The OpsRamp platform provides a rich set of APIs. By using these APIs, customers can build solutions to automate various IT Operations Management (ITOM) workflows. These could include discovery and monitoring, event and incident management, or remediation and automation. Testing these APIs and the resulting automated solution is critical to ensure the reliability and stability of the workflow.

PyTest is a powerful Python testing framework and it is widely used to test APIs. Fixtures are one of the most important capabilities of the PyTest framework. In this article, I will discuss some advanced techniques for testing OpsRamp APIs using PyTest fixtures.

What fixtures are

PyTest fixtures are a special type of Python function that provisions a fixed baseline for testing. With the help of this baseline, you can ensure tests are run reliably and produce consistent results. In addition, you can ensure that the same tests are repeatable.

Install PyTest

Run the below command.

pip install -U pytest

Verify PyTest installation

Run the below command.

pytest --version
pytest 7.4.0

Define fixture

You can define fixtures by decorating a simple Python function with @pytest.fixture.

Python Decorators are a very powerful and convenient way to alter the behaviour of functions. For more details, refer to Python Decorators.

import pytest


@pytest.fixture
def hello_world():
    return 'Hello, Happy testing'

Invoke fixtures

Test functions can invoke fixtures just by declaring the required fixtures as arguments.

def test_hello_world(hello_world):
    assert hello_world == 'Hello, Happy testing'

Invoke OpsRamp API

As shown below, execute OpsRamp Get Access Token API.

import http.client
import json
import traceback
import urllib
from asyncio.log import logger
import pytest


def get_auth_header(api_endpoint, client_id, client_secret, grant_type="client_credentials"):
    """
    Returns bearer token string in the blow format
    bearer adfa-afdaf-1599-402c-a3ee-1ed24f597cc8
    """
    api_connection = http.client.HTTPSConnection(api_endpoint)
    params = {"client_id": client_id, "client_secret": client_secret, "grant_type": "client_credentials"}
    payload = urllib.parse.urlencode(params)


    headers = {
        'accept': "application/json",
        'content-type': "application/x-www-form-urlencoded",
    }

Define fixture to get access token

You can transform the above function as a PyTest fixture by simply decorating the code as shown below.

@pytest.fixture
def get_auth_header(api_endpoint, client_id, client_secret, grant_type="client_credentials"):

Autouse fixtures

Sometimes some of your fixtures are required for all other tests. The above Get access token is the perfect use case for this. Every API call depends on this API. In this case, you can designate this fixture as an “autouse fixture” by passing in autouse=True to the fixture’s decorator. The autouse fixture capability will reduce a lot of redundant requests.

@pytest.fixture(autouse=True)
def get_auth_header(api_endpoint, client_id, client_secret, grant_type="client_credentials"):

Provisioning of third-party plugin fixtures

You do not necessarily have to define all fixtures on your code. Fixtures can be provided by third-party plugins as well and you are free to use them. You just need to install the required plugins. You can see an example with the pytest-datadir plugin.

Install this plugin as shown below.

pip install pytest-datadir

@pytest.fixture(autouse=True)
def get_bearer_token(shared_datadir,request):
    json_path = (shared_datadir / "auth_header_input.json")

The get_bearer_token fixture is using shared_datadir fixture from a third-party plugin. The Shared_datadir fixture provisions the data folder as pathlib.Path object.

Fixture finalization

Usually, you do teardown or cleanup activities for each test. There are many good reasons for this, such as:

• The executed tests will not impact other tests
• Do not want to have the testing environment piled up with tons of test data
• Every test should execute in a clean state

Fixtures offer a very powerful teardown/cleanup capability known as finalization.

@pytest.fixture()
def create_client(shared_datadir, get_bearer_token, request):
    client = Client()
    client.logger.debug("IN***")


    payload_file = (shared_datadir / "create_client1.json")
    input_payload = client.http_helper.to_json(payload_file)


    # client_response=client.http_helper.do_post(url, input_payload, get_bearer_token)
    client_response = client.create_client(input_payload, get_bearer_token)


    def terminate_client():
        client.logger.debug("IN***")
        util = Utils()


        clientId = Client.clientId


        if len(str(clientId)) > 0:
            oauth_token = get_bearer_token
            client.terminate_client(clientId, oauth_token)
        client.logger.debug("OUT***")


    request.addfinalizer(terminate_client)


    client.print_log(client_response)
    client.logger.debug("OUT***")
    return client_response

You should add the finalizer function to the request’s context object of the test as shown below:

request.addfinalizer(terminate_client)

In the above example, you do all the test set-up on the client and execute the test. Once the test is complete, the finalizer function runs and cleans up the whole thing.

Conclusion

In this article, I have explained what PyTest fixtures are and how to define, invoke and implement fixtures. I also showed you how to leverage the testing of OpsRamp APIs using these fixtures and how to do teardown activities using the finalizer functions. Check back often on the HPE Developer Community blog to find more blog posts on OpsRamp.