Allow Chimera selection specifications

[jaimergp]

Instead of requesting Molecule/residue_position or Molecule/atom_serial_number, allow Chimera selection commands when possible. Maybe this can be bone at the gaudi.genes.molecule.Compound level, but we first have to figure out how to implement it without breaking the existing find_* calls.

[jaimergp]

Partial support added to gaudi.genes.torsion in PR #4 to see how it would work... Seems OK so far, so this will get implemented soon, I'd say.