mm2: added back metadata handling. Added new "Enterprise Metadata

Framework"

python_apps/media-monitor2/media/metadata/__init__.py

@@ -0,0 +1 @@
+

python_apps/media-monitor2/media/metadata/definitions.py

@@ -0,0 +1,103 @@
+# -*- coding: utf-8 -*-
+import media.monitor.process as md
+from os.path import normpath
+from media.monitor.pure import format_length, file_md5
+
+with md.metadata('MDATA_KEY_DURATION') as t:
+    t.default(u'0.0')
+    t.depends('length')
+    t.translate(lambda k: format_length(k['length']))
+
+with md.metadata('MDATA_KEY_MIME') as t:
+    t.default(u'')
+    t.depends('mime')
+    t.translate(lambda k: k['mime'].replace('-','/'))
+
+with md.metadata('MDATA_KEY_BITRATE') as t:
+    t.default(u'')
+    t.depends('bitrate')
+    t.translate(lambda k: k['bitrate'])
+
+with md.metadata('MDATA_KEY_SAMPLERATE') as t:
+    t.default(u'0')
+    t.depends('sample_rate')
+    t.translate(lambda k: k['sample_rate'])
+
+with md.metadata('MDATA_KEY_FTYPE'):
+    t.depends('ftype') # i don't think this field even exists
+    t.default(u'audioclip')
+    t.translate(lambda k: k['ftype']) # but just in case
+
+with md.metadata("MDATA_KEY_CREATOR") as t:
+    t.depends("artist")
+    # A little kludge to make sure that we have some value for when we parse
+    # MDATA_KEY_TITLE
+    t.default(u"")
+    t.max_length(512)
+
+with md.metadata("MDATA_KEY_SOURCE") as t:
+    t.depends("album")
+    t.max_length(512)
+
+with md.metadata("MDATA_KEY_GENRE") as t:
+    t.depends("genre")
+    t.max_length(64)
+
+with md.metadata("MDATA_KEY_MOOD") as t:
+    t.depends("mood")
+    t.max_length(64)
+
+with md.metadata("MDATA_KEY_TRACKNUMBER") as t:
+    t.depends("tracknumber")
+
+with md.metadata("MDATA_KEY_BPM") as t:
+    t.depends("bpm")
+    t.max_length(8)
+
+with md.metadata("MDATA_KEY_LABEL") as t:
+    t.depends("organization")
+    t.max_length(512)
+
+with md.metadata("MDATA_KEY_COMPOSER") as t:
+    t.depends("composer")
+    t.max_length(512)
+
+with md.metadata("MDATA_KEY_ENCODER") as t:
+    t.depends("encodedby")
+    t.max_length(512)
+
+with md.metadata("MDATA_KEY_CONDUCTOR") as t:
+    t.depends("conductor")
+    t.max_length(512)
+
+with md.metadata("MDATA_KEY_YEAR") as t:
+    t.depends("date")
+    t.max_length(16)
+
+with md.metadata("MDATA_KEY_URL") as t:
+    t.depends("website")
+
+with md.metadata("MDATA_KEY_ISRC") as t:
+    t.depends("isrc")
+    t.max_length(512)
+
+with md.metadata("MDATA_KEY_COPYRIGHT") as t:
+    t.depends("copyright")
+    t.max_length(512)
+
+with md.metadata("MDATA_KEY_FILEPATH") as t:
+    t.depends('path')
+    t.translate(lambda k: normpath(k['path']))
+
+with md.metadata("MDATA_KEY_MD5") as t:
+    t.depends('path')
+    t.optional(False)
+    t.translate(lambda k: file_md5(k['path'], max_length=100))
+
+# MDATA_KEY_TITLE is the annoying special case
+with md.metadata('MDATA_KEY_TITLE') as t:
+    # Need to know MDATA_KEY_CREATOR to know if show was recorded. Value is
+    # defaulted to "" from definitions above
+    t.depends('title','MDATA_KEY_CREATOR')
+    t.max_length(512)
+

python_apps/media-monitor2/media/metadata/process.py

@@ -0,0 +1,140 @@
+# -*- coding: utf-8 -*-
+from contextlib import contextmanager
+from media.monitor.pure import truncate_to_length, toposort
+import mutagen
+
+
+class MetadataAbsent(Exception):
+    def __init__(self, name): self.name = name
+    def __str__(self): return "Could not obtain element '%s'" % self.name
+
+class MetadataElement(object):
+    def __init__(self,name):
+        self.name = name
+        # "Sane" defaults
+        self.__deps          = set()
+        self.__normalizer    = lambda x: x
+        self.__optional      = True
+        self.__default       = None
+        self.__is_normalized = lambda _ : True
+        self.__max_length    = -1
+
+    def max_length(self,l):
+        self.__max_length = l
+
+    def optional(self, setting):
+        self.__optional = setting
+
+    def is_optional(self):
+        return self.__optional
+
+    def depends(self, *deps):
+        self.__deps = set(deps)
+
+    def dependencies(self):
+        return self.__deps
+
+    def translate(self, f):
+        self.__translator = f
+
+    def is_normalized(self, f):
+        self.__is_normalized = f
+
+    def normalize(self, f):
+        self.__normalizer = f
+
+    def default(self,v):
+        self.__default = v
+
+    def get_default(self):
+        if hasattr(self.__default, '__call__'): return self.__default()
+        else: return self.__default
+
+    def has_default(self):
+        return self.__default is not None
+
+    def path(self):
+        return self.__path
+
+    def __slice_deps(self, d):
+        return dict( (k,v) for k,v in d.iteritems() if k in self.__deps)
+
+    def __str__(self):
+        return "%s(%s)" % (self.name, ' '.join(list(self.__deps)))
+
+    def read_value(self, path, original, running={}):
+        # If value is present and normalized then we don't touch it
+        if self.name in original:
+            v = original[self.name]
+            if self.__is_normalized(v): return v
+            else: return self.__normalizer(v)
+
+        # A dictionary slice with all the dependencies and their values
+        dep_slice_orig    = self.__slice_deps(original)
+        dep_slice_running = self.__slice_deps(running)
+        full_deps         = dict( dep_slice_orig.items()
+                                + dep_slice_running.items() )
+
+        # check if any dependencies are absent
+        if len(full_deps) != len(self.__deps) or len(self.__deps) == 0:
+            # If we have a default value then use that. Otherwise throw an
+            # exception
+            if self.has_default(): return self.get_default()
+            else: raise MetadataAbsent(self.name)
+        # We have all dependencies. Now for actual for parsing
+        r = self.__normalizer( self.__translator(full_deps) )
+        if self.__max_length != -1:
+            r = truncate_to_length(r, self.__max_length)
+        return r
+
+def normalize_mutagen(path):
+    """
+    Consumes a path and reads the metadata using mutagen. normalizes some of
+    the metadata that isn't read through the mutagen hash
+    """
+    m = mutagen.File(path, easy=True)
+    md = {}
+    for k,v in m.iteritems():
+        if type(v) is list: md[k] = v[0]
+        else: md[k] = v
+    # populate special metadata values
+    md['length']      = getattr(m.info, u'length', 0.0)
+    md['bitrate']     = getattr(m.info, 'bitrate', u'')
+    md['sample_rate'] = getattr(m.info, 'sample_rate', 0)
+    md['mime']        = m.mime[0] if len(m.mime) > 0 else u''
+    md['path']        = path
+    return md
+
+class MetadataReader(object):
+    def __init__(self):
+        self.clear()
+
+    def register_metadata(self,m):
+        self.__mdata_name_map[m.name] = m
+        d = dict( (name,m.dependencies()) for name,m in
+                self.__mdata_name_map.iteritems() )
+        new_list = list( toposort(d) )
+        self.__metadata = [ self.__mdata_name_map[name] for name in new_list
+                if name in self.__mdata_name_map]
+
+    def clear(self):
+        self.__mdata_name_map = {}
+        self.__metadata       = []
+
+    def read(self, path, muta_hash):
+        normalized_metadata = {}
+        for mdata in self.__metadata:
+            try:
+                normalized_metadata[mdata.name] = mdata.read_value(
+                        path, muta_hash, normalized_metadata)
+            except MetadataAbsent:
+                if not mdata.is_optional(): raise
+        return normalized_metadata
+
+global_reader = MetadataReader()
+
+@contextmanager
+def metadata(name):
+    t = MetadataElement(name)
+    yield t
+    global_reader.register_metadata(t)

python_apps/media-monitor2/media/monitor/pure.py

@@ -272,8 +272,8 @@ def normalized_metadata(md, original_path):
     else:
         # Read title from filename if it does not exist
         default_title = no_extension_basename(original_path)
-        #if re.match(".+-%s-.+$" % unicode_unknown, default_title):
-            #default_title = u''
+        if re.match(".+-%s-.+$" % unicode_unknown, default_title):
+            default_title = u''
         new_md = default_to(dictionary=new_md, keys=['MDATA_KEY_TITLE'],
                             default=default_title)