Re: Driving JSON-LD framing specs from SHACL model

This is very helpful thank you. I am particularly impressed by the work of
the WoT Working Group [1], very similar to my problematics, both in terms
of coupling RDF modelling with JSON schemas, and in terms of model
documentation. So far we are providing SHACL-generated documentation, but
it could make sense to provide this JSON-based documentation as well, or as
a complement.

[1] :
https://docs.google.com/presentation/d/193OFcFaxD0GqrRuOggwZe5eorgL1C1Epe2cAYN3JEkk/edit#slide=id.g2c05cb296b1_0_113


Le mar. 11 juin 2024 à 17:52, Pierre-Antoine Champin <pierre-antoine@w3.org>
a écrit :

> Hi Thomas,
>
> we had two breakout sessions, in September 2023 [1] (during TPAC) and in
> March 2024 [2] (during Breakout day) around this topic. You might find the
> minutes and slides presented there interesting. We might continue the
> discussion during the upcoming TPAC or further events (this is still in
> discussion).
>
> More comments belo
> [1]
> https://www.w3.org/events/meetings/b8617206-4c2b-4355-80da-fcf4eb628d1a/
> [2]
> https://www.w3.org/2024/03/breakouts-day-2024/#b-fb7b6d9f-fc2f-456f-9fff-3a516f3ec7ab
>
>
> On 07/06/2024 12:31, Thomas Francart wrote:
>
> Hello SHACL & JSON-LD lovers
>
> I have SHACL specs. I also have APIs that return JSON-LD. The JSON-LD is
> built from RDF using a manually maintained context + JSON-LD framing specs.
>
> My question is how I could bridge my SHACL with my JSON-LD context +
> framing, to have a completely model-driven architecture. In other words can
> I generate the context + framing from the SHACL (probably using some extra
> annotations).
>
> Generating a default context from the SHACL is relatively easy [1]. My
> question is on the framing part, specifically on the hierarchical structure
> of the JSON. How can I indicate in the SHACL model :
>
>    1. which shapes are the roots of the JSON-LD structure ?
>    2. which properties use embedding in the JSON-LD structure ?
>    ("@embed": "@always") and which should be simply references to URIs ?
>
> Intuitively, I would use a similar approach to the one you used in your
> SHACL-to-context conversion: a dedicated predicate used to annotate the
> node shapes (for 1) and the property shapes (for 2).
>
> The ability to specify the (or a) hierarchical view of my graph structure
> could be useful, even for an RDF/XML serialization, or an HTML rendering of
> a dataset.
>
> +1
>
>
> Any ideas or pointers on this ?
>
> Cheers
> Thomas
>
> [1] : https://shacl-play.sparna.fr/play/context#documentation - a
> simplistic JSON-LD context derivation from a SHACL spec.
>
> --
>
> *Thomas Francart* -* SPARNA*
> Web de *données* | Architecture de l'*information* | Accès aux
> *connaissances*
> blog : blog.sparna.fr, site : sparna.fr, linkedin :
> fr.linkedin.com/in/thomasfrancart
> tel :  +33 (0)6.71.11.25.97, skype : francartthomas
>
>

-- 

*Thomas Francart* -* SPARN**A*
linked *data* | domain *ontologies* | *knowlegde* graphs
blog : blog.sparna.fr, site : sparna.fr, linkedin :
fr.linkedin.com/in/thomasfrancart
tel :  +33 (0)6.71.11.25.97