Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo
Is there currently any library support available to MediaTomb that allows it to extract the Tag or Keyword details from a Jpeg file on Import so I can create a Keyword structure under Photos?
I've looked at libexif and libextractor, but I've not been successful as yet in getting them to find and extract the right data from the jpeg files.
If I try the 'exif', or 'extract' programs on the photos separately, they're not showing the extended metadata which is where I think the keywords are stored (using F-Spot, thats where they're listed, under the Subjects and Keywords label), which means I'm just guessing as to what values to specify in config.xml for libextractor's add-data element (although 'extract -L' does list some potentials). It would also suggest that this information will not be available to MediaTomb using either of those two libraries anyway.
Also is there any logging that can be used or enabled thats accessible from the import.js file? It would be handy to be able to dump out the contents of the obj.meta or obj.aux to a log file so I could see what values were available.
libexif and libextractor are the only libs I can suggest, I do not know much about jpeg keywords, so far I have been only working with EXIF data - probably because I usually work with photos from my digicam and all digicams add some sort of exif info to the images.
I do not know if you can really dump obj.meta or obj.aux, so you are right - that indeed would be handy; obj.meta is clearly defined, so you could look up the tags, for obj.aux you have to know what you put into the object (that's the stuff you specify in config.xml in the aux section)
the AUX tags themselves - i.e. the ones that you can use are defined by the appropriate libraries.
but... if exif/extract do not show anything on your photos - then the fields will be empty; those command line utilities use the same libraries as we do.
the only thing I can suggest - find a program that would read your jpeg keywords and add EXIF tags. of course, if you know a library for reading jpeg keywords - tell me, I might have a look and maybe add support for that.
below is a list of auxtags that we support for libexif (this is information is also available through libexif documentation)
and here is the list for libextractor:
the ones below only available in libextractor > 0.5.2
From what I can see, the keywords I've got in my photos are in the newer Extensible Metadata Platform (XMP) format ("Adobe's XML-based specification for metadata geared toward digital images and multimedia").
I had used Vistas Photo Gallery to Tag all my photos, so they're all categorized, which is really useful apart from the fact that nothing, including the Windows Media Players Network Sharing, can actually see the tags/keywords which is no use at all.
I ideally want to replace XP on my Media Server box with Linux but retain the functionality I had before, which is the ability to view my photos through my PS3.
After discovering MediaTomb this morning, it all seemed possible. I've already tweaked the import.js to allow me to browse by Year/Month/Day so I've got the same functionality as I had through XP. I just need to get the browse by Tags working next to make it even better.
If you are interested in trying to incorporate the extra functionality, Adobe do have a downloadable XML SDK toolkit under the BSD license and the only library/program I've come across today that uses it to access XMP metadata is exempi (http://libopenraw.freedesktop.org/wiki/Exempi).
I managed to get it built and installed (I think it just installs the library libexempi), and using the dumpmainxmp test program, I could see my tags.
looks interesting, there's even a package for Fedora so I guess the library is stable/good enough since it made it into the distro. Could you please send me one photo that is tagged accordingly, I need something for testing :> please send it to: jin at mediatomb dot cc
also, is it possible that an image has both, XML and EXIF data? I guess I should figure out what to do in that case, or who gets more priority since double extraction would not really make sense.
I'll look at the API of the library, our interface to adding new metadata extractors is very simple, so if the library is easy to use I could quickly add support for it.
A sample should be on its way to you. It'll have EXIF data and the XMP data.
My camera puts the EXIF data in and the Vista Photo Gallery app puts the XMP data in (although there must be other apps, like Adobe Photoshop that do this too), although I don't know if it also duplicates the already present EXIF data when it saves the XMP metadata.
Hello, Sorry for opening that topic after such a long time, but how does it go with XMP support in MediaTomb?