Checking an error is raised by a test

[p-j-smith]

trafficstars

In the Defensive Programming section, there is a test for patient_normalise that checks a ValueError is raised when negative values are passed to the function:

with pytest.raises(raises):
    npt.assert_almost_equal(patient_normalise(np.array(test)), np.array(expected), decimal=2)

This could be simplified to:

with pytest.raises(raises):
    patient_normalise(np.asarray(test))

Also, it's best to check the specific error message being thrown:

with pytest.raises(raises, match='Inflammation values should not be negative'):
    patient_normalise(np.asarray(test))

And there is a check for whether raises is ValueError or None:

if raises:

However, it's better to always explicitly check that a variable is not None:

if raises is not None:

It makes no difference here, but other times some statements (e.g. []) could evaluate to False and lead to the wrong code block executed.

[steve-crouch]

Matt: Just need check the proposed solution still lines up.

[bielsnohr]

@p-j-smith thanks for this detailed issue and apologies it has taken us so long to get around to it! I've got a PR #451 up that should address everything. I have opted to put the stuff about "matching" in an optional exercise because I feel like it is just a bit of a stretch for everyone who might be taking the course, but could be really helpful for the more advanced learners. The point about using if expected_raises is not None was already incorporated into the course material at some other time, but regardless, thanks for pointing it out.