Merge pull request #2845 from RosettaCommons/everyday847/revive_cif_io
Revive CIF I/O and fix some big bug classes
At final tally, this branch fixes about 8K new PDBs (~1/2 of broken PDBs) and 1K cifs (and thousands of others it moves to a more 'chemically realistic' bug class').
Massive bugfix to `metalloprotein_abrelax` which had been loading in the CCD type `pdb_CYZ` for its centroid_rot phase, and to various enumerative_sampling and related tests that had been loading in `pdb_VRT`.
A couple trajectory changes are observed due to floating point issues.
Some logfiles change due to spacing alterations in atom names. (Use stripped names in more places -- the RT will always have them in addition to the with-spaces alias, while the CCD residue will ONLY have the stripped version.)
Also prevented `PartitionContactScreener` from wiping out protein trajectories -- residues often won't make adjacent-residue contacts in typical (say) cyclic peptides.