Changelog

Additions

• Added support for Python 3.14. (#830)
• Added structure.dihedral_side_chain() analogous to dihedral_backbone(). (#827)
• Added structure.get_all_residue_positions() and structure.get_atom_name_indices() to residue level functionalities. (#827)
• Expanded list of 'water' residue names in structure.filter_solvent(). (#832)

Changes

structure.io.pdbx.get_structure(use_author_fields=False) assigns an incrementing res_id for hetero residues instead of -1. (#823)

• Previously in edge cases assigning -1 to hetero residues lead to missing distinction of subsequent residues. (#553)
• A different res_name is not used anymore to distinct two subsequent residues in the residue-level functionalities. (#825)
  • Previously, if two altlocs of the same residue had a different res_name, both would be incorrectly kept. (#824)

Fixes

• Fixed reading BCIF files written by py-mmcif in structure.io.pdbx.BinaryCIFFile. (#833)
• structure.io.pdbx.BinaryCIFData.array is now writable if it was encoded only with ByteEncoding. (#833)
• Fixed issue where for some invalid UniProt IDs database.rcsb.fetch() would still downloaded another entry. (#833)
  • For example instead of the invalid A0A12345, A0A123 would have been downloaded.
• structure.io.pdb.get_structure() now only infers bonds from CCD for proteins and nucleic acids, not for other ligands. (#820)
  • This allows using residue names for compounds that clash with the compound names in the CCD without getting wrong bonds.