> scan hdisk for images and create thumbnails
Well, Konqueror can do that, for one. I'm not sure what format the
thumbnails are stored in...
OK, they're stored as JPG images, in .kde/share/thumbnails. I don't
immediately see any rhyme or reason to the series of oddly-named
subdirectories you have to wade through to get to them, but they are indeed
on disk, and in a format viewable with the GIMP. I could not get any of
them to display with Image Magick for some reason.
Quote:> Does anyone know of a program that can scan the hdisk for all image
> files and then produce a browsable interface with thumbnails?
Don't know about that right off. Using what's readily available, one
solution might be to write a script to find all image files, then put
symlinks to them in one common place, and then use Konqueror to thumbnail
them. That would probably have limitations in terms of number though.
Quote:> will it be necessary to copy the images into a database and then
> manually categorize them? if so, has anyone with experience doing
> this anything to recommend?
Manually categorizing them is the only way to go unless they have meaningful
filenames. What sort of classification depends on what you need to do with
them.
I'm currently cataloging about 15,000 images. It's slow going, but I've
picked up the pace thanks to an invention that I conceived and kind folks
here helped me figure out how to realize. It's many times faster than
viewing thumbnails and dragging them around by hand.
I have a series of scripts that work with each other. The sort script takes
a directory and a delay factor as arguments. It runs a for/do loop on the
directory, then passes each filename to the classify script.
The classify script uses Image Magick via the display command to throw each
image up on-screen, then it delays for the assigned delay factor, then
kills the display program, giving the focus back to the console. Then the
classify script scans the keyboard for a keypress and passes that to a case
statement which then moves the image just displayed to a new directory
based on the key pressed. The arrangement works well, though it's not very
flexible, as all of these keypresses and target directories are hard-coded
into the classify script. When I run into something that could fit into
multiple categories, I have to just pick the closest one and move on, or
else stop and hack the classify script and the target directory structure
to add a new category.
I'm finding that almost all of my categories are too broad, and I'm going to
have to go back and refine each of the intermediate holding areas with a
new hard-coded script, or else develop some more flexible way of handilng
that whole aspect of the process. Be that as it may, as least they're
loosely grouped now, instead of being an enormous hodgepodge.
It would be better if the image could remain on-screen until you're ready to
classify it, but I couldn't find a way to give the focus back to the
console. The delay is a kludge, but it's workable.
All in all, it's about the best I could do short of having a program that
handled the keyboard input internally, then returned it to the calling
program as its exit status or some such. Perhaps a hacked version of Image
Magick or the like, but that would be involved.
My solution is something short of elegant, but I only have about 9,000
images left unclassified. It's at least getting me there.
--
Michael McIntyre zone 6b in SW VA
Silvan Pagan
umount /mnt/windows&&mke2fs /dev/hde1
www.geocities.com/Paris/Rue/5407/index.html