Re: Combining @id with @vocab / combining @graph with @included / JSON API compatibility

Hi Thomas,

On 17/01/2023 18:04, Thomas Francart wrote:
> Hello
>
> Considering the JSON-LD below, which combines @graph, @included and 
> @vocab, I have 3 questions :
>
> 1. No rdf:type triple is produced with https://exemple.eu/entity_1 as 
> subject. I intended the "id" attribute to be found in the context, as 
> I marked the "id" key with "@type = @vocab". This works correctly with 
> the type key.
> I understand this is because JSON-LD mandates @id to be IRIs or blank 
> node identifiers. But why wouldn't it be possible to find the 
> corresponding IRI in the @context ?
> I would like to do that so that the key is the same when I refer to 
> "entity_1" in the "refers_to" key and when I describe "entity_1" in 
> the "included" section.

I can see why you would want that, but as a general rule, you can not 
override how keywords (@id, @type, @language...) interpret their values. 
Which means that attempting to do that should maybe raise a warning... 
(your "@type": "@vocab" in the definition of "id").

I believe that this would open a dangerous pandora box (imagine 
replacing "@type": "@vocab" with "@type": "xsd:string" -- what would 
that mean?).

>
> 2. When combining @graph with @included, the triples from the @graph 
> are put in an anonymous graph instead of the default graph. Triples 
> from the @included section are put in the default graph (if I 
> workaround the first issue by using an explicit IRI). I find that 
> counter intuitive. The spec reads "/When a JSON-LD document's 
> top-level structure is a map that contains no other keys than @graph 
> and optionally @context (...), @graph is considered to express the 
> otherwise implicit default graph/", but I suggest to change it to 
> "/...contains no other keys than |@graph and @included...|/"
>
> Maybe there is another way to have the data+included structure at the 
> top-most level while maintaining a single default graph ?
easy: alias "graph" to "@included" instead of "@graph" :)
>
> 3. This is done in an attempt to mimic the JSON API (jsonapi.org 
> <http://jsonapi.org>) structure while keeping JSON-LD compatibility. 
> Do you know of other attempts to do so, or is there some kind of 
> default "JSONAPI to JSON-LD" @context or something similar ?

I have a vague recollection of somebody mentioning that while the WG was 
working on the 1.1 specifications, but don't remember who that was. 
Maybe Benjamin?

   pa

>
> Thanks !
> Thomas
>
>
> ```
> {
>   "@context": {
>
>     "id": {
>       "@id": "@id",
>       "@type":"@vocab"
>     },
>
>     "type": {
>       "@id": "rdf:type",
>       "@type": "@vocab"
>     },
>
>     "refers_to": {
>       "@id": "https://exemple.eu/ontology#refers_to",
>       "@type": "@vocab"
>     },
>
>     "data": "@graph",
>     "included": "@included",
>
>     "entity_1": "https://exemple.eu/entity_1",
>
>     "Test" : "https://exemple.eu/ontology#Test"
>   },
>
>
>   "data": [
>     {
>       "id": "https://exemple.eu",
>       "type": "Test",
>       "refers_to": "entity_1"
>     }
>   ],
>
>   "included" : [
>     {
>       "id": "entity_1",
>       "type": "Test"
>     }
>   ]
> }
> ```
>
> -- 
> *
> *
> *Thomas Francart* -*SPARNA*
> Web de _données_ | Architecture de l'_information_ | Accès aux 
> _connaissances_
> blog : blog.sparna.fr <http://blog.sparna.fr>, site : sparna.fr 
> <http://sparna.fr>, linkedin : fr.linkedin.com/in/thomasfrancart 
> <https://fr.linkedin.com/in/thomasfrancart>
> tel :  +33 (0)6.71.11.25.97, skype : francartthomas