Description

• Switch output file format from PDB to CIF
• Use fasta headers as entity descriptions in output CIF files
• Add the following under LocalLDDT qa_metric: C\alpha LDDT for proteins, C1' LDDT for nucleic acids
• Keep a per-atom LDDT under B-factor column as the ligand LDDT computed per-atom is more informative than the average over the ligand. Per-atom metrics can't be written as "Local" quality assessment metrics in python-modelcif API (https://python-modelcif.readthedocs.io/en/latest/qa_metric.html#modelcif.qa_metric.Local)

Test plan

Tested on a diversity of inputs: proteins, modified residues, DNA, RNA, ligand, complexes and multiple chains from same entity.