adding mamba install option

[dmitryhits]

Is there a reason to use 'bash' as the processing shell? For one of my nbdev projects I need a package that is not available on pypi but is available on conda (mamba). I have an idea on how to add a mamba install option, but it needs 'bash -l' as the processing shell (login shell) in order to activate mamba environment.
At them moment I am using a hack (https://github.com/dmitryhits/HighResAnalysis/blob/master/.github/workflows/deploy.yaml) but I think it would be nicer if this option was available through a central nbdev repo.

P.S. I am referring to quarto-ghp action.

[michaelaye]

Mamba is best set up via https://github.com/conda-incubator/setup-miniconda
Have a look at my longer workflow that I needed to come up with so that I can have all dependencies installed quickly via mamba here: https://github.com/michaelaye/nbplanetary/blob/master/.github/workflows/deploy.yml

However, I am just today learning how to setup micromamba, which should (could?) accelerate this further, specifically if one can use pre-calculated envs from prefix.dev.

[dmitryhits]

I think that in spirit your solution is similar to mine ;-)
I more wondering whether the nbdev developers want to incorporate this in their workflows.