Accelerate mmCIF parsing

[rhiju]

Templates for tRNAs (which are super common!) often reside in ribosome mmCIF files, which take forever to parse with Biopython.

• Write parsers 'by hand' that rapidly grep for the _pdbx_poly_seq_scheme information and xyz coordinates. If we go this route, let's separate out the functions for doing this into something like cif_util.py so we can re-use in other codes.
• Less elegant: could use pre-split PDB chains and, if needed, pre-split _pdbx_poly_seq_scheme information. If we go this route, let's add the python scripts and workflow for the pre-splitting