According to Martin, we can use papermill to inject parameters in slow notebooks to avoid excessively long computations, i.e. using papermill to set max_nfev=1 in the pumping test notebooks, which will then just test if the calibration runs 1 step. That way we could include all notebooks in the tests, but configure the slow ones to run quickly in the tests.
According to Martin, we can use papermill to inject parameters in slow notebooks to avoid excessively long computations, i.e. using papermill to set
max_nfev=1in the pumping test notebooks, which will then just test if the calibration runs 1 step. That way we could include all notebooks in the tests, but configure the slow ones to run quickly in the tests.