doctor

This module uses python types to validate request and response data in Flask Python APIs. It uses python 3 type hints to validate request paramters and generate API documentation. It also supports generic schema validation for plain dictionaries. An example of the generated API documentation can be found in the docs.

Install

doctor can easily be installed using pip:

$ pip install doctor

Quick Start

Define some types that will be used to validate your request parameters.

# mytypes.pyfromdoctorimporttypes# doctor provides helper functions to easily define simple types.FooId=types.integer('The foo ID.')
FetchBars=types.boolean('A flag that indicates if we should fetch bars')
# You can also inherit from type classes to create more complex types.classFoo(types.Object):
description='A Foo object'example= {'foo_id': 1}
properties= {'foo_id': FooId}
required= ['foo_id']
additional_properties=False

Define the logic function that our endpoint will route to:

# foo.pyfrommytypesimportFoo, FooId, FetchBars# Note the type annotations on this function definition. This tells Doctor how# to parse and validate parameters for routes attached to this logic function.# The return type annotation will validate the response conforms to an# expected definition in development environments. In non-development# environments a warning will be logged.defget_foo(foo_id: FooId, fetch_bars: FetchBars=False) ->Foo:
"""Fetches the Foo object and optionally related bars."""returnFoo.get_by_id(foo_id, fetch_bars=fetch_bars)

Now tie the endpoint to the logic function with a route:

fromflaskimportFlaskfromflask_restfulimportApifromdoctor.routingimportcreate_routes, get, Routefromfooimportget_fooroutes= (
Route('/foo/<int:foo_id>/', methods=(
get(get_foo),)
),
)
app=Flask(__name__)
api=Api(app)
forroute, resourceincreate_routes(routes):
api.add_resource(resource, route)

That's it, you now have a functioning API endpoint you can curl and the request is automatically validated for you based on your schema. Positional arguments in your logic function are considered required request parameters and keyword arguments are considered optional. As a bonus, using the autoflask sphinx directive, you will also get automatically generated API documentation.

Documentation

Documentation and a full example is available at readthedocs.

Running Tests

Tests can be run with tox. It will handle installing dependencies into a virtualenv, running tests, and rebuilding documentation.

Then run Tox:

cd doctor
tox

You can pass arguments to pytest directly:

tox -- test/test_flask.py