Test with parameters#

We called the tests "naive" before, simply because there are tons of features in pytest that can improve your tests.

We tested both European and African species in the first test. Following this logic, adding an Asian species will require adding an additional two lines there. There should be a more elegant way.

In the second test, we only check a single value of cargo_weight, how do we know if the other values also lead to correct results?

Let's investigate the decorator @pytest.mark.parametrize. This decorator, placed before a test, helps giving a range of values to use as parameters to the test. For instance:

@pytest.mark.parametrize("species", ["european", "african"])
@pytest.mark.parametrize("cargo_weight", [0, 0.1, 0.2, 0.3, 0.4])
def test_swallow_velocity(species, cargo_weight):
    """Test that the velocity of the swallow is correct."""
    swallow = Swallow(species=species)
    swallow.cargo_weight = cargo_weight

    # Can you tell why this test is not good?
    assert swallow.get_speed() == pytest.approx(60.0 / (1 + cargo_weight))


@pytest.mark.parametrize(
    "species, is_migratory", [("european", True), ("african", False)]
)
def test_swallow_migration(species, is_migratory):
    """Test that the European swallow is migratory, while the African swallow is
    not."""
    swallow = Swallow(species=species)
    assert swallow.is_migratory() == is_migratory

As you might have noticed, a parameter defined in @pytest.mark.parametrize needs to appear in the test function signature. Secondly, you can defined multiple parameters in one decorator, their values are then passed as tuples (("african", False)). If there are multiple @pytest.mark.parametrize decorators, then all values of one are run against the values of the other!

Run the tests!