I found the reason MP3 files weren't getting parsed out, heads up on this if you use Python 2.7! It only requires changing one line of code. First, here's what's going on...
Example:
fileName = "Z:\mp3\Art Blakey - Orgy In Rhythm\03 Toffi.mp3"
# fileName actually provided by navigating into a folder containing mp3s, which are processed
# in a "for" loop in mymedia.py, getdoc(), which calls the getart() and file2item() functions,
# which both ultimately instantiate the Mp3AudioFile class below
In eyeD3 module,
tag.py:
def isMp3File(fileName):
(type, enc) = mimetypes.guess_type(fileName)
return type == "audio/mpeg"
...
...
...
class Mp3AudioFile(TagFile):
def __init__(self, fileName, ... )
...
if not isMp3File(fileName):
raise InvalidAudioFormatException("File is not mp3")
In the Python 2.7 shell...
>>> import mimetypes
>>> mimetypes.guess_type("Z:\mp3\Art Blakey - Orgy In Rhythm\03 Toffi.mp3")
('audio/x-mpg', None)
mimetypes.guess_type() documentationAll you have to do is replace
return type == "audio/mpeg" with
return type == "audio/mpeg" or "audio/x-mpg" (line 1709 in
tag.py). According to
this, 'audio/x-mpg' is actually just an alias for 'audio/mpeg', the proper mp3 mimetype. Not sure why Python 2.7 returns this alias, causing
isMp3File(fileName) to return
false.
I notice someone else made this change to eyeD3
tag.py here, for the same issue. Actually, a
few other changes were made to purportedly make eyeD3 play better with Python 2.7.
So, I have to assume mimetypes.guess_type() will return ("audio/mpeg", None) in the Python 2.6 interpreter, which is a more appropriate result I think. When I get a chance I'll try confirming this on another computer.
