Re: [Vala] using tracker for other tasks

[Raul Gutierrez Segales <raul gutierrez segales collabora co uk>]

Hi, 

On Sun, 2011-03-06 at 19:20 +1300, banjo wrote:
> hi all
>
> Ive been working on a vala app that requires complex data,
> and found RDF was close to what i need.
>
> I was pondering Redland and Rasqal but they seem overkill
> for what i want.
>
> This app is purely local and doesn't need externally
> defined ontologies or any of the Semantic Web overhead.
>
> So i started writing an RDF-ish store over sqlite
> and was pondering whether i needed a SPARQL-ish queryer.
>
> Then i found tracker which also uses vala
>    http://projects.gnome.org/tracker/
> and has implemented a store and SPARQL queryer.
>
> In trackers' README i see...
>
>    * Comprehensive one stop solution for all applications needing
>    an object database, powerful search (via RDF Query), first class
>    methods, related metadata and user-definable metadata/tags.
>
> This is what i need.
> But i can't find instructions on how to use it.
> (though ive only been looking for a day).
>
> Also i don't want to store my apps data in the same store
> as trackers desktop meta-data.
>
> So my question is...
> Can i cleanly connect to the tracker code in vala
> and just use it's store & query?
> Can i store that data elsewhere from wherever tracker stores it's data?

I've been using tracker-builder (to build queries) and tracker-sparql
(to send those queries through D-Bus) to add a Tracker back-end for
libfolks [1]. 

I suspect that tracker-sparql is highly coupled with Tracker's specfic
DB but then again maybe it it isn't that hard (and might be quite
useful) to decouple the RDF to SQLite layer. 

Cheers,
Raúl 

[1] http://telepathy.freedesktop.org/wiki/Folks