I added in some automation and testing that might make it faster + easier to review the currently open PRs in this project. It might be good to start adding tests to the project to verify the correctness and prevent regressions.
Adds pytest for testing (seems easiest to use + good docs and examples)
Write basic integration type test based off smogn_example_1_beg.ipynb. Could be useful to detect basic regressions but further tests needed to guarantee correctness on the basis of the actual algorithm substance. Only 1/3 examples are executed but it at least sets up the framework for automated testing of this library 😬
Setup simple test + setup execution through Makefile; eg: just run make test to fire off tests... could be useful to create a GH Action in the future to build, package, and upload the repo automatically on main pushes
note: might be useful to use auto lint or code format later since flake8 was logging lots of warnings on my GH Actions test runs
Create requirements.txt with proper version freezing (per numpy, pandas in roughly April 2020) for pip in order to use dependency resolution when installing via pip.
I added in some automation and testing that might make it faster + easier to review the currently open PRs in this project. It might be good to start adding tests to the project to verify the correctness and prevent regressions.
Adds
pytestfor testing (seems easiest to use + good docs and examples)Write basic integration type test based off
smogn_example_1_beg.ipynb. Could be useful to detect basic regressions but further tests needed to guarantee correctness on the basis of the actual algorithm substance. Only 1/3 examples are executed but it at least sets up the framework for automated testing of this library 😬Setup simple test + setup execution through Makefile; eg: just run
make testto fire off tests... could be useful to create a GH Action in the future to build, package, and upload the repo automatically onmainpushesAdds rudimentary GitHub Actions (adapted from official example) automation to run tests that runs automatically on branch push, using matrix builds for all currently supported Python versions. Might be worth looking into tox for doing this locally though...
Create
requirements.txtwith proper version freezing (per numpy, pandas in roughly April 2020) for pip in order to use dependency resolution when installing via pip.Removed deprecated fields in
setup.py