You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I propose that we integrate tests into our documentation through doctests rather than rely on converting jupyter notebooks to python and then running them.
This has a few benefits:
We can remove the need to keep and track the jupyter notebooks in the examples directory. These do provide some benefit, so we may want to move them somewhere, or figure out if we can generate a notebook from the documentation.
The main issue with keeping jupyter notebooks and example documentation is that we have to update both in the event of API changes. The notebooks are difficult to diff unless you use nbdime. I am open to keeping the notebooks, but then we have to make sure that, when there are API changes, both the documentation and the example notebooks get updated.
The text was updated successfully, but these errors were encountered:
I propose that we integrate tests into our documentation through doctests rather than rely on converting jupyter notebooks to python and then running them.
This has a few benefits:
pytest
via ENH Transition to pytest for all testing #325, we can pick these up in the project testing and CI -pytest
doctest integrationThe main issue with keeping jupyter notebooks and example documentation is that we have to update both in the event of API changes. The notebooks are difficult to diff unless you use
nbdime
. I am open to keeping the notebooks, but then we have to make sure that, when there are API changes, both the documentation and the example notebooks get updated.The text was updated successfully, but these errors were encountered: