Pytest

Unit Test, is something I ignore for a long time. I know it existence but barely initiate one. If it’s already in the system, I don’t mind add one, like for Canopy.
But recent requests I get, it always ask for unit tests, so here are some finding regarding pytest.

1. Test with Pytest

• Run code with pytest is straightforward, just pytest code.py
• Use assert for test check, or define an assert_function
• Use with pytest.raises(ValueError): for konwn exceptions
• Use with pytest.warns(UserWarning): to catch known warnings
• Use -s flag to show stand output

2. Pytest attributes

• @pytest.fixture for predefine items can be used as function argument
  • Use autouse=True argument to auto requested by all tests, so you MUST supply augments of the fixture (if any) by mark.parameterize
  • See example below

    @pytest.fixture(autouse=True)
    def mock_v1_local_models(requests_mock: Mocker, base_url: str) -> None:
      requests_mock.get(
          f"{base_url}/models",
          json={"data": 123},
      )
    
    @pytest.mark.parametrize(
        "base_url",["bogus"],
    )
    def test_function(public_class: type):
      assert public_class().flag
    

• use pytest.mark.parameterize to set groups of arguments for test function or fixtures.
  • Example of (v11, v12) and (v21, v22) value groups for argument (param1, param2)

    @pytest.mark.parametrize("param1, param2,
    [("v11", "v12"),
     ("v21", "v22")])
    

  • It can also be used to set parameters for fixtures