Implement pdb_shiftres in preprocessing gear

[joaomcteixeira]

I decided to leave pdb_shiftres out from #144 because it is difficult to decide on an effective behaviour and the implications of all possible implementations. So, I am making a new issue for this. Possible questions:

• should all residues be shited to 1?
• should residue shift be a parameter?
• can we implement shift res without implementing automatic update .tbl when preprocessing PDBs #495?

Definitively, pdb_reres should not be implemented because it would break the continuity of residue numbering across chains.

See also #143