Cell optimisation constraints need updating #394
Description
When optimising the simulation cell using fixed side-length ratios (e.g. fixing AtomMove.OptCell.Constraint c/a) the stress needs to be adjusted to maintain the ratio. There is some freedom in what we do with the stress, so I have decided to maintain the average stress of the two directions constrained.
For AtomMove.OptCellMethod 3 we need a further tweak so that the strains in the two directions are the same (this approach optimises the strain rather than the cell side lengths). Here I have just averaged the two.