The GSoC official period started and i’m back again to the Media Center. Today i’ll show you a one-night-coded engine 
After notmart’s suggestion (great mentor, great) i got a pretty nice working version of a YouTubeEngine!! 
You can easily use it to query YouTube for videos. Try the following queries for example (and thank sebas for the query suggestions!)
asian backstreet boys
numa numa guy
Remember to prepend “query/” to your queries as from the screenshots. Each source will have a key for each video id found. And with each id you’ll be able to retrieve a Plasma::DataEngine::Data containing video informations accessible through the keys specified in the docs (title, comment, keywords). I’ll add new keys asap but it’s already pretty nice and working. I need to do more checking stuff on the queries though.
Aaanyway, enjoy!
Cheers
This is cool, but could you add the support for all RSS/ATOM-based podcasts and then change the dataengine name to something else?
I’m asking because this is a grey area we had with twitter data engine – since twitter is not free software, it was decided that its d.a. will not be shipped by default, but since it started supporting laconica based sites (identi.ca) it is a completely different story.
Sounds great. I have an idea for a plasma widget with that engine:
Get the cinema charts from somewhere and add the trailer, which is probably on youtube, with /query some_film trailer (german|french|spanish)
(of course one could also get the trailers from somewhere else or just embed a webpage which already has the trailers on it)
But whatever way it is done, such a cinema trailers plasmoid would be absolutely great, and I think this dataengine would be a great starting point!
Same point as Ivan really. Even if you don’t do it now, it’s likely that at some point someone will want to expand this to support other sources of video search, it’s probably best to design and name it with that in mind.
@anon … sounds cool … could also check IMDB for trailers
Thanks to Ivan today i discussed with my mentor about the chance to extend it to more services, not only youtube. And we agree that making this engine working with plugins would be really neat. Anyway now i won’t focus on this, but i’ll for sure try a simple plugin implementation asap
small nitpick on the query/ syntax: could you use query| instead? it’s becoming something of a common de-facto approach to such things in plasma dataengines.
suuure mr Aaron, will do!
It would be great if this could totally bypass the flash infested web site. Youtube requires maybe a 1.2 Ghz cpu to play videos in 1:1 size. Running them with a native mplayer instance (wrapped in a kde app) would allow running youtube on older machines, too.