pyOpenSci
Test examples
Begin work on #249 (and #248 to some degree) by adding a few Pytest-based unit tests to the "pure-hatch" example code. I added some dummy data to use to test the actual "mean" calculations, since I thought a more complex mocking / faking example would probably be out of scope.
Moving forward, I'd like to add the tests to CI and add nox tests for the actual packaging steps.
As this is my first PR for the project, please let me know anything I might need to clean up!
Summoning @lwasser: I thought we had recommended importlib.resources for exactly this scenario: 5kb of CSV files are used in testing. But our tests page Only recommends a "repository such as Figshare or Zenodo" and then using Pooch. But that seems like over kill for cases such as this. And although the package here is artificial, I don't think this case is very unusual.
Summoning @lwasser: I thought we had recommended
importlib.resourcesfor exactly this scenario: 5kb of CSV files are used in testing. But our tests page Only recommends a "repository such as Figshare or Zenodo" and then using Pooch. But that seems like over kill for cases such as this. And although the package here is artificial, I don't think this case is very unusual.
Summoned 🧞♀️ 😆 hi! So the data section of our guide needs some real work - there is a pr here that we can work on together. What i suggest here @ucodery is that if you think importlib.resources is the best approach, let's use it
then let's open an issue to update the tests page to reference that. And then let's work on that data pr that @NickleDave started to get it merged in the upcoming months (maybe by the end of SciPy even!)
Does that sound reasonable to you? I trust your thoughts on best ways to include data in packages.
