[KPhotoAlbum] Likelihood of duplicate images?

Tim McCormack kpa at brainonfire.net
Sat Nov 12 22:38:26 CET 2011


On Sat, 1 Oct 2011 07:51:38 +0300, Miika Turkia wrote:
> On Sat, Oct 1, 2011 at 5:21 AM, Tim McCormack <kpa at brainonfire.net>
> wrote:
> > * Does KPA have a documented behavior (that I should mimic) for
> >  handling duplicates?
> 
> Nope. If image is moved we detect that by the MD5 sum and update the
> metadata (however, I have no idea how well that works with duplicates)

OK, thanks. I think my algorithm will change to be this:

1. For all images in index.xml that have exact md5/path matches in my
   shadow.db3, mark as processed.
2. For all remaining images in index.xml that have a single partial
   match (just path or just md5) in shadow.db3, accept and mark as moved
   or edited. If multiple partial matches, either exit program or ask
   user to disambiguate.
3. For all remaining images in index.xml that have no match, accept as
   new.
4. Delete all images in shadow.db3 that did not have any match in
   index.xml.

Previously I was able to do a single sweep through index.xml --
allowing duplicates requires a preliminary sweep.

 - Tim McCormack


More information about the KPhotoAlbum mailing list