Documentation for nested config stores

[markkvdb]

Hi,

First of all, thank you for this tool. Even though, all the concepts within Hydra are a bit hard to grasp in the beginning, I'm beginning to see the philosophy and what the tool unlocks.

I'm trying hydra-zen out to see if and how it would fit in my data science projects and I'm wondering how you would handle the configuration of a scikit-learn pipeline. I would like to do experiments to see what happens when I change parameters of a pipeline step and/or removing steps in the pipeline entirely.

How would you setup the hydra configuration in this case?

I hope my question is clear and I'd be happy to explain myself further if needed.

Cheers, Mark