Test interp nominal lon

JoranAngevaare commented 3 months ago

Fix test for `_interp_nominal_lon`

This PR is some leftover work from #296 (and #300)

I've added a dummy test - which is a slimmed down version of the example issue in #295.
Add type-hints to #296
Fix #358

PR checks

[x] Closes #295, #358
[x] Tests added
[x] Passes pre-commit run --all-files
[x] User visible changes (including notable bug fixes) are documented in whats-new.rst
[x] New functions/methods are listed in api.rst

JoranAngevaare commented 2 months ago

@jbusecke of course value-errors result in failed tests and this syntax does trigger the tests:

(py39) aangevaare@surfclimahuge:~/software/xMIP$ pytest tests/test_preprocessing.py -k TestReplaceXYNominalLatLon
================================================= test session starts ==================================================
platform linux -- Python 3.9.0, pytest-8.2.2, pluggy-1.5.0 -- /home/aangevaare/miniconda3/envs/py39/bin/python
cachedir: .pytest_cache
rootdir: /home/aangevaare/software/xMIP
configfile: pyproject.toml
collected 140 items / 138 deselected / 2 selected

tests/test_preprocessing.py::TestReplaceXYNominalLatLon::test_old_fails PASSED                                   [ 50%]
tests/test_preprocessing.py::TestReplaceXYNominalLatLon::test_new_works PASSED                                   [100%]

=================================================== warnings summary ===================================================
xmip/preprocessing.py:4
  /home/aangevaare/software/xMIP/xmip/preprocessing.py:4: UserWarning: Import(s) unavailable to set up matplotlib support...skipping this portion of the setup.
    import cf_xarray.units  # noqa: F401

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
===================================== 2 passed, 138 deselected, 1 warning in 8.37s =====================================

But sure, let me write assert statements if that's more conform the way things are done for xmip

JoranAngevaare commented 2 months ago

@jbusecke as requested:

I replaced the raise ValueError with assert 9faf75f.
To show that it works explicitly, I pushed 0270f4b which reverts the logic, and shows that the old code would have raised an error.
I think you asked me to remove the showcase of the old test. In that case, using a class is of course completely redundant so I wrote it as a single function in fda6fbc.

jbusecke commented 2 months ago

of course value-errors result in failed tests and this syntax does trigger the tests

Thanks for verifying this. And it obviously makes sense (Ill keep that in mind for future test designs, and thanks for teachibg me this!). I should have tested this myself (but am using gh on mobile so am a bit restricted here).

jbusecke / xMIP