[Python] How can I measure code coverage of API integration tests?

In my company we have a Django project that has API endpoints. We are using Pytest as testing framework and some of our tests executing requests to such endpoints using requests.Session(). For example:

The test perform a GET request to /api/account/data

content of root/dir/tests/test_api.py


def test_some_api(self):
client = requests.Session()
response = client.get('/api/account/data')
assert response.status_code == 200


When performing the request, the backend execute this function: `

root/models/account/api.py


def user_data(request, format=None):
"""
@api {get} /api/account/data
"""
if request.method == "GET":
return APIResponse(code=200)


We would like to measure the code executed in `user_settings` by the test, but I failed to make it work. We are using:

coverage==4.5.4 # forced to use an old version as we have dependencies conflict

pytest-cov==2.10.1

To measure coverage I run this command from root

pytest -v --cov-report html:html-report --cov=dir --cov-config=dir/.coveragerc

.coveragerc - has only files to exclude


I verified in backend logs that when the test runs, it execute the if block inside user_settings

I have tried adding this code to conftest.py as written in coverage docs, but it still didn't measure user_settings code execution.

@pytest.fixture(scope='session', autouse=True)
def run_cov():
process = coverage.process_startup()
source_path = os.path.join(_home_path, "/models/account")
cov = coverage.Coverage(config_file='root/.coveragerc', source=[source_path])
cov.start()
yield cov # Tests runs here
cov.stop()
cov.save()
cov.html_report(directory='html-report')


I've seen this solution How to generate coverage report for http based integration tests?, but failed to understand how to implement it on my end.

I expected user_settings code to be measured by coverage report.

Continue reading...