Add code coverage to tests

[sappelhoff]

related to JOSS review (openjournals/joss-reviews#2621)

It would be great to add a tool that measures what proportion of your code is actually covered by the tests.

It's helpful to:

• give users a better impression of how far they can "trust" the tests
• give the developers a better idea of what aspects of the code are tested, and what aspects are not tested

This would be quite easy to implement, e.g., using pytest extensions and codecov as in this repository: https://github.com/sappelhoff/pyprep

• you'd need pytest-cov (https://github.com/sappelhoff/pyprep/blob/da8e3a64cac7759e553d9dc219a88edc13f3b6cb/requirements-dev.txt#L10)
• you'd need to extend your GitHub Actions workflow (example: https://github.com/sappelhoff/pyprep/blob/da8e3a64cac7759e553d9dc219a88edc13f3b6cb/.github/workflows/python_tests.yml#L56-L60)

Finally, you could add a coverage badge next to your tests:

[TomDonoghue]

Somewhat related: it might be worth adding a note somewhere on the requirements for running the tests locally. As well as the standard of needing pytest, when trying to run the tests I got an error that fixture 'qtbot' not found, which I tracked to being because pytest-qt is also required to run the tests. This isn't clearly noted anywhere.

Somewhere, I would add a note of test-requirements, and maybe some instructions on running tests locally (this is perhaps most relevant in a contributor guide, as you would likely want contributors to run and check tests.

As a note, I was able to run tests locally. I'm not sure of the coverage (adding that is a good idea), though the GUI tests looked (in terms of how many things popped up) fairly comprehensive.

[JanCBrammer]

it might be worth adding a note somewhere on the requirements for running the tests locally

@TomDonoghue, I've included testing instructions in the contributor guide (see 6685c6e).

[JanCBrammer]

It would be great to add a tool that measures what proportion of your code is actually covered by the tests.

@sappelhoff, I now collect and report code coverage. However, the coverage is deceptively low (59%) since pytest-cov does not seem to pick up on code that is run within threads (see here). Looking into it.

[JanCBrammer]

Closing and transfering to #21, since original issue has been addressed.