python-ihm 2.7 release

- Representative models for :class:`ihm.model.ModelGroup` can now be
  denoted using the :class:`ihm.model.ModelRepresentative` class.
- Bugfix: fix reference counts of Python bool objects.

python-ihm 2.6 release

- The new :class:`ihm.License` and :class:`ihm.Disclaimer` classes
  allow describing how the data in the file can be used, and map to
  the ``pdbx_data_usage`` mmCIF table (#171).
- Two :class:`ihm.Assembly` classes are now considered equal if they
  contain the same set of constituents, even if those constituents are
  in differing orders or are split differently into domains (#170).
- The new :class:`ihm.format.ChangeFuncValueFilter` class allows modifying
  mmCIF values by passing them through an arbitrary Python function.
- Checks for non-canonical atom names no longer use Ligand Expo to obtain
  CCD info, as it is being retired (#169).
- :class:`ihm.metadata.CIFParser` and :class:`ihm.metadata.BinaryCIFParser`
  now return the same ``entity_source`` mapping that
  :class:`ihm.metadata.PDBParser` does (#168).

python-ihm 2.5 release

- Assemblies are now checked at file-output time to ensure that each
  asym referenced in an assembly is represented by atoms and/or spheres
  in at least one model (#165).
- ``make_mmcif`` has a new ``--check_atom_names`` option which, if set,
  will check for any non-canonical atom names (#166).
- ``pip install`` should no longer fail to install if the C extension
  module cannot be built; it will fall back to use the slower pure Python
  implementation.

python-ihm 2.4 release

- Add support for fits of model groups or ensembles to cross-links
  using the new :class:`ihm.restraint.CrossLinkGroupFit` class.

python-ihm 2.3 release

- The C-accelerated BinaryCIF parser now supports interval quantization
  encoding and a wider range of data types.
- ``make_mmcif`` now has a new ``--histidines`` option to convert any
  non-standard histidine names to HIS (#164).
- Bugfix: don't change name or description of the complete assembly
  in ``make_mmcif``.
- Bugfix: if :class:`ihm.reference.SeqDif` is used to annotate an
  insertion or deletion, don't erroneously claim the alignment is incorrect.

python-ihm 2.2 release

- The C-accelerated BinaryCIF parser now supports a wider range of data types.

python-ihm 2.1 release

- Minor fixes to the C-accelerated BinaryCIF parser.

python-ihm 2.0 release

- python-ihm no longer supports Python 2; the latest supported version
  is now Python 3.6 (#161).
- BinaryCIF files are now read in using a C-accelerated parser, which is
  faster and more memory efficient than the older Python parser (#160).
- The tokenizers for mmCIF and BinaryCIF now return data of the correct
  type (int, float, bool, or string); previously, all values were returned
  as strings. :class:`ihm.reader.Handler` subclasses now request data of
  the correct type using Python type annotations. The API of the
  C-accelerated parsers has changed accordingly (#162).
- The new :class:`ihm.metadata.BinaryCIFParser` class can extract metadata
  such as database IDs and template information from BinaryCIF files, in
  a similar fashion to the existing :class:`ihm.metadata.CIFParser`.
- Information about a deposited structure, such as the deposition date,
  can now be read from :attr:`System.database_status` (#163).
- The new :class:`ihm.format.ReplaceCategoryFilter` class can be used to
  completely replace or remove an mmCIF category when using
  :class:`ihm.format.CifTokenReader`.

python-ihm 1.8 release

- Support added for datasets containing EPR, X-ray diffraction, footprinting
  or predicted contacts using the :class:`ihm.dataset.EPRDataset`,
  :class:`ihm.dataset.XRayDiffractionDataset`,
  :class:`ihm.dataset.HydroxylRadicalFootprintingDataset`,
  :class:`ihm.dataset.DNAFootprintingDataset` and
  :class:`ihm.dataset.PredictedContactsDataset` classes (#157).
- Revision information (in the ``pdbx_audit_revision_*`` mmCIF tables)
  can now be read or written using the new :class:`ihm.Revision`
  and :class:`ihm.RevisionDetails` classes (#156).
- The new :class:`ihm.location.BMRbigLocation` class allows for
  referencing datasets stored in the BMRbig database.
- All references to the old PDB-Dev database are now updated to PDB-IHM.

python-ihm 1.7 release

- Sanity checks when writing out a file can now be disabled if desired,
  using the new ``check`` argument to :func:`ihm.dumper.write` (#153).
- Data that have been split over multiple mmCIF or BinaryCIF files can now
  be combined into a single :class:`ihm.System` object using the new
  ``add_to_system`` argument to :func:`ihm.reader.read`.
- Input files that assign multiple :class:`ihm.location.Location` objects
  to a single :class:`ihm.dataset.Dataset` can now be read (#151).
- Bugfix: multiple :class:`ihm.restraint.EM3DRestraint` and
  :class:`ihm.restraint.SASRestraint` objects can now be created for a
  single dataset, as long as they act on different assemblies, as allowed
  by the dictionary.
- Bugfix: allow for non-standard residues in the ``struct_ref`` table (#154).