[ckan-dev] Metadata representation within CKAN

[Rufus Pollock]

On 22 September 2011 15:17, Antoine Logean <antoine.logean at innoq.com> wrote:
> Hi,

Welcome!

> First of all I would like to thanks the community of developer who has
> written CKAN. By following the Amazone EC2 tutorial I had in 10 minutes
> a running directory ! Nice job.
>
> Now I have a question related to the way metadata are
> represented/managed/accessed within CKAN. With metadata I mean the
> description of the datasets (semantic of the data (culture, population,
> etc...), format, URL, size, copyright, ...). Where are this metadata
> definition located ? Are they represented by a clear interface/schema in

So the metadata itself is obviously stored with the CKAN database. It
can be accessed and edited via the API and via the web user interface
(or directly from code if you are working on the same machine).

Regarding metadata definitions see:

<http://wiki.ckan.org/Domain_Model>

In particular for the Dataset (Package) schema see:

<http://wiki.ckan.org/Domain_Model/Package>

(As stated there you can see this reified in code here:
https://bitbucket.org/okfn/ckan/src/tip/ckan/logic/schema.py)

> a central place that I can export and use within an other application ?
> How are these metadata formally described within CKAN ? Is it possible
> to set up a sparql endpoint on CKAN so that I can navigate in the
> metadata defined within CKAN and export them in RDF ?

It is easy to export CKAN metadata to RDF (we already have an
converter library) and setting up a SPARQL endpoint is easy. For
example all of thedatahub.org's data (and more) can be SPARQL queried
in:

<http://semantic.ckan.net/>

Regards,

Rufus