Comment 1 for bug 1046687

This is very good to have, and other people have also requested it, but is quite some work to implement and I don't have much time at the moment. Still, it will definitely land in Variety at some point.

Steps necessary:

1. Hardest one. Implement keywords/tags extraction for every online source:

For Flickr this means additional API calls which may turn out a problem in the future. Flickr limits the number of free API calls allowed per day and at the moment we are very frugal in this aspect - one API call to fill the download queue with tens or hundreds of images. If we need to extract tags, that's another API call for every image in the queue.

For Wallbase and Wallpapers.net this means more scraping code and potential breakage in the future.

MediaRSS (support not yet implemented) does not have a clear convention for listing tags. media:category seems to be the usual place for placing tags, but can we rely on it - needs investigation.

2. Save tags into the images - IPTC Keywords seems like the good place for these, and we may want to duplicate in a Variety-specific tag. Needs more investigation.

3. Implement filtering based on the saved tags. As with any other filtering, this slows down the searching for images and means more CPU/HDD usage from time to time when the prepared queue is exhausted.