Handle UNK residues in template sequences

[benmwebb]

As noted in #3, in some cases the alignment file contains some X residues in the template (the example given is from the Arabidopsis thaliana bulk download in NP_001030619.1_1.ali.xml, where the template is 4buj chain E).

We need the 3-letter name to output entity_poly_seq. Simply mapping X in the alignment to UNK, as proposed in #3, doesn't completely handle this though since we need to output (UNK) for entity_poly.pdbx_seq_one_letter_code but X for _entity_poly.pdbx_seq_one_letter_code_can (and I'm not sure what we should be writing out for ma_alignment.sequence; see ihmwg/ModelCIF#2). We'd also need to add UNK to chem_comp.