cc-4105: refactored metadata into it's own file

This commit is contained in:
Rudi Grinberg 2012-07-19 15:41:09 -04:00
parent 09303a96d3
commit 1c09907fd5
5 changed files with 149 additions and 22 deletions

View File

@ -1,9 +1,8 @@
# -*- coding: utf-8 -*-
import os
import mutagen
import abc
from media.monitor.exceptions import BadSongFile
from media.monitor.pure import LazyProperty
from media.monitor.monitor import Metadata
class PathChannel(object):
"""a dumb struct; python has no record types"""
@ -13,27 +12,12 @@ class PathChannel(object):
# It would be good if we could parameterize this class by the attribute
# that would contain the path to obtain the meta data. But it would be too much
# work for little reward
# work
class HasMetaData(object):
# TODO : add documentation for HasMetaData
__metaclass__ = abc.ABCMeta
# doing weird bullshit here because python constructors only
# call the constructor of the leftmost superclass.
@LazyProperty
def metadata(self):
# Normally this would go in init but we don't like
# relying on consumers of this behaviour to have to call
# the constructor
try: f = mutagen.File(self.path, easy=True)
except Exception: raise BadSongFile(self.path)
metadata = {}
for k,v in f:
# Special handling of attributes here
if isinstance(v, list):
if len(v) == 1: metadata[k] = v[0]
else: raise Exception("Weird mutagen %s:%s" % (k,str(v)))
else: metadata[k] = v
return metadata
return Metadata(self.path)
class BaseEvent(object):
__metaclass__ = abc.ABCMeta

View File

@ -0,0 +1,132 @@
# -*- coding: utf-8 -*-
import mutagen
import math
import copy
from media.monitor.exceptions import BadSongFile
from media.monitor.log import Loggable
import media.monitor.pure as mmp
"""
list of supported easy tags in mutagen version 1.20
['albumartistsort', 'musicbrainz_albumstatus', 'lyricist', 'releasecountry',
'date', 'performer', 'musicbrainz_albumartistid', 'composer', 'encodedby',
'tracknumber', 'musicbrainz_albumid', 'album', 'asin', 'musicbrainz_artistid',
'mood', 'copyright', 'author', 'media', 'length', 'version', 'artistsort',
'titlesort', 'discsubtitle', 'website', 'musicip_fingerprint', 'conductor',
'compilation', 'barcode', 'performer:*', 'composersort', 'musicbrainz_discid',
'musicbrainz_albumtype', 'genre', 'isrc', 'discnumber', 'musicbrainz_trmid',
'replaygain_*_gain', 'musicip_puid', 'artist', 'title', 'bpm', 'musicbrainz_trackid',
'arranger', 'albumsort', 'replaygain_*_peak', 'organization']
"""
airtime2mutagen = {
"MDATA_KEY_TITLE": "title",
"MDATA_KEY_CREATOR": "artist",
"MDATA_KEY_SOURCE": "album",
"MDATA_KEY_GENRE": "genre",
"MDATA_KEY_MOOD": "mood",
"MDATA_KEY_TRACKNUMBER": "tracknumber",
"MDATA_KEY_BPM": "bpm",
"MDATA_KEY_LABEL": "organization",
"MDATA_KEY_COMPOSER": "composer",
"MDATA_KEY_ENCODER": "encodedby",
"MDATA_KEY_CONDUCTOR": "conductor",
"MDATA_KEY_YEAR": "date",
"MDATA_KEY_URL": "website",
"MDATA_KEY_ISRC": "isrc",
"MDATA_KEY_COPYRIGHT": "copyright",
}
# Some airtime attributes are special because they must use the mutagen object
# itself to calculate the value that they need. The lambda associated with each
# key should attempt to extract the corresponding value from the mutagen object
# itself pass as 'm'. In the case when nothing can be extracted the lambda
# should return some default value to be assigned anyway or None so that the
# airtime metadata object will skip the attribute outright.
airtime_special = {
"MDATA_KEY_DURATION" : lambda m: getattr(m.info, "length", 0.0),
"MDATA_KEY_BITRATE" : lambda m: getattr(m.info, "bitrate", 0),
"MDATA_KEY_SAMPLERATE" : lambda m: format_length(getattr(m.info, "sample_rate", 0)),
"MDATA_KEY_MIME" : lambda m: m.mime[0] if len(m.mime) > 0 else u'',
}
mutagen2airtime = dict( (v,k) for k,v in airtime2mutagen.iteritems() if isinstance(v, str) )
truncate_table = {
'MDATA_KEY_GENRE' : 64,
'MDATA_KEY_TITLE' : 512,
'MDATA_KEY_CREATOR' : 512,
'MDATA_KEY_SOURCE' : 512,
'MDATA_KEY_MOOD' : 64,
'MDATA_KEY_LABEL' : 512,
'MDATA_KEY_COMPOSER' : 512,
'MDATA_KEY_ENCODER' : 255,
'MDATA_KEY_CONDUCTOR' : 512,
'MDATA_KEY_YEAR' : 16,
'MDATA_KEY_URL' : 512,
'MDATA_KEY_ISRC' : 512,
'MDATA_KEY_COPYRIGHT' : 512,
}
def format_length(mutagen_length):
"""Convert mutagen length to airtime length"""
t = float(mutagen_length)
h = int(math.floor(t / 3600))
t = t % 3600
m = int(math.floor(t / 60))
s = t % 60
# will be ss.uuu
s = str(s)
seconds = s.split(".")
s = seconds[0]
# have a maximum of 6 subseconds.
if len(seconds[1]) >= 6: ss = seconds[1][0:6]
else: ss = seconds[1][0:]
return "%s:%s:%s.%s" % (h, m, s, ss)
def truncate_to_length(item, length):
if isinstance(item, int): item = str(item)
if isinstance(item, basestring):
if len(item) > length: return item[0:length]
else: return item
class Metadata(Loggable):
def __init__(self, fpath):
try: full_mutagen = mutagen.File(self.path, easy=True)
except Exception: raise BadSongFile(self.path)
# TODO : Simplify the way all of these rules are handled right not it's
# extremely unclear and needs to be refactored.
metadata = {}
# Load only the metadata avilable in mutagen into metdata
for k,v in full_mutagen:
# Special handling of attributes here
if isinstance(v, list):
if len(v) == 1: metadata[k] = v[0]
else: raise Exception("Unknown mutagen %s:%s" % (k,str(v)))
else: metadata[k] = v
self.__metadata = {}
# Start populating a dictionary of airtime metadata in __metadata
for muta_k, muta_v in metadata.iteritems():
# We must check if we can actually translate the mutagen key into
# an airtime key before doing the conversion
if muta_k in mutagen2airtime:
airtime_key = mutagen2airtime[muta_k]
# Apply truncation in the case where airtime_key is in our
# truncation table
muta_v = truncate_to_length(muta_v, truncate_table[airtime_key]) \
if airtime_key in truncate_table else muta_v
self.__metadata[ airtime_key ] = muta_v
# Now we extra the special values that are calculated from the mutagen
# object itself:
for special_key,f in airtime_special:
new_val = f(full_mutagen)
if new_val is not None:
self.__metadata[special_key] = f(full_mutagen)
# Finally, we "normalize" all the metadata here:
self.__metadata = mmp.normalized_metadata(self.__metadata, fpath)
# Now we must load the md5:
self.__metadata['MDATA_KEY_MD5'] = mmp.fild_md5(fpath)
def extract(self):
return copy.deepcopy(self.__metadata)

View File

@ -14,8 +14,7 @@ class Organizer(ReportHandler,Loggable):
"""Intercept events where a new file has been added to the organize
directory and place it in the correct path (starting with self.target_path)"""
try:
normal_md = mmp.normalized_metadata(event.metadata, event.path)
new_path = mmp.organized_path(event.path, self.target_path, normal_md)
new_path = mmp.organized_path(event.path, self.target_path, event.metadata.extract())
mmp.magic_move(event.path, new_path)
self.logger.info('Organized: "%s" into "%s"' % (event.path, new_path))
except BadSongFile as e:

View File

@ -193,7 +193,10 @@ def organized_path(old_path, root_path, normal_md):
return filepath
def file_md5(path,max_length=100):
"""Get md5 of file path (if it exists)"""
"""
Get md5 of file path (if it exists). Use only max_length characters to save time and
memory
"""
if os.path.exists(path):
with open(path, 'rb') as f:
m = hashlib.md5()

View File

@ -1,5 +1,6 @@
# -*- coding: utf-8 -*-
import unittest
import os
import media.monitor.pure as mmp
class TestMMP(unittest.TestCase):
@ -51,4 +52,12 @@ class TestMMP(unittest.TestCase):
# for recorded it should be something like this
# ./recorded/2012/07/2012-07-09-17-55-00-Untitled Show-256kbps.ogg
def test_file_md5(self):
p = os.path.realpath(__file__)
m1 = mmp.file_md5(p)
m2 = mmp.file_md5(p,10)
self.assertTrue( m1 != m2 )
self.assertRaises( ValueError, lambda : mmp.file_md5('/bull/shit/path') )
self.assertTrue( m1 == mmp.file_md5(p) )
if __name__ == '__main__': unittest.main()