Hi Tobias,

On 19 October 2010 17:18, Tobias Girschick <tobias.girschick at in.tum.de>wrote:

> Hi Nina,
>
> I am not sure I get the whole thing.


Currently, given an ontology service, one can query for algorithms, models,
features, etc., but not vice versa - there is no automatic way to find out
which ontology service may contain entries of e.g. TUMserver/algorithm


> Just to clarify. If I issue a GET
> on, e.g., TUMserver/algorithm/CDK/sparqlendpoint the result would be the
> URI of all ontology servers the algorithm is registered with?
>

Yes, list of ontology servers , the algorithm is registered with.   I am not
sure whether it is feasible to have the list always consistent and perhaps
it is better to assume it might be incomplete (e.g. the service is
registered somewhere, but the entry is not updated).


> Point 3) is done "automatically" on registering with an ontology
> service?
>
>
I can think of several options:

1)  As you said, update is done automatically on registration (i.e. the
ontology service issues POST/PUT/DELETE in relevant options)

2) The list is initialized internally by the service itself, with an address
of a preferred ontology service.  If a service registers itself
automatically (as discussed yesterday during the virtual meeting), then it
will update the entries accordingly.

3) The entries are updated by a client, which decides to register particular
service (e.g. a validated model)

I don't really have a preference to any of the options, the idea was to
offer a convenient way to find the ontology server, offering sparql endpoint
, given URL of any OpenTox REST service.

Best regards,
Nina



> best regards,
> Tobias
>
> On Tue, 2010-10-19 at 16:21 +0300, Nina Jeliazkova wrote:
> > All,
> >
> > As the answer of the question in the subject of this email doesn't have
> an
> > easy answer via the current API, here is a suggestion (originally by
> Egon)
> > how to extend the API :
> >
> > 1) Each service implements  /{service}/sparqlendpoint
> >
> > e.g. /compound/sparqlendpoint  , /model/sparqlendpoint , etc.
> >
> > 2) The /{service}/sparqlendpoint  returns on GET zero, one or mode URLs
> of
> > ontology services, using text/uri-list   mime type;
> >
> > 3) Adding, updating and deleting entries is done via POST , PUT and
> DELETE
> > to /{service}/sparqlendpoint
> > (whether it will use www-form or text/uri-list as content-type is to be
> > specified )
> >
> > Similar approach might be adopted for returning URLs of AA services ,
> e.g.
> > /{service}/auth.
> >
> > Any feedback is welcome.
> >
> > Best regards,
> > Nina
> > _______________________________________________
> > Development mailing list
> > Development at opentox.org
> > http://www.opentox.org/mailman/listinfo/development
>
> --
> Dipl.-Bioinf. Tobias Girschick
>
> Technische Universität München
> Institut für Informatik
> Lehrstuhl I12 - Bioinformatik
> Bolzmannstr. 3
> 85748 Garching b. München, Germany
>
> Room: MI 01.09.042
> Phone: +49 (89) 289-18002
> Email: tobias.girschick at in.tum.de
> Web: http://wwwkramer.in.tum.de/girschick
>
> _______________________________________________
> Development mailing list
> Development at opentox.org
> http://www.opentox.org/mailman/listinfo/development
>