Serving the schema

get_schema_view and the SchemaView class

The get_schema_view() function and the SchemaView class it returns (click links for documentation) are intended to cover the majority of use cases one might want to configure. The class returned by get_schema_view() can be used to obtain view instances via SchemaView.with_ui(), SchemaView.without_ui() and SchemaView.as_cached_view() - see 1. Quickstart in the README for a usage example.

You can also subclass SchemaView by extending the return value of get_schema_view(), e.g.:

SchemaView = get_schema_view(info, ...)

class CustomSchemaView(SchemaView):
    generator_class = CustomSchemaGenerator
    renderer_classes = (CustomRenderer1, CustomRenderer2,)

Renderers and codecs

If you need to modify how your Swagger spec is presented in views, you might want to override one of the renderers in renderers or one of the codecs in codecs. The codec is the last stage where the Swagger object arrives before being transformed into bytes, while the renderer is the stage responsible for tying together the codec and the view.

You can use your custom renderer classes as kwargs to SchemaView.as_cached_view() or by subclassing SchemaView.

Management command

If you only need a swagger spec file in YAML or JSON format, you can use the generate_swagger management command to get it without having to start the web server:

$ python generate_swagger swagger.json

See the command help for more advanced options:

$ python generate_swagger --help
usage: generate_swagger [-h] [--version] [-v {0,1,2,3}]
   ... more options ...


The DEFAULT_INFO setting must be defined when using the generate_swagger command. For example, the README quickstart code could be modified as such:


   'DEFAULT_INFO': '',


api_info = openapi.Info(
   title="Snippets API",
   ... other arguments ...

schema_view = get_schema_view(
   # the info argument is no longer needed here as it will be picked up from DEFAULT_INFO
   ... other arguments ...