[tdwg-tapir] conceptual binding

[Renato De Giovanni]

Hi Markus,

This is really good news for the PyWrapper!

Now concerning your question, I think we should really try to avoid 
defining concepts with compound XML schemas. It's a lot cleaner and 
more elegant using gcp#accession/FullScientificName to map local 
databases and to use in filters. As you said, the WFS response would 
just be one of the possible output models making use of the concepts.

The real xpaths to WFS elements (including namespaces) would only be 
used in the output model mapping section: 

<node 
path="/wfs:FeatureCollection/gml:featureMember/gcp:accession/gcp:FullS
cientificName">
<concept id="gcp#accession/FullScientificName"/>
</node>

Best Regards,
--
Renato

On 28 Feb 2006 at 11:38, "Döring, Markus" wrote:

> Hello,
> I recently added support for schema imports in pywrapper mainly to support multiple namespaces in the repsonse.
> Me (and Javier) wanted to allow something like this:
> 
> <wfs:FeatureCollection 
>     xmlns:wfs="http://www.opengis.net/wfs"
>     xmlns:gml="http://www.opengis.net/gml"
>     xmlns:gcp="http://www.ipgri.org/schemas/gcp_passport_gml/1.0">
>     <gml:featureMember>
>         <gcp:accession fid="accession.12">
>             <gcp:GermplasmID>12</gcp:GermplasmID>
>             <gcp:FAOInstituteCode>ES</gcp:FAOInstituteCode>
>             <gcp:FullScientificName>Quercus ilex</gcp:FullScientificName>
>         </gcp:accession>
>     </gml:featureMember>
> </wfs:FeatureCollection>
> 
> Here comes my problem: 
> How do we refer to the FullScientificName concept in TAPIR?
> 
> As a convention for schemas we said to create the fully qualified concept from the namespace and the simple xpath to the element. So somethinkg like this:
> 
> http://www.opengis.net/wfs#/FeatureCollection/featureMember/accession/FullScientificName
> 
> A real xpath to that element would of course include several namespaces:
> /wfs:FeatureCollection/gml:featureMember/gcp:accession/gcp:FullScientificName
> 
> Could it be that we never need to refer to "compound" concepts anyway? A provider would probably map his db to the GCP schema alone. The WFS response is only the responded view/data model and we would still just use gcp#accession/FullScientificName as the qualified concept in filters. This is also true for our simple extension schemas I think (EFG for ABCD, all extensions for DwC). But arent there other cases?
> 
> Thanks,
> Markus