Mechanism for relaxing to subgroup symmetry

[Jakob-Unfried]

We have some implementations, e.g. change_symmetry.

In particular, the Space.change_symmetry framework is not ready though, since it assumes a sector is mapped to a single sector only, not a direct sum, as e.g. SU(2) -> U(1) would mean.

Consult discussion with Ian!