drf-openapi-tester 2.1.0

Test utility for validating OpenAPI response documentation

DRF OpenAPI Tester

This is a test utility to validate DRF Test Responses against OpenAPI 2 and 3 schema. It has built-in support for:

• OpenAPI 2/3 yaml or json schema files.
• OpenAPI 2 schemas created with drf-yasg.
• OpenAPI 3 schemas created with drf-spectacular.

Installation

pip install drf-openapi-tester

Usage

First instantiate one or more instances of SchemaTester:

from openapi_tester import SchemaTester

schema_tester = SchemaTester()

If you are using either drf-yasg or drf-spectacular this will be auto-detected, and the schema will be loaded by the SchemaTester automatically. If you are using schema files though, you will need to pass the file path to the tester:

from openapi_tester import SchemaTester

# path should be a string
schema_tester = SchemaTester(schema_file_path="./schemas/publishedSpecs.yaml")

Once you instantiate a tester, you can use it to test responses:

from openapi_tester.schema_tester import SchemaTester

# you need to create at least one instance of SchemaTester.
# you can pass kwargs to it
schema_tester = SchemaTester()


def test_response_documentation(client):
    response = client.get('api/v1/test/1')
    assert response.status_code == 200
    schema_tester.validate_response(response=response)

If you are using the Django testing framework, you can create a base APITestCase that incorporates schema validation:

from rest_framework.response import Response
from rest_framework.test import APITestCase

from openapi_tester.schema_tester import SchemaTester

schema_tester = SchemaTester()


class BaseAPITestCase(APITestCase):
    """ Base test class for api views including schema validation """

    @staticmethod
    def assertResponse(response: Response, **kwargs) -> None:
        """ helper to run validate_response and pass kwargs to it """
        schema_tester.validate_response(response=response, **kwargs)

Then use it in a test file:

from shared.testing import BaseAPITestCase


class MyAPITests(BaseAPITestCase):
    def test_some_view(self):
        response = self.client.get("...")
        self.assertResponse(response)

Options

You can pass options either globally, when instantiating a SchemaTester, or locally, when invoking validate_response:

from openapi_tester import SchemaTester, is_camel_case
from tests.utils import my_uuid_4_validator

schema_test_with_case_validation = SchemaTester(
    case_tester=is_camel_case,
    ignore_case=["IP"],
    validators=[my_uuid_4_validator]
)

Or

from openapi_tester import SchemaTester, is_camel_case
from tests.utils import my_uuid_4_validator

schema_tester = SchemaTester()


def my_test(client):
    response = client.get('api/v1/test/1')
    assert response.status_code == 200
    schema_tester.validate_response(
        response=response,
        case_tester=is_camel_case,
        ignore_case=["IP"],
        validators=[my_uuid_4_validator]
    )

case_tester

The case tester argument takes a callable that is used to validate the key casings of both schemas and responses. If nothing is passed, case validation is skipped.

The library currently has 4 built-in case testers:

• is_pascal_case
• is_snake_case
• is_camel_case
• is_kebab_case

You can of course pass your own custom case tester.

ignore_case

List of keys to ignore when testing key casing. This setting only applies when case_tester is not None.

validators

List of custom validators. A validator is a function that receives two parameters: schema_section and data, and returns either an error message or None, e.g.:

from typing import Any, Optional
from uuid import UUID


def my_uuid_4_validator(schema_section: dict, data: Any) -> Optional[str]:
    schema_format = schema_section.get("format")
    if schema_format == "uuid4":
        try:
            result = UUID(data, version=4)
            if not str(result) == str(data):
                return f"Expected uuid4, but received {data}"
        except ValueError:
            return f"Expected uuid4, but received {data}"
    return None

field_key_map

You can pass an optional dictionary that maps custom url parameter names into values, for cases where this cannot be inferred by the DRF EndpointEnumerator. A concrete use case for this option is when the django i18n locale prefixes.

from openapi_tester import SchemaTester

schema_tester = SchemaTester(field_key_map={
  "language": "en",
})

Schema Validation

When the SchemaTester loads a schema, it runs it through OpenAPI Spec validator which validates that the schema passes without specification compliance issues. In case of issues with the schema itself, the validator will raise the appropriate error.

Django testing client

The library includes an OpenAPIClient, which extends Django REST framework's APIClient class. If you wish to validate each response against OpenAPI schema when writing unit tests - OpenAPIClient is what you need!

To use OpenAPIClient simply pass SchemaTester instance that should be used to validate responses and then use it like regular Django testing client:

schema_tester = SchemaTester()
client = OpenAPIClient(schema_tester=schema_tester)
response = client.get('/api/v1/tests/123/')

To force all developers working on the project to use OpenAPIClient simply override the client fixture (when using pytest with pytest-django):

from pytest_django.lazy_django import skip_if_no_django

from openapi_tester.schema_tester import SchemaTester


@pytest.fixture
def schema_tester():
    return SchemaTester()


@pytest.fixture
def client(schema_tester):
    skip_if_no_django()

    from openapi_tester.clients import OpenAPIClient

    return OpenAPIClient(schema_tester=schema_tester)

If you are using plain Django test framework, we suggest to create custom test case implementation and use it instead of original Django one:

import functools

from django.test.testcases import SimpleTestCase
from openapi_tester.clients import OpenAPIClient
from openapi_tester.schema_tester import SchemaTester

schema_tester = SchemaTester()


class MySimpleTestCase(SimpleTestCase):
    client_class = OpenAPIClient
    # or use `functools.partial` when you want to provide custom
    # ``SchemaTester`` instance:
    # client_class = functools.partial(OpenAPIClient, schema_tester=schema_tester)

This will ensure you all newly implemented views will be validated against the OpenAPI schema.

Known Issues

• We are using prance as a schema resolver, and it has some issues with the resolution of (very) complex OpenAPI 2.0 schemas. If you encounter issues, please document them here.

Contributing

Contributions are welcome. Please see the contributing guide