Average trajectory positions

[amritagos]

Thermal fluctuations could possibly result in the erroneous identification of a hydrogen bond as "broken", even with a permissive criterion. This, in turn, affects the identification of ion pairs, since this is accomplished by finding the shortest path among connections (defined by distance cutoffs and hydrogen bond).

As an alternative to quenching trajectories, it is possible to also average trajectories over a small number of time-steps (say, 10), which should alleviate the issue with thermal fluctuations. There should be some convenience functions to produce a new 'averaged trajectory' (essentially a vector of System objects in James). This could be in a new file in soluanalysis, perhaps called util.py.