The current tests are standard unit tests, running methods of each class with certain inputs, programatically expecting certain ouputs; accounting for certain edge-cases present within the intrinsic nature of magpylib. These make sure that the codebase is behaving as expected whenever changes are introduced.
It is important that for every introduced method there should be at least one test case showcasing its main functionality.
Tests are implemented utilizing pytest. The automation interface tox is utilized to create and install the library under a new environment, making sure the installation process works before running the tests. With our configuration, a code coverage html report is also generated when calling tox, demonstrating how much of the codebase has been tested.
To run rests, simply run pytest or run the tox interface .
$ pytest
$ tox
Automated "checks" are done by the continuous integration service CircleCI whenever a push is made to any branch on the remote repository at GitHub. This makes sure the Codebase always remains tested.
CircleCI will utilize the configuration file, running all the steps necessary (seen above) to properly perform the tests in a cloud environment.
If a test in a latest commit fails in the cloud environment, a ❌ will appear next to responsible branch being tested, letting maintainers know the code is behaving unexpectedly and should be looked into before merged into a development or release branch.