Mutagen

Where can we find examples on how to use the Mutagen library?

http://svn.sacredchao.net/svn/quodlibet/trunk/mutagen/TUTORIAL

Any plans for python 3.0 compatibility?

a small example querying a flac file

#!/usr/bin/env python
from mutagen.flac import FLAC
from pprint import pprint
import sys

DEBUG=False

if len(sys.argv) > 1:
    fname = sys.argv[1]
else:
    fname = "/Users/philou/Music/Lhasa/La Llorona/01 De Cara A La Pared.flac"

audio = FLAC(fname)
if DEBUG:
    print "DD: instancied a mutagen object of type %s for file %s" % (type(audio), fname)
    print "DD: pprint(instance):"
    pprint(audio)
    print "/DD"
    print

if audio.has_key('artist'):      print "Artist:              ", audio['artist'][0]
if audio.has_key('album'):       print "Album:               ", audio['album'][0]
if audio.has_key('tracktotal'):
    tracktot = int(audio['tracktotal'][0])
else:
    tracktot = None
if audio.has_key('tracknumber'):
    trackn = int(audio['tracknumber'][0])
    if tracktot != None:
        track_str = "%s / %s" % (trackn, tracktot)
    else:
        track_str = str(trackn)
    print "Track:                %s" % track_str
if audio.has_key('title'):       print "Title:               ", audio['title'][0]
if audio.has_key('mcn'):         print "MCN:                 ", audio['mcn'][0]
if audio.has_key('isrc'):        print "ISRC:                ", audio['isrc'][0]

print "Sample rate:         ", audio.info.sample_rate
print "Nr. of channels:     ", audio.info.channels
print "Bits per sample:     ", audio.info.bits_per_sample
print "Total nr. of samples:", audio.info.total_samples
print "Length in secs:      ", audio.info.length
print "MD5:                 ", audio.info.md5_signature

An apparent source of confusion is the relationship between mutagen.mp3, id3, and easyid3. mp3.MP3 should be used to open an MP3 file; in addition to reading ID3v2 tags using either of the two ID3 parsers in Mutagen, it also reads additional information from the actual MPEG stream which may not be in the ID3 tags (like bitrate and duration).

An mp3.MP3 object behaves like a dict, like ID3 and EasyID3; in fact, the methods for getting and setting keys and values on an MP3 appear to simply map to the underlying ID3 parser's. If the ID3 tags you're working are on MP3 files, stick to using mp3.MP3 instead of using ID3 objects directly.

An mp3.MP3 object also has the method add_tags(), which allows you to add ID3 tags to a file that doesn't have any. This is the only way to add ID3 tags within Mutagen, AFAIK.

Also: mp3.add_tags has an optional argument 'ID3', which works just like the same arg on the constructor of an mp3 object, viz:

from mutagen.mp3 import MP3
from mutagen.easyid3 import EasyID3
import mutagen.id3

def test(filename):
   """Example which shows how to automatically add tags to an MP3 using EasyID3."""

   m = MP3(filename, ID3=EasyID3)

   try:
       m.add_tags(ID3=EasyID3)
       print("Added tags to %s" % filename)
   except mutagen.id3.error:
       print("%s already had tags" % filename)

   m['title'] = 'Test!'
   m.save()
   print(m.pprint())
   print('\n')

test('has-tags.mp3')
test('no-tags.mp3')

Any chance this library could ever be licensed as LGPL?

An example of writing tags to an ASF/WMA file in a format that ffmpeg knows how to decode:

from mutagen.asf import ASF 
audio = ASF("myfilename.wma")
audio["Title"] = "MyTitle"
audio["Author"] = "MyArtist"
audio["WM/AlbumTitle"] = "MyAlbum"
audio["WM/TrackNumber"] = "12"
#print ASF.pprint(audio)
#print ASF.keys(audio)
ASF.save(audio)

I use the "Author" tag for "Artist", as that happens to be the mapping used by musicpd (mpd) when it uses ffmepg AVFormatContext() to get the tags from the wma file.

How about a new page for MutagenExamples? ? I'd contribute a simple image extractor:

#!/usr/bin/python
import os
import sys
import locale
from optparse import OptionParser

def main(argv):
    from mutagen.id3 import ID3

    parser = OptionParser()
    (options, args) = parser.parse_args(argv[1:])
    if not args:
        raise SystemExit(parser.print_help() or 1)

    enc = locale.getpreferredencoding()
    for filename in args:
        print "--", filename
        try:
            f = ID3(filename);
            for frame in f.getall("APIC"):
                print frame.pprint().encode(enc, 'replace')
                ext = ".img"
                if (frame.mime == "image/jpeg") or (frame.mime == "image/jpg"): ext = ".jpg"
                if frame.mime == "image/png": ext = ".png"
                if frame.mime == "image/gif": ext = ".gif"
                (proot,pext) = os.path.splitext(filename)
                imgfilename = proot + ext
                if os.path.exists(imgfilename) : print "File " + imgfilename + " exists. Skipping!"
                else:
                    print "Writing image to " + imgfilename + " !"
                    myfile = file(imgfilename, 'wb')
                    myfile.write(frame.data)
                    myfile.close()
        except AttributeError: print "- Unknown file type"
        except KeyboardInterrupt: raise
        except Exception, err: print str(err)
        print

main(sys.argv)

I'd like a way to determine if there is an id3v1 tag, even when an id3v2 tag is also present. For example, mutagen.id3.ID3.version will return (2,3,0) when both v1 and v2 tags are present. Any suggestions on how to check if an id3v1 tag is also there? Thanks.

What about COMM tags? As it is now, it converts only tags starting with T, but COMM too can contain text. I personally did something like this to make it work on my MP3s, but have no idea if it makes sense:

Index: mid3iconv
===================================================================
--- mid3iconv   (revision 4351)
+++ mid3iconv   (working copy)
@@ -62,9 +62,9 @@
                 print str(err)
             continue

-        for tag in filter(lambda t: t.startswith("T"), id3):
+        for tag in id3:
             frame = id3[tag]
-            if isinstance(frame, mutagen.id3.TimeStampTextFrame): # non-unicode fields
+            if isinstance(frame, mutagen.id3.TimeStampTextFrame) or not hasattr(frame, "text"): # non-unicode fields
                 continue

             try:

Is there any way to read data from an MP3 file that is already in memory, stored in a file-like object? I've managed to pass the object to MPEGInfo just fine, but have been unable to do so for MP3 - which has made retrieving the ID3 information kind of difficult.

Hi, here a quick hint how to add iTunes compatible lyrics to an MP3:

lyrics = u"""
Some text
here
does not really
mattah
"""

tagg = id3.ID3("some.mp3")
tagg[u"USLT::'eng'"] = id3.USLT()     # add a new unsynchronized lyrics element
tagg[u"USLT::'eng'"].text = lyrics
tagg[u"USLT::'eng'"].encoding=1       # i guess it means utf8
tagg[u"USLT::'eng'"].lang='eng'       # seems to be used for everything
tagg[u"USLT::'eng'"].desc=u''         # dont know what the description is good for
tagg.save()

Works for me, didnt took the time to read the specs.

Hey r3inforce... adding lyrics that way to an mp3 does not work if the lenght of the lyrics is bigger than (somewhat) characters...at least iTunes shows blank lyrics. Do u know another way?

Hey grabelo, I think mutagen compress all tags greater than 2K. So iTunes can't work with this tags. I found this out as I can't see the album art in some players. As I comment out the rows in id3.py, which do the "zipping", it works :-)

Is there a way to check the lang of an existing USLT frame?

I have a collection of mp3 files that have XXX,eng and nld as the lang. It might be possible that if my program I have is used in another locale or something, it will not be able to read the USLT frame and this won't get any lyrics.