Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

Powered by ReviewNB

This builds successfully locally, but asks you to choose a kernel when opening a notebook:

I find that kind of annoying and possibly confusing to newcomers. Wish there was a way to simply overwrite kernelname="python3"...

Is it the case that following instructions here https://github.com/xarray-contrib/xarray-tutorial/blob/main/CONTRIBUTING.md#create-a-python-environment, you'll still end up with the different kernel names if nb_conda_kernels is installed?

also noting that this PR addresses #78

If write a script to add name and display_name back to data.metadata.kernelspec, could set it up as a local pre-commit hook after nbstripout.

If using Jupyter Lab from the xarray-tutorial conda environment, name should be the default python3.

AH I just checked.

1. If I run from my base environment with nb-conda-kernels installed it asks to choose a kernel.
2. If I run from xarray-tutorial it just opens up the "default" kernel. I assume this will happen on binder too, but we can triple-check after merging.

Seems like we could recommend tutorial users open jupyter lab or jupyter notebook in the xarray-tutorial environment and then things will work well?

I assume this will happen on binder too, but we can triple-check after merging.

You can just launch binder pointing to this nbstripout branch: https://mybinder.org/v2/gh/xarray-contrib/xarray-tutorial/nbstripout?urlpath=lab

Sweet got the default kernel on binder.