I currently use MRAN snapshots to help maintain some my R enviroments, which are clones of the CRAN R package repository throughout different times in history. This way I can make sure all the package/dependencies are in sync and I can have reproducible results.
Here is a link -> https://mran.microsoft.com/documents/rro/reproducibility
I also maintain different versions of python and was wondering if something like this existed for python. I currently just do conda install numpy but I realized that eventually the dependencies can get out of sync.
Any recommendations or thoughts?