JSON-LD 1.1 conversion of FHIR/JSON to RDF

Harold goes through slides: https://lists.w3.org/Archives/Public/www-archive/2019Dec/att-0028/HCLSDec.pdf

eric: Gregg Kellogg uses framing but then has to finish the job with something else

harold: ShEx is starting to get some serious use. Someone mentioned that it is everything he wanted except for inability to put in a uniqueness constraint. I asked him to file an issue report.
... He said that's the one thing he can do in SHACL but not in ShEx.

(slide 3)

harold: this is the current setup.

(slide 4)

we need to generate ALL of the *.shex file, not only the one big fhir.shex file.

harold: Takes about 20 minutes to convert all shexc to shexj.

(slide 6)

harold: Then we generate the *.context.jsonld
... github has the instructions for doing it.

eric: We might have to submit a PR to make the JSON-LD processor accept file: URLs.

harold: we'll want people to be able to edit @contexts and see the results.
... We'll need a way for someone to run a local server or something.

(slide 8)

https://hackmd.io/@BsAUvzYBR86xn2LTPFZuVg/

SJ4R4Qu0H

harold: One issue: There are lots of UNKNOWN URLs. if you have a primitive type,

eric: I use something online called RDFConverter to turtlize.

harold: Basic problem: when using primitive types, to extend it an underscore is put in front of the JSON property.
... Second issue is harder: nested resources. IN a number of places in FHIR, there can be a resource inside that can be any resource in FHIR.
... DomainResource has a thing called "contained", which can contain any number of Resources.
... So somewhere the @context has to recognize the resource type to figure out what context to use.

guoqian: Examples of use?

harold: We tried all of the examples at fhir.org and then looked at the ones that failed to build.
... There are examples in there. And a reference can either be by name or by value.
... Might need to preprocess in order to get something out.

eric: I could write a faster C parser that would allow all of FHIR to be in a resource.

harold: Or we could pre-look to see what resource is in there, and put it in. Or there might be a JSON-LD 1.1 feature that would help, but I still need to investigate.
... If yuou have a reference then yuou have to use a bundle.
... parameters are yet another metamodel in FHIR.
... Third issue: some things are not coming out right. I need to investigate more.
... fourth issue: integer64 is somehow showing up. Don't know where it's coming from.
... THe hackmd URL above gives instructions for trying this.
... summary of things to do.

david: Added these TODO items to the github issues list?

harold: Not yet, but I will.
... on the FHIRCat issues list.
... Also noticed JSON-LD Best Practices on github https://github.com/w3c/json-ld-bp
... I added an issue about maintaining list ordering.
... I should have asked for at least having an index variable.
... We can do our solution and post it on the JSON-LD BP issue discussion so that they can see it.

ADJOURNED

This is scribe.perl Revision: 1.154  of Date: 2018/09/25 16:35:56  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

Present: Daniel_Stone Harold_Solbrig David_Booth Guoqian_Jiang EricP Gopi
No ScribeNick specified.  Guessing ScribeNick: dbooth
Inferring Scribes: dbooth

WARNING: No date found!  Assuming today.  (Hint: Specify
the W3C IRC log URL, and the date will be determined from that.)
Or specify the date like this:
<dbooth> Date: 12 Sep 2002

People with action items: 

WARNING: IRC log location not specified!  (You can ignore this 
warning if you do not want the generated minutes to contain 
a link to the original IRC log.)