implement structure minimization on the workflow

[David-Araripe]

As discussed previously, we would like to enable toggling an edge-dependent structure minimization of the protein's amino acids, while keeping the backbone rigid so we don't change the protein structure too much.

There was a demo on how to do something similar at the OMSF symposium in Naturalis. We could adapt the code from their jupyter notebook and implement on our codebase as well.

A bottleneck is that we need to be careful not to introduce clashes with the water sphere and I'm not sure if we could load our topologies with water molecules as shown in the example, since we use the spherical boundary condition.

[David-Araripe]

Probably more related, but still need to check if it would be ok to run this with 2 ligands (as we have on our topology) and to figure out if water molecules could be included:
https://github.com/openforcefield/anything-goes/blob/main/minimization-with-nagl.ipynb