#2: Media Metadata

Raw tags

At the most basic layer, it simply gets the metadata encoded within a file, hiding the details of this from the application programmer. In your code you get the tool, call 'mdtool.getMetadataProxies(file)' (where file is the blob you'd like info on). 'file' might be ogg/mp3/wma/avi/divx/...

A metadata proxy returns a metadata object, which subclasses 'dict'. It also provides other methods to get e.g. the type of the metadata returned.

Parsed tags

Wrappers or proxies at a higher level may augment this metadata object by mapping the literal tags in the media files to a single set of names preferred by PloneMultimedia, e.g. derived from MusicBrainz metadata. (See MusicBrainz Metadata Initiative 2.1, and MusicBrainz Metadata Vocabulary). mmpython has its own mappings of tags to names (see mmpython.mediainfo).

Adding application data

An MP3 file contains metadata about the file in isolation. However, the file might form part of a radio show, in which case additional properties become relevant, e.g. date of broadcast and name of presenter. In this case, the application would register a proxy which returns this data.

Metadata objects

A metadata object subclasses dict. In addition, it knows what type of metadata it represents:

>>> print metadata.getMetadataType()
'id3'
>>> print metadata
{'WXXX': u'(User defined URL link): (): ',
'TCOP': u'(Copyright message): ',
'TOPE': u'(Original artist(s)/performer(s)): ',
...

MetadataTool

The tool provides getMetadataProxies, which takes either a Plone content object or a filename and mime type specification, and looks at the registered proxies, returning all the matching ones. Multiple proxies might match, e.g. id3_proxy and itunes_proxy. It's up to the application to choose among them.

MetadataProxy

A metadata proxy is registered for a mime type. It knows its type, and you can iterate and subscript it like a dictionary. If there is a marshaller registered for the mime type, it provides the default string representation. If multiple marshallers are registered, the application can choose among them.

Marshallers

In addition to extraction adapters, the metadata tool allows registration of marshalling adapters. Marshalling adapters are registered for metadata types. E.g. a MusicBrainzRDFMarshaller might be registered for id3 tags, for ogg tags, or for a higher-level metadata type such as the tags used by mmpython.

Here we are with a proxy which has MusicBrainzRDFMarshaller as primary marshaller:

>>> metadata
<?xml version="1.0" encoding="UTF-8"?>
<rdf:RDF xmlns:rdf=" http://www.w3.org/1999/02/22-rdf-syntax-ns#"
        xmlns:dc=" http://purl.org/dc/elements/1.1/"
        xmlns:mq=" http://musicbrainz.org/mm/mq-1.1#"
        xmlns:mm=" http://musicbrainz.org/mm/mm-2.1#"
        xmlns:ar=" http://musicbrainz.org/ar/ar-1.0#"
        xmlns:az=" http://www.amazon.com/gp/aws/landing.html#">
  <mq:Result>
    <mq:status>OK</mq:status>
    <mm:albumList>
      <rdf:Bag>
        <rdf:li rdf:resource="http://musicbrainz.org/mm-2.1/album/1073abfc-768e-455b-9937-9b41b923c746"/>
      </rdf:Bag>
    </mm:albumList>
  </mq:Result>
  <mm:Album rdf:about="http://musicbrainz.org/mm-2.1/album/1073abfc-768e-455b-9937-9b41b923c746">
    <dc:title>Beaucoup Fish</dc:title>
    ...
    <az:Asin>B00000IFTF</az:Asin>
  </mm:Album>
</rdf:RDF>

Let's see what else we have:

>>> metadata.getMarshallers()
[MusicBrainzRDF, RFC822, BitziRDF]

>>> metadata.asRFC822()
Title: Girl
Creator: Beck
AlbumName: Guero
Date: 2005
FileName: 03 - Beck - Girl - www.torrentazos.com.mp3
...

Bundled proxies and marshallers

A set of standard proxies and marshallers can be bundled as a separate product, similar to ATExtensions and MoreFieldsAndWidgets that bundle fields and widgets. Specific products (e.g. PloneRadioShowVapourware) provide their own specialised proxies and marshallers.

Sample session

Here's a sample session at the zopectl debug Python interpreter:

>>> # A Zope object
>>> song = container['somefile.mp3']
>>> song.portal_type
'ATAudio'

>>> photo = container['somephoto.jpg']
>>> # e.g. CMFPhoto or ATImage

>>> # See if the metadata_tool has adapters for us
>>> song_metadata_proxies = metadata_tool.getMetadataProxies(song)
>>> photo_metadata_proxies = metadata_tool.getMetadataProxies(photo)
>>> # E.g. one for EPC, one for IPTC data
>>> len(photo_metadata_proxies)
2

>>> a = photo_metadata_proxies[0]

>>> # Gather metadata from all the metadata_proxies
>>> metadata = [a.getMetadata() for a in photo_metadata_proxies]

>>> metadata[0][u'title']
u'asdfdsaf sadfsadf' # Return unicode
>>> # Does this make sense? Shouldn't we just return the bytes that were
>>> # included in the file?

>>> metadata[0].getMetadatatype()
'exif'

>>> exif_metadata = metadata[0]

>>> metadata[1].getMetadataType()
'iptc'

>>> iptc_metadata = metadata[1]

>>> exif_metadata.keys() # dict interface
[u'Aperture', u'Make', u'Model', ....]

>>> exif_metadata['Aperture']
u'8' # We can't assume the type of the value, so make it string

>>> itunes_song = container['indistinguishablefromaplaintrack.mp3']

>>> song_metadata_proxies = metadata_tool.getMetadataProxies(itunes_song)

>>> metadata_list = [a.getMetadata() for a in song_metadata_proxies]

>>> for metadata in metadata_list:
>>>     print '---'
>>>     print metadata.getMetadataType()
>>>     print metadata
---
'id3'
{'WXXX': u'(): ',
'TCOP': u'',
'TOPE': u'',
'TCOM': u'',
'TRCK': u'4',
'TIT2': u'Hooplas Involving Circus Tricks',
'TENC': u'',
'COMM': u'(iTunNORM)[eng]:  0000102D 00000707 000069F9 00003FAF 000119F0 000119D6 00008A67 000089ED 00011CCB 0000D98D',
'TPE1': u'Say Hi To Your Mom',
'TALB': u'Numbers & Mumbles',
'TYER': u'2004',
'TCON': u'(131)Indie',
# Repeated tag as list
'COMM': [u'()[eng]: http://www.sayhitoyourmom.com',
    u'(ID3v1 Comment)[XXX]: http://www.sayhitoyourmom.co',
    u'(iTunes_CDDB_1)[eng]: 0200D201+213+1+150',
    u'(iTunes_CDDB_TrackNumber)[eng]: 1',
    ]
    }
---
'itunes'
{'WXXX': u'(): ',
'TCOP': u'',
'TOPE': u'',
'TCOM': u'',
'TRCK': u'4',
'TIT2': u'Hooplas Involving Circus Tricks',
'TENC': u'',
# Include original tag
'COMM': u'(iTunNORM)[eng]:  0000102D 00000707 000069F9 00003FAF 000119F0 000119D6 00008A67 000089ED 00011CCB 0000D98D',
# Parse it into an additional tag. Should this be distinguished from the
# id3 tags, or is that something the application should take care of?
'(iTunNORM)[eng]': u'0000102D 00000707 000069F9 00003FAF 000119F0 000119D6 00008A67 000089ED 00011CCB 0000D98D',
'TPE1': u'Say Hi To Your Mom',
'TALB': u'Numbers & Mumbles',
'TYER': u'2004',
'TCON': u'(131)Indie',
'COMM': [u'()[eng]: http://www.sayhitoyourmom.com',
    u'(ID3v1 Comment)[XXX]: http://www.sayhitoyourmom.co',
    u'(iTunes_CDDB_1)[eng]: 0200D201+213+1+150',
    u'(iTunes_CDDB_TrackNumber)[eng]: 1', ],
'(iTunes_CDDB_1)[eng]': u'0200D201+213+1+150',
'(iTunes_CDDB_TrackNumber)[eng]': '1', }
}

References