jayvdb/drf-spectacular

Sane and flexible OpenAPI 3 schema generation for Django REST framework.

drf-spectacular

Sane and flexible OpenAPI 3 schema generation for Django REST framework.

This project has 3 goals:

1. Extract as much schema information from DRF as possible.
2. Provide flexibility to make the schema usable in the real world (not only toy examples).
3. Generate a schema that works well with the most popular client generators.

The code is a heavily modified fork of the DRF OpenAPI generator, which is/was lacking all of the below listed features.

Features

• Serializers modelled as components. (arbitrary nesting and recursion supported)

• @extend_schema decorator for customization of APIView, Viewsets, function-based views, and @action

  • additional parameters
  • request/response serializer override (with status codes)
  • polymorphic responses either manually with PolymorphicProxySerializer helper or via rest_polymorphic's PolymorphicSerializer)
  • ... and more customization options

• Authentication support (DRF natives included, easily extendable)
• Custom serializer class support (easily extendable)
• MethodSerializerField() type via type hinting or @extend_schema_field
• Tags extraction
• Description extraction from docstrings
• Sane fallbacks where no Serializer is available (free-form objects)
• Sane operation_id naming (based on path)
• Easy to use hooks for extending the spectacular AutoSchema
• Optional schema serving with SpectacularAPIView

• Included support for:

  • django-polymorphic / django-rest-polymorphic
  • SimpleJWT
  • DjangoOAuthToolkit

Incomplete features (in progress):

• optional separate component versions for PATCH serializers (no required fields)
• optional input/output serializer component split

License

Provided by T. Franzel, Cashlink Technologies GmbH. Licensed under 3-Clause BSD.

Requirements

• Python >= 3.6
• Django (2.2, 3.0)
• Django REST Framework (3.10, 3.11)

Installation

Install using pip…

$ pip install drf-spectacular

then add drf-spectacular to installed apps in settings.py

INSTALLED_APPS = [
    # ALL YOUR APPS
    'drf_spectacular',
]

and finally register our spectacular AutoSchema with DRF

REST_FRAMEWORK = {
    # YOUR SETTINGS
    'DEFAULT_SCHEMA_CLASS': 'drf_spectacular.openapi.AutoSchema',
}

Take it for a spin

drf-spectacular is KISS. It only generates a valid OpenAPI 3 specification for you and nothing else. You can easily view your schema with the excellent Swagger UI or any other compliant UI or tool:

$ ./manage.py spectacular --file schema.yml
$ docker run -p 80:8080 -e SWAGGER_JSON=/schema.yml -v ${PWD}/schema.yml:/schema.yml swaggerapi/swagger-ui

or serve the schema directly from your API with

from drf_spectacular.views import SpectacularAPIView
urlpatterns = [
    # YOUR PATTERNS
    url(r'^api/schema$', SpectacularAPIView.as_view(), name='schema')
]

Usage

drf-spectacular works pretty well out of the box. You might also want to set some metadata for your API. Just create a SPECTACULAR_SETTINGS dictionary in your settings.py and override the defaults. Have a look at the available settings.

The toy examples do not cover your cases? No problem, you can heavily customize how your schema will be rendered.

Customization by using @extend_schema

Most customization cases should be covered by the extend_schema decorator. We usually get pretty far with specifying OpenApiParameter and splitting request/response serializers, but the sky is the limit.

from drf_spectacular.utils import extend_schema, OpenApiParameter
from drf_spectacular.types import OpenApiTypes

class AlbumViewset(viewset.ModelViewset)
    serializer_class = AlbumSerializer

    @extend_schema(
        request=AlbumCreationSerializer
        responses={201: AlbumSerializer},
    )
    def create(self, request):
        # your non-standard behaviour
        return super().create(request)

    @extend_schema(
        # extra parameters added to the schema
        parameters=[
            OpenApiParameter(name='artist', description='Filter by artist', required=False, type=str),
            OpenApiParameter(
                name='release',
                type=OpenApiTypes.DATE,
                location=OpenApiParameter.QUERY,
                description='Filter by release date',
            ),
        ],
        # override default docstring extraction
        description='More descriptive text',
        # provide Authentication class that deviates from the views default
        auth=None,
        # change the auto-generated operation name
        operation_id=None,
        # or even completely override what AutoSchema would generate. Provide raw Open API spec as Dict.
        operation=None,
    )
    def list(self, request):
        # your non-standard behaviour
        return super().list(request)

    @extend_schema(
        request=AlbumLikeSerializer
        responses={204: None},
    )
    @action(detail=True, methods=['post'])
    def set_password(self, request, pk=None):
        # your action behaviour

Customization by overriding AutoSchema

Still not satisifed? You want more! We still got you covered. We prepared some convenient hooks for things that are probably up to taste. If you are careful, you can change pretty much anything.

Don't forget to register your custom AutoSchema in the DEFAULT_SCHEMA_CLASS.

from drf_spectacular.openapi import AutoSchema

class CustomAutoSchema(AutoSchema):
    def get_operation_id(self, path, method):
        return 'YOUR-ID'.replace('-', '_')

Extras

got endpoints that yield list of differing objects? Enter PolymorphicProxySerializer

@extend_schema(
    responses=PolymorphicProxySerializer(
        component_name='MetaPerson',
        serializers=[SerializerA, SerializerB],
        resource_type_field_name='type',
    )
)
@api_view()
def poly_list(request):
    return Response(list_of_multiple_object_types)

Testing

Install testing requirements.

$ pip install -r requirements.txt

Run with runtests.

$ ./runtests.py

You can also use the excellent tox testing tool to run the tests against all supported versions of Python and Django. Install tox globally, and then simply run:

$ tox