Skip to content

Latest commit

 

History

History
36 lines (27 loc) · 1.71 KB

TESTING.md

File metadata and controls

36 lines (27 loc) · 1.71 KB
Raw

Running tests

The Datadog Python API client is mostly using BDD tests to verify its behavior. They are using recodings to store HTTP interactions, allowing to run the tests without talking to the API. We also store the test time to be able to freeze.

You can control the behavior with the RECORD environment variable:

  • RECORD=false, the default value, means replaying HTTP requests from recordings.
  • RECORD=true creates or updates recordings. This will need valid credentials in DD_TEST_CLIENT_API_KEY and DD_TEST_CLIENT_APP_KEY.
  • RECORD=none ignores recordings. This will also runs tests that we call integration-only, i.e. tests that we don't record for security reasons. It also needs valid credentials.

Recording and freeze files are stored in the directory tests/$VERSION/cassettes/test_scenarios/

To run the tests, it's generally better to use a virtual environment. tox makes it easier to manage them. For example to run the tests for Python 3.9:

tox -epy39

To pass extra arguments to run the tests, you need to close tox arguments with --. You can get more verbose information with the -v flag to pytest, filter the tests using the -k or specifying the test full path. For example:

tox -epy39  -- -k test_get_a_list_of_all_incident_teams_returns_ok_response -vvv

-k takes a regular expression, so you don't have to specify the whole exact string. See https://docs.pytest.org/ for more information.

The first time you run a test that needs recordings, it will fail with: Time file '$PATH/tests/$VERSION/cassettes/test_scenarios/$TEST_NAME.frozen' not found: create one setting 'RECORD=true' or ignore it using 'RECORD=none'.