Ihm

=================
- :class:`ihm.Citation` now takes a ``is_primary`` argument, which can
be used to denote the most pertinent publication for the modeling.
- Improved support for non-standard residues, and for standard amino acids
used as nonpolymers.

=================
- Minor documentation improvements.
- Add support for the _struct.pdbx_structure_determination_methodology
mmCIF data item.

=================
- :func:`ihm.dumper.write` and :func:`ihm.reader.read` both now take
a ``variant`` argument which can be used to control the set of tables
that are read/written. This can be used by other libraries (such as
python-ma) to support other mmCIF extensions.

=================
- :func:`ihm.dictionary.Dictionary.validate` will now report errors for
any keywords or categories in the file that are not present in the
dictionary.
- :class:`ihm.LPeptideAlphabet` now supports the ASX and GLX ambiguous
residue types.

=================
- :class:`ihm.AsymUnit` now supports insertion codes in its
``auth_seq_id_map``. The target of this mapping can either be an
author-provided residue number (as previously) or a 2-element tuple
containing this number and an insertion code.
- :class:`ihm.AsymUnit` now allows the PDB or author-provided strand/chain ID
to be different from the regular ID.
- Bugfix: if two :class:`ihm.dictionary.Dictionary` objects both contain
information about a given category, adding the two dictionaries together
now combines the category information, rather than just using that from
one dictionary.
- Bugfix: :class:`ihm.dictionary.Dictionary` should now be able to validate
BinaryCIF files containing integer or float values (66).

=================
- Bugfix: _struct_ref.pdbx_seq_one_letter_code is now treated as the subset
of the reference (e.g. UniProt) sequence that overlaps with our Entities,
not the entire sequence (64).