Subsampling documentation

[Bernadette-Mohr]

Would it be possible to add examples (ideally for multidimensional df) to the documentation of alchemlyb.preprocessing.subsampling?
At least I am not able to guess the usage of statistical_inefficiency() correctly.

[dotsdl]

Hi @Bernadette-Mohr, welcome to alchemlyb!

I agree that this is sorely needed. It's been on my radar for some time, but your immediate need pushes the priority up for this. I have added producing a PR addressing #63 to my list for this weekend; this should serve as a good concrete example addressing for this issue.

Are you interested in providing feedback once the PR is up?

[Bernadette-Mohr]

Of course, I'll be glad to help!
And yes, generally adding examples for multidimensional dataframes would be awesome.

[Bernadette-Mohr]

Hi @dotsdl ,

Thank you for adding the multi-lambda example to the parser documentation! It confirms what I thought the contents were.
But it still doesn't help me with my original question: How would I extract uncorrelated samples from this dataframe with the subsampling module?
No matter what I tried, it either throws an error or returns me the same 2-3 rows for every lambda step, so probabely just acting like slicing.

[Bernadette-Mohr]

Sorry to complain again, but a hint at how to extract the individual contributions of VdW and Coulomb interactions would be helpful too.
I guess I would have to split the multiindex dataframe by coul-lambda=1.0 AND vdw-lambda != 0.0, but that's just a guess.