Enable Water Sampling in Simulations

[badisa]

• Disables water sampling for solvent, doesn't seem to make a difference
• Some random clean up of the example scripts that were from trying to use them to test water sampling
• Runs TIBDExchangeMove on all windows
• Defaults to 1000 MC proposals per 400 steps of MD which makes for about 10% additional runtime cost. Performance optimizations intended in a follow up.

Follow up questions

• How do we effectively target a region that will produce effective moves (note that probability of placing in the inner region is the furthest from the center)? Currently defaults to 1.0nm from the centroid of the ligand (want this to be ligand based, to be generic), probably in practice take the max pair wise distance of the ligands and add a padding (~0.4nm).
  • Select ligand atoms that are 'buried' in the protein to define sphere where we think it is important, unclear how to define
  • Randomize the center of the sphere/radius (no API available short of reconstructing the entire class)

Todo

• [x] Add a test with a 3 atom ligand (including dummy atoms)

[proteneer]

awesome :) we now have a legit water sampler!