Re: [weekly report] Week 1 : Allow Tracker to index remote resources

[Alexandre <airmind gmail com>]

On Thu, Jun 4, 2009 at 12:09, Adrien BUSTANY <madcat mymadcat com> wrote:

Hi there,

I didn't have a lot of time so far to work on the GSOC project due to school exams, but we (with Rob Taylor) nevertheless sorted out a few design issues. Bridges between Tracker and remote sites will be independent processes that fetch metadata and insert it into Tracker. At first, the sync will be dummy (ie not incremental), though it might change in the future.
All the bridge are managed by a separate process which is used to monitor the status of the different bridges and initiate the identification process with the remote sites. The manager uses DBus to "talk" to the bridges.

So far, I have a basic version of the manager working which can list the available bridges and get their status, as well as identify bridges against remote websites. The only bridge so far is for Flickr, it can currently authenticate against the site and pull all the metadata. It does not insert the data into Tracker yet, but I'll do that shortly.

What is planned for next week :
1. Get a fully functionnal sync for Flickr
2. See how we integrate the bridge manager with Flickr

Note : All the details about my project are kept on a page on gnome live : http://live.gnome.org/Tracker/TrackerWebService

Hi,

Somehow I missed your project from the accepted proposals page.
I'm working with Conduit this Summer of Code and one thing we have been discussing on Conduit is how to make  our dataproviders more useful to external apps. One of the ideas we had was to use Tracker for our metadata needs, so it's much faster to obtain information from files and we could use all the crazy stuff Tracker provides (ontologias, the query language, etc)
Well, our dataproviders already implement parts of what you are implementing, and one of our ideas was to get Conduit to index a user's account (such as Flickr) into Tracker. I'm very happy with your project tough, as it means Tracker will have that ability in itself and we could use that information instead of our own dataprovider.
One thing will still be missing is how to push files into a webservice account, which is not currently a goal for your project. But that is something we could work together in the future.

My GSoC project is to make Conduit a service itself. It is already availiable through DBus and I'd be more then happy to create an API that Tracker could use to index files. We already do account configuration and authentication, we already provide means to obtain metadata from items and Conduit would greatly benefit from ontologies.

We have several possibilities where we could work together. What do you think?
 
Regards,
Alexandre

Cheers

Adrien

_______________________________________________
gnome-soc-list mailing list
gnome-soc-list gnome org
http://mail.gnome.org/mailman/listinfo/gnome-soc-list