Parametrizing pytest with json

pytest parameterization is a very powerful tool for python unit tests.

Sometimes the standard decorator technique starts to get unwieldy:

@pytest.mark.parametrize(
    "test_input,expected",
    [
        ("3+5", 8),
        ("1+1", 2),
        ("2*6", 12),
        ("10-2", 8),
        ("100/10", 10),
        ("2+4", 6),
        ("6*9", 42),
    ],
)
def test_eval(test_input, expected):
    assert eval(test_input) == expected

By using json files for tests cases you can keep unit test code simple and small. Test code is almost as important at the 'actual' code, so it's critical to keep test code understandable.

Example:

This is an example of using a json file to parameterize pytest.

It reads in the json file and creates parameters for each test case.

One important thing to note is the id function

def id_func(item: dict) -> str:
    """ Returns a unique id for each test, just make testing clear """
    return item["input"]

This will give each test an id so it's easy to see what failed with a pytest (in this case jbr)

$ pytest -v
==================================================================== test session starts =====================================================================
platform darwin -- Python 3.7.4, pytest-5.1.1, py-1.8.0, pluggy-0.12.0 -- /usr/local/opt/python/bin/python3.7
cachedir: .pytest_cache
rootdir: /example-pytest-json-input
plugins: dash-1.0.2, cov-2.7.1, requests-mock-1.6.0
collected 3 items                                                                                                                                            

test_json_input.py::test_greeter[John] PASSED                                                                                                          [ 33%]
test_json_input.py::test_greeter[jbr] FAILED                                                                                                           [ 66%]
test_json_input.py::test_greeter[Zachariah] PASSED                                                                                                     [100%]

========================================================================== FAILURES ==========================================================================
_____________________________________________________________________ test_greeter[jbr] ______________________________________________________________________

test_case = {'input': 'jbr', 'output': 'Hi jbr!'}

    def test_greeter(test_case):
        """ Do we introduce customers correctly? """
>       assert greeter(test_case["input"]) == test_case["output"]
E       AssertionError: assert 'Howdy jbr!' == 'Hi jbr!'
E         - Howdy jbr!
E         + Hi jbr!

test_json_input.py:27: AssertionError
================================================================ 1 failed, 2 passed in 0.38s =================================================================

This simple example uses json, but it's easy to adapt this to other test case formats.