Open bobturneruk opened 3 years ago
e.g. automation with GitHub actions https://github.com/pykale/pykale/blob/main/.github/workflows/test.yml
Ironically, despite being testers, we rarely test our own research code! I agree, though, that it's probably sensible and good practice to test what we can. Certainly the causecumber core (whatever that turns out to be) should be well tested, and probably contain more input validation to make it a bit less fragile and give more meaningful errors.
I'm usually the one who knows most about tests in a team. Not here.
My default suggestion would be to use pytest and have a
tests
folder that contains tests covering the code in the (TBD)causcumber
folder (see #12). That these tests should be configured to run automatically prior to code being integrated into themain
branch. It seems that there are plenty of functions that would be amenable to relatively straightforward unit tests.