music-tag is a library for editing audio metadata with an interface that does not depend on the underlying file format. In other words, editing mp3 files shouldn not be any different than flac, m4a, ... This library is just a layer on top of mutagen, which does all the heavy lifting.
The following file formats are actively tested.
aac
aiff
dsf
flac
m4a
mp3
ogg
opus
wav
wv
Metadata is available using a dictionary-like interface with the following keys.
Keys are not case sensitive and can contain arbitrary whitespace, '-', and '_'
characters. In other words, Album Artist
, album-artist
, and
album_artist
are all synonyms for albumartist
. Also, disk
is synonymous with disc
.
album
albumartist
artist
artwork
comment
compilation
composer
discnumber
genre
lyrics
totaldiscs
totaltracks
tracknumber
tracktitle
year
isrc
#bitrate
(read only)#codec
(read only)#length
(read only)#channels
(read only)#bitspersample
(read only)#samplerate
(read only)import music_tag
f = music_tag.load_file("music-tag/sample/440Hz.m4a")
# dict access returns a MetadataItem
title_item = f['title']
# MetadataItems keep track of multi-valued keys
title_item.values # -> ['440Hz']
# A single value can be extracted
title_item.first # -> '440Hz'
title_item.value # -> '440Hz'
# MetadataItems can also be cast to a string
str(title_item) # -> '440Hz'
# tags can be set as if the file were a dictionary
f['title'] = '440Hz'
# additional values can be appended to the tags
f.append_tag('title', 'subtitle')
title_item.values # -> ['440Hz', 'subtitle']
title_item.first # -> '440Hz'
title_item.value # -> '440Hz, subtitle'
str(title_item) # -> '440Hz, subtitle'
del f['title']
f.remove_tag('title')
Album artwork is wrapped in an object that keeps track of some of the extra metadata associated with images. Note that some album art functionality requires the Pillow (modern day PIL) library.
# get artwork
art = f['artwork']
# Note: `art` is a MetadataItem. Use ``art.value`` if there is
# only one image embeded in the file. This will raise a
# ValueError if there is more than one image. You can also
# use ``art.first``, or iterate through ``art.values``.
art.first.mime # -> 'image/jpeg'
art.first.width # -> 1280
art.first.height # -> 1280
art.first.depth # -> 24
art.first.data # -> b'... raw image data ...'
# set artwork
with open('music_tag/test/sample/imgA.jpg', 'rb') as img_in:
f['artwork'] = img_in.read()
with open('music_tag/test/sample/imgB.jpg', 'rb') as img_in:
f.append_tag('artwork', img_in.read())
# Make a thumbnail (requires Pillow)
art.first.thumbnail([64, 64]) # -> pillow image
art.first.raw_thumbnail([64, 64]) # -> b'... raw thumbnail data ...'
# finally, you can bounce the edits to disk
f.save()
By default, tags are validated and normalized. For instance, track numbers
and years are return as integers. Some tag formats store everything as strings
to enable things like leading zeros in tracknumbers (i.e., track '01'). I think
this is ugly, but you can use the file object's raw
property if you like
this kind of thing.
f.raw['tracknumber'] = '01'
f.raw['tracknumber'].value # -> '01'
Some tags may not exist in a file, but there could be enough information to
discern the correct value. For instance, the album artist
tag is probably
equal to the artist
tag, or "Various Artists"
if the compilation
flag is set. Here are some examples,
f['album artist'] = 'Brian'
f.resolve('album artist') # <- 'Brian'
f['artist'] = 'Brian'
del f['album artist']
f['compilation'] = False
f.resolve('album artist') # <- 'Brian'
f['compilation'] = True
f.resolve('album artist') # <- 'Various Artists'
del f['compilation']
f['album artist'] = 'Various Artists'
f.resolve('compilation') # <- True
f['album artist'] = 'Brian'
f.resolve('compilation') # <- False
The music_tag package can be used as a CLI to get / set tags. Here are some examples,
# Print tags from all audio files in sample directory
python -m music_tag --print ./sample
# Print specific tags from all audio files in sample directory
python -m music_tag --print --tags="Title : Album" ./sample
# Write tags from all audio files in sample directory to a csv file
python -m music_tag --to-csv tags.csv ./sample
# Write specific tags from all audio files in sample directory to a csv file
python -m music_tag --tags="Title : Album" --to-csv tags.csv ./sample
# Set a couple tags for multiple files
python -m music_tag --set "genre:Pop" --set "comment:cli test" \
./sample/440Hz.aac ./sample/440Hz.flac
# Write tags from csv file to audio files (assuming file paths in
# the csv file are relative to the sample directory
python -m music_tag --from-csv tags.csv