Skip to main content
Version: Laravel: 4.x (current)

Supported annotations

PHP 8 attributes vs docblock tags

Scribe v4 introduces PHP 8 attributes that provide the same functionality as the old docblock tags. Here's a quick comparison of what they look like:


/**
* Healthcheck
*
* Check that the service is up. If everything is okay, you'll get a 200 OK response.
*
* Otherwise, the request will fail with a 400 error, and a response listing the failed services.
*
* @response 400 scenario="Service is unhealthy" {"status": "down", "services": {"database": "up", "redis": "down"}}
* @responseField status The status of this API (`up` or `down`).
* @responseField services Map of each downstream service and their status (`up` or `down`).
*/
public function healthcheck() {
return [
'status' => 'up',
'services' => [
'database' => 'up',
'redis' => 'up',
],
];
});

Attributes have the following advantages over tags:

  1. They're less error-prone (for us). With docblocks, everything is a string, so we have to parse it, and try to guess your intentions. There's a lot that can go wrong there. With attributes, you give us the exact values you want.
  2. They're less error-prone (for you). Attributes are a language feature, so you have IDE help built in. For instance, typing in #[Response( will bring up the list of parameters, so you don't need to memorize the specific order or field names.
  3. They're programmable. Since attributes are actual PHP code (with some limits), you can do more. For instance, you can have #[ResponseFromApiResource(paginate: self::PAGINATION_CONFIG)]. You can create your own attributes to avoid repeating the same things.

On the flip side, attributes:

  1. can be bulky. They especially don't look good for long text, such as descriptions.

  2. don't look good on inline (closure) routes.

    // This isn't valid PHP
    #[Authenticated]
    Route::get('/endpoint', function () { ... });

    // This works, but it's not very nice visually.
    Route::get(
    '/endpoint',

    #[Authenticated]
    function () { ...
    });

The good news is that you can mix them!

That means you can write an endpoint like this:

/**
* Healthcheck
*
* Check that the service is up. If everything is okay, you'll get a 200 OK response.
*
* Otherwise, the request will fail with a 400 error, and a response listing the failed services.
*/
#[Response(["status" => "down", "services" => ["database" => "up", "redis" => "down"]], status: 400, description: "Service is unhealthy")]
#[ResponseField("status", "The status of this API (`up` or `down`).")]
#[ResponseField("services", "Map of each downstream service and their status (`up` or `down`).")]
public function healthcheck() {
return [
'status' => 'up',
'services' => [
'database' => 'up',
'redis' => 'up',
],
];
});

This way, the text part stays textual, while the structured part uses the defined attributes.

If you'd like to try attributes, we made a Rector rule to automatically convert your tags to attributes. It will convert parameter tags to attributes, but leave text like endpoint titles and descriptions untouched.

Format

Annotations in docblocks typically consist of a tag (@-something) followed by text in a certain format. Some important details: Attributes are written like a regular PHP function call, and you can use named parameters to make the code clearer.

Some things to note about tags:

  • The @hideFromAPIDocumentation, @authenticated and @unauthenticated tags are the only boolean tags; they don't take any text after them.

  • In the "Format" sections below, ? indicates an optional value.

  • Tags typically default required to false

  • Most annotations are written in a "natural" format, @tag value1 value2, where Scribe figures out what value1 and value2 represent, based on the order. However, some tags also support fields (@tag key1=value1 value2 or @tag value2 key1=value1).

    Tag fields don't have to follow a specific order; they can be at the start or end of the tag (but they generally cannot be in the middle). Tag attribute values which consist of multiple words should use quotes (eg @tag key1="this is value1" value2).

Some things to note about attributes:

  • They all live under the Knuckles\Scribe\Attributes namespace. So you can either write #[Knuckles\Scribe\Attributes\Header], or write #[Header] and have an import statement (use Knuckles\Scribe\Attributes\Header).
  • Since they're regular PHP, you can easily find out the arguments with your IDE, like you would for a normal function call. We won't list all the arguments here.
  • Attributes typically default required to true

Here's a list of all the docblock annotations Scribe supports.

Metadata annotations

tip

All metadata annotations can be used on the method or class docblock. Using them on the class will make them apply to all endpoints in that class.

TagDescriptionFormat
@hideFromAPIDocumentationExcludes an endpoint from the docs@hideFromAPIDocumentation
@groupAdds an endpoint to a group@group <groupName>
Example: @group User management
@authenticatedIndicates that an endpoint needs auth@authenticated
@unauthenticatedOpposite of @authenticated@unauthenticated

Request parameter annotations

@header / #[Header]

Describes a request header.

Format: @header <name> <example?>

Examples:

@header Api-Version
@header Content-Type application/xml

@urlParam/#[UrlParam]

Describes a URL parameter.

Format: @urlParam <name> <type?> required? <description?> Example: <example?>

Notes:

  • If you don't supply a type, string is assumed.
  • To prevent Scribe from including this parameter in example requests:
    • end the description with No-example when using tags
    • pass"No-example"as the example parameter when using attributes
  • You can also use this on Form Request classes.

Examples:

@urlParam id
@urlParam id int
@urlParam id int required
@urlParam id int required The user's ID.
@urlParam id int required The user's ID. Example: 88683
@urlParam id int The user's ID. Example: 88683
@urlParam id int The user's ID. No-example

@queryParam/#[QueryParam]

Describes a query parameter.

Format: @queryParam <name> <type?> required? <description?> Example: <example?>

Notes:

  • If you don't supply a type, string is assumed.
  • To prevent Scribe from including this parameter in example requests:
    • end the description with No-example when using tags
    • pass"No-example"as the example parameter when using attributes
  • You can also use this on Form Request classes.

Examples:

@queryParam date The date. Example: 2022-01-01
@queryParam page int
@queryParam page int The page number.
@queryParam page int required The page number. Example: 4
@queryParam page int The page number. No-example

@bodyParam/#[BodyParam]

Describes a request body parameter.

Format: @bodyParam <name> <type> required? <description?> Example: <example?>

Notes:

  • To prevent Scribe from including this parameter in example requests:
    • end the description with No-example when using tags
    • pass"No-example"as the example parameter when using attributes
  • You can also use this on Form Request classes.

Examples:

@bodyParam room_id string
@bodyParam room_id string required The room ID.
@bodyParam room_id string The room ID. Example: r98639bgh3
@bodyParam room_id string Example: r98639bgh3

// Objects and arrays
@bodyParam user object required The user data
@bodyParam user.name string required The user's name.
@bodyParam user.age int Example: 1000
@bodyParam people object[] required List of people
@bodyParam people[].name string Example: Deadpool

// If your entire request body is an array
@bodyParam [] object[] required List of things to do
@bodyParam [].name string Name of the thing. Example: Cook

Response annotations

@response/#[Response]

Describes an example response.

Format: @response <status?> <response>

Notes:

  • If you don't specify a status, Scribe will assume 200.
  • Supported fields: scenario, status

Examples:

@response {"a": "b"}
@response 201 {"a": "b"}
@response 201 {"a": "b"} scenario="Operation successful"
@response status=201 scenario="Operation successful" {"a": "b"}
@response scenario=Success {"a": "b"}
@response 201 scenario="Operation successful" {"a": "b"}

@responseFile/#[ResponseFile]

Describes the path to a file containing an example response. The path can be absolute, relative to your project directory, or relative to your Laravel storage directory.

Format: @responseFile <status?> <filePath>

Notes:

  • If you don't specify a status, Scribe will assume 200.
  • Supported fields: scenario, status

Examples:

@responseFile /an/absolute/path
@responseFile 400 relative/path/from/your/project/root
@responseFile status=400 scenario="Failed" path/from/your/laravel/storage/directory
@responseFile 400 scenario="Failed" path/from/your/laravel/storage/directory

@responseField/#[ResponseField]

Describes a field in the response.

Format: @responseField <name> <type?> <description?>

Notes:

  • You can omit the type; Scribe will try to figure it out from your example responses.
  • You can also use this on Eloquent API resource toArray() methods.

Examples:

@responseField total The total number of results.
@responseField total int The total number of results.

@apiResource

Tells Scribe how to generate a response using an Eloquent API resource. Must be used together with @apiResourceModel.

Format: @apiResource <status?> <resourceClass>

Notes:

  • If you don't specify a status, Scribe will assume 200.

Examples:

@apiResource App\Http\Resources\UserApiResource
@apiResourceModel App\Models\User

@apiResource 201 App\Http\Resources\UserApiResource
@apiResourceModel App\Models\User

@apiResourceCollection

Tells Scribe how to generate a response using an Eloquent API resource collection. Must be used together with @apiResourceModel.

Format: @apiResourceCollection <status?> <resourceClass>

Notes:

  • If you don't specify a status, Scribe will assume 200.

Examples:

@apiResourceCollection App\Http\Resources\UserApiResource
@apiResourceModel App\Models\User

@apiResourceCollection App\Http\Resources\UserApiResourceCollection
@apiResourceModel App\Models\User

@apiResourceCollection 201 App\Http\Resources\UserApiResourceCollection
@apiResourceModel App\Models\User

@apiResourceModel

Tells Scribe the model to use when generating the Eloquent API resource response. Must be used together with @apiResource or @apiResourceCollection.

Format: @apiResourceModel <modelClass>

Notes:

  • Supported fields:
    • states: Comma-separated list of states to be applied when creating an example model via factory.
    • with: Comma-separated list of relations to be loaded with the model. Works for factory (Laravel 8+) or database fetching.
    • paginate: The number of items per page (when generating a collection). To use simple pagination instead, add ,simple after the number.
@apiResource App\Http\Resources\UserApiResource
@apiResourceModel App\Models\User

@apiResourceCollection App\Http\Resources\UserApiResource
@apiResourceModel App\Models\User states=editor,verified

@apiResource App\Http\Resources\UserApiResource
@apiResourceModel App\Models\User with=accounts,pets

@apiResourceCollection App\Http\Resources\UserApiResource
@apiResourceModel App\Models\User paginate=5

@apiResourceCollection App\Http\Resources\UserApiResourceCollection
@apiResourceModel App\Models\User paginate=5,simple

@apiResource App\Http\Resources\UserApiResource
@apiResourceModel App\Models\User with=accounts states=editor,verified

@apiResourceAdditional

Specifies additional metadata for an API resource. Can only be used with @apiResource and @apiResourceCollection. The additional metadata is specified as fields (key-value pairs).

Format: @apiResourceAdditional <key1>=<value1> ... <keyN>=<valueN>

Notes:

  • Supported formats for key-value pairs:
    • key=value
    • key="long text with spaces"
    • "key with spaces"="long text with spaces"

Examples:

@apiResource App\Http\Resources\UserApiResource
@apiResourceModel App\Models\User
@apiResourceAdditional result=success message="User created successfully"

#[ResponseFromApiResource]

All-in-one attribute alternative to @apiResource, @apiResourceCollection, @apiResourceModel and @apiResourceAdditional.

Examples:

use App\Models\User;
use App\Http\Resources\UserApiResource;
use App\Http\Resources\UserApiResourceCollection;

#[ResponseFromApiResource(UserApiResource::class, User::class)]
#[ResponseFromApiResource(UserApiResource::class, User::class, status: 201)]
#[ResponseFromApiResource(UserApiResource::class, User::class, 201, description: "User details")]
public function endpoint() {...}

// Collections
#[ResponseFromApiResource(UserApiResource::class, User::class, 201, collection: true)]
#[ResponseFromApiResource(UserApiResourceCollection::class, User::class, 201)]
public function endpoint() {...}

// Specifying factory states and relations
#[ResponseFromApiResource(UserApiResource::class, User::class,
factoryStates: ['editor', 'verified'], with: ['accounts', 'pets'])]
public function endpoint() {...}

// Pagination
#[ResponseFromApiResource(UserApiResourceCollection::class, User::class, paginate: 10)]
#[ResponseFromApiResource(UserApiResource::class, User::class, collection: true, paginate: 10)]
#[ResponseFromApiResource(UserApiResource::class, User::class, collection: true, simplePaginate: 10)]
public function endpoint() {...}

// Additional data
#[ResponseFromApiResource(UserApiResource::class, User::class, 201,
merge: ["result" => "success", "message" => "User created successfully")]
public function endpoint() {...}

@transformer

Tells Scribe how to generate a response using a Fractal transformer. Can be used together with @transformerModel.

Format: @transformer <status?> <transformerClass>

Notes:

  • If you don't specify a status, Scribe will assume 200.
  • If you don't specify transformerModel, Scribe will use the first argument to your method.

Examples:

@transformer App\Http\Transformers\UserTransformer
@transformerModel App\Models\User

@transformer 201 App\Http\Transformers\UserTransformer
@transformerModel App\Models\User

@transformerCollection

Tells Scribe how to generate a response using a Fractal transformer collection. Can be used together with @transformerModel and @transformerPaginator.

Format: @transformerCollection <status?> <transformerClass>

Examples:

@transformerCollection App\Http\Transformers\UserCollectionTransformer
@transformerModel App\Models\User

@transformerCollection 201 App\Http\Transformers\UserCollectionTransformer
@transformerModel App\Models\User

@transformerModel

Tells Scribe the model to use when generating the Fractal transformer response. Can only be used together with @transformer or @transformerCollection (along with @transformerPaginator).

Format: @transformerModel <modelClass>

Notes:

  • Supported fields:
    • states: Comma-separated list of states to be applied when creating an example model via factory.
    • with: Comma-separated list of relations to be loaded with the model. Works for factory (Laravel 8+) or database fetching.
    • resourceKey: The resource key to be used during serialization.
@transformer App\Http\Transformers\UserTransformer
@transformerModel App\Models\User

@transformerCollection App\Http\Transformers\UserTransformer
@transformerModel App\Models\User states=editor,verified

@transformer App\Http\Transformers\UserTransformer
@transformerModel App\Models\User with=accounts,pets

@transformerPaginator

Tells Scribe the paginator to use when generating the Fractal transformer response. Can only be used together with @transformerCollection.

Format: @transformerPaginator <adapterClass> <perPage?>

Examples:

@transformerCollection App\Http\Transformers\UserCollectionTransformer
@transformerModel App\Models\User
@transformerPaginator League\Fractal\Pagination\IlluminatePaginatorAdapter

@transformerCollection App\Http\Transformers\UserCollectionTransformer
@transformerModel App\Models\User
@transformerPaginator League\Fractal\Pagination\IlluminatePaginatorAdapter 15

#[ResponseFromTransformer]

All-in-one attribute alternative to @transformer, @transformerCollection, @transformerModel and @transformerPaginator.

Examples:

use App\Models\User;
use App\Http\Transformers\UserTransformer;
use App\Http\Transformers\UserCollectionTransformer;
use League\Fractal\Pagination\IlluminatePaginatorAdapter;

#[ResponseFromTransformer(UserTransformer::class, User::class)]
#[ResponseFromTransformer(UserTransformer::class, User::class, status: 201)]
#[ResponseFromTransformer(UserTransformer::class, User::class, 201, description: "User details")]
#[ResponseFromTransformer(UserTransformer::class, User::class, 201, resourceKey: "uuid")]
public function endpoint() {...}

// Collections
#[ResponseFromTransformer(UserTransformer::class, User::class, 201, collection: true)]
#[ResponseFromTransformer(UserCollectionTransformer::class, User::class, 201)]
public function endpoint() {...}

// Specifying factory states and relations
#[ResponseFromTransformer(UserTransformer::class, User::class,
factoryStates: ['editor', 'verified'], with: ['accounts', 'pets'])]
public function endpoint() {...}

// Pagination
#[ResponseFromTransformer(UserCollectionTransformer::class, User::class,
paginate: [IlluminatePaginatorAdapter::class])]
#[ResponseFromTransformer(UserTransformer::class, User::class, collection: true,
paginate: [IlluminatePaginatorAdapter::class, 15])]
public function endpoint() {...}