Kantag

* Ported to Python 3; support for Python 2 has been dropped.
* Removal of non-ASCII punction is now also applied to artist names.
* Added an option to initkan to pass the MusicBrainz release ID on the command
line, rather than reading it exclusively from existing tags.
* Added support for Compilation tags.
* Added support for ISRC/TSRC tags.
* Added support for AcoustID tags.
* Added support for an AlbumVersion tag to distinguish between different
editions of the same album.
* Fixed handling of ID3V2 UFID tags used for MusicBrainz recording ids. Also,
since foobar2000 doesn't support UFID tags, a TXXX:MusicBrainz Track ID tag
is also written to the file.
* All reading of ID3 tags is now case-insensitive.
* Fixed an exception generated when requesting MusicBrainz lookup when no
track number could be determined.

* Ported from deprecated python_musicbrainz2 to python_musicbrainzngs. "NGS"
was a very big change for MusicBrainz, and some kantag APIs needed to change
drastically as a result. The upside is more features, better data, and
greatly improved performance in many cases.
* initkan: Add '--locale' to retrieve locale-specific artist aliases when
using '--call-musicbrainz'. May not apply to all artists.
* initkan: Argument '--parse-disc' now defaults off, as it was designed for
a release title style that was deprecated at MusicBrainz long ago. Still
available for dealing with legacy data in local files.
* initkan: Add '--track-artists' argument to support multi-value 'Artists' and
'AritstsSort' tags for tracks.
* initkan: Add '--ascii-punctuation' argument to support translation of
non-ascii punction in titles (release title, track title, work title, etc.)
with ascii equivalents. Code borrowed from MusicBrainz Picard. Enabled by
default.
* initkan/showkan: No longer dump large blocks of data for embedded images.
* initkan: Condense items in the artist and work blocks, such that tracks
with the same tag/value will output with a single multi-track line instead
of one line for each track.
* Use MusicBrainz recording disambiguation as the Version tag, if available.

* Fix detection of various artist releases.
* Add support for acoustid_id, Writer, musicbrainz_trackid,
musicbrainz_releasegroupid, and musicbrainz_discid tags.
* Change shebangs back to '/usr/bin/env python2' for systems (e.g., funtoo)
that default to python3.
* initkan: Output OriginalDate at track level if it cannot be rolled to disc or
release level (e.g., a re-release that has multiple albums on a single disc).
* initkan: Fix output of OriginalDate if there is no Date at the same level.
* initkan: Fix double output of Date tag with multi-disc releases.
* When reading non-MP3 files, map Remixer to Arranger. This was already
happening for MP3s, as well as for remixer relationships pulled from the
MusicBrainz web service.
* Add ISRC to recognized tags; unsupported by initkan.
* Add support for ID3 TDOR frame.
* Add support for additional tags supported by Picard, such as multi-value
AlbumArtists tags, work title, musicbrainz_releasetrackid,
musicbrainz_albumstatus, and musicbrainz_albumtype.
* Add basic support for total tracks/total discs. initkan will not write them
by default, but will read them without warning.
* Fix bug that caused an error in path inference function.
* initkan: Fix crash if no musicbrainz ids are present.
* Change tag inference to set title if the exsting title starts with "Track ".
* Add MP4 file support.
* Add support for ID3 TXXX:original date.

* Change shebangs to '/usr/bin/env python' since cygwin does not have a python2
symlink.
* Remove module imports from __init__.py so that tools showkan can work without
python-musicbrainz2.

* Initial release.