Exceptions;
use Flugg\Responder\Exceptions\Http\HttpException;
class SoldOutException extends HttpException
{
/**
* The HTTP status code.
*
* @var int
*/
protected $status = 400;
/**
* The error code.
*
* @var string|null
*/
protected $errorCode = 'sold_out_error';
/**
* The error message.
*
* @var string|null
*/
protected $message = 'The requested product is sold out.';
}
You can also add a data
method returning additional error data:
/**
* Retrieve additional error data.
*
* @return array|null
*/
public function data()
{
return [
'shipments' => Shipment::all()
];
}
If you're letting the package handle exceptions, you can now throw the exception anywhere in your application and it will automatically be rendered to an error response.
throw new SoldOutException();
Contributions are more than welcome and you're free to create a pull request on Github. You can run tests with the following command:
vendor/bin/phpunit
If you find bugs or have suggestions for improvements, feel free to submit an issue on Github. However, if it's a security related issue, please send an email to [email protected] instead.
The package is completely free to use, however, a lot of time has been put into making it. If you want to show your appreciation by leaving a small donation, you can do so by clicking here. Thanks!
Laravel Responder is free software distributed under the terms of the MIT license. See license.md for more details.