[Digikam-users] Keyword import
Gilles Caulier
caulier.gilles at kdemail.net
Mon Sep 18 06:29:17 BST 2006
On Sunday 17 September 2006 18:32, Marcel Wiesweg wrote:
> Am Sonntag 17 September 2006 16:16 schrieb Daniel Seifert:
> > Hi,
> >
> > I want to import my iPhoto library into digikam. Most of the meta data
> > is stored in a file called AlbumData.xml and using the iphoto2 gem for
> > Ruby, I can extract this data quite nicely. So now I have a lot of
> > images and the data - how do I get this into digikam?
> >
> > Using exiv2 to put the data into the IPTC/EXIF fields of the images is
> > possible, but when importing the images into digikam not all fields are
> > taken over. Especially the keywords section is not imported
> > automatically - but which I need to have this sorted into albums later.
> >
> > I found
> > http://mail.kde.org/pipermail/digikam-users/2006-June/001642.html which
> > is basically the same problem. Two solutions are presented, one is to
> > access digikam's sqlite database (ugh) and the other to use digikam
> > 0.9.x (I have 0.9.0 beta1).
> >
> > I can't believe that every digikam user who starts using digikam has no
> > previous photos or spends days re-categorizing them, so I would suppose
> > that there is some way out there to properly import existing photos.
> > Direct import from iPhoto would be nice
> > (https://bugs.kde.org/show_bug.cgi?id=132353), but for now anything
> > (script, hidden commands, etc) that allows me to import images with
> > standard IPTC/EXIF fields (including comment, time and keyword) into
> > digikam would be highly appreciated.
>
> Currently only comment, time and rating is imported from IPTC, keyword
> importing is planned but not yet done. This will be the most portable way
> to import tags.
> For now, I dont know if anyone has written a script to import from iPhoto
> to digikam's db.
>
> Marcel
Marcel,
Just to be clear, the code is ready into DMetadata class to extract iptc
keywords from pictures, but an interface with the database is missing to
perform import (scanlib in fact).
The problem is to use the iptc keywords properlly, especially if non-ascii
char are used. I think that the database must be untouched in this case to
prevent dummy Tags import.
Also, a rule must be done when an existing tag already exist in database.
Gilles
More information about the Digikam-users
mailing list