On Tue, Jul 26, 2011 at 4:26 PM, Florent André <[email protected]> wrote:
> Bigs thanks Rupert for this detailed explanation and your time.
>
> It really clear now.
> In fact I was confuse by the "local" term and imported (cached in fact)
> referenced sites (I think this lasts one are the "locals").
>
> Investigate documentation with this insight this bring me to some
> thoughts/questions. Stop me if I'm wrong :
>
> - entityhub/lookup allow to search entities in local and in referenced site.
> It's the unique entry point when we want request "home made" symbols and
> referenced entities.
>
Currently there is no way to search the local and all referenced
sites. Basically lookup does exactly that, but it would also import an
entity found on an referenced site therefore it is not really a way to
search only.
However if you would need this feature just write a feature request.
Maybe /lookup should only search and one should add an /import?

> - entityhub/mapping/{entity|symbol} : make the link between local and
> referenced things [ ha ! a symbol here ! :) ]
>

Exactly. Mappings are created when you import an entity. In future it
is also planed to automatically detect them based on owl:sameAs links
to other entities present on referenced sites but that is currently
not implemented and I do not have it on my immediately TODO list.
The "symbol" is still there because I was not able to find a good
name. Maybe It would be possible to remove both mapping/entity and
mapping/symbol and just go with /mapping.

> - entityhub/entity perform CRUD operation for adding entities to the local
> repository.
>

Yes this is the Idea. I will also write integration tests for that in
the coming days

> About the creation of local entities, documentation just speak about the id
> parameter.

Basically you parse an RDF graph in the POST (create) and PUT (update)
requests to /entity. Such an RDF graph can contain triples about many
resources. Think about parsing the RDF data of an whole thesaurus.
If no id parameter is parsed, than all the resources used as subjects
within triples in the parsed RDF graph will be created as Entities. In
case of the thesaurus all concepts (skos:Concept) and the resource for
the thesaurus itself (skos:ConceptScheme) would be created as
entities.
If you parse the URI of a single concept as id, than only this Entity
(all triples with this URI as subject) would be imported and all the
triples would get ignored.

> There is a way to add any kind of relations to this entities (like
> skos:definition) ? It's planned ?

Updating an entity works like this:

* GET the entity
* add the data you want
* PUT the entity with the additional data.

best
Rupert Westenthaler

-- 
| Rupert Westenthaler             [email protected]
| Bodenlehenstraße 11                             ++43-699-11108907
| A-5500 Bischofshofen

Reply via email to