Add symmetry constraints for atomic coordinates

[rozyczko]

Currently, symmetry is internally enforced only for cell vectors. Atomic coordinates (fractional) are free to vary during the minimization.

Symmetry constraints need to also be added to atomic coordinates, possibly is Wyckoff positions.

Probaly the best idea is to check how Yurii did it in CrysPy and copy the solution, assuming it is correct.
Also, discussing this with Paul Erhart from Chalmers is a good idea.