Would be great to add some unittests for the production bits of the code. Perhaps as a resonable first go, aim for >50% coverage.
Perhaps a way to breka up the test could be
The prepocess of data step should be tests
the running of the model
the saving of the data (this can come later)
You'' have to think a bit how to mock the input data. You can either save some small files in git, or add some conftest.py pytest fixtures that will make these
Would be great to add some unittests for the production bits of the code. Perhaps as a resonable first go, aim for >50% coverage.
Perhaps a way to breka up the test could be
You'' have to think a bit how to mock the input data. You can either save some small files in git, or add some conftest.py pytest fixtures that will make these
pytestis a good frame work for running this.CI can be set up like this