Meeting minutes
Introductions
max: From Medtronic, building interactions w FHIR and EHRs. Investigating using RDF for representing our data. Primarliy work on arch. Saw the video w Harold on uing JSON-LD into RDF.
dbooth: Still plannig to use JSON-LD?
max: Yes, using the framing apect of JSON-LD -- framing service for JSON data.
… RDF comes into the backend side for transferring.
eric: Is your framing target isomorphic to the graph structure.
max: IDK yet.
eric: JSON-LD does not invent nodes.
Concept IRIs
gaurav: No progress since last week.
… Still wrinting a prog to pull all the code sys and naming sys from the repo.
HAPI
eric: Resurrected the HAPI server. Still focused on jena shex. Also working w Deepak on validating the examples against the shex.
… There's a bug in Jim's generated RDF. For codings, the datatype is a fhir:code, and the XML of that is an xsd:token, but they were represented as xsd:strings, so they were rejected in the validation.
… But since it was only codes, deepak hand edited the schema for code to allow either string or xsd:token.
OWL to FHIR transformers
gaurav: THere were some old ones aehrc/
… Foundational model of anatomy
https://
… And TSMG is discussing this, also about CPT.
#narrow/stream/179202-terminology/topic/AMA.2C.20HL7.20Partner.20on.20FHIR.20CPT.20Integration.20to.20Drive.20Interop
… It sounds like AMA agreed to let CPT to be shared in FHIR.
eric: Only as a code system, or also logical models?
gaurav: Code sys.
… FMA is not going to terminology.hl7.org. Someone just wanted to use it.
… I think as a code system.
… I think that's what the OWL-to-FHIR tools do.
eric: As giant valuesets?
gaurav: The new tool was demoed, and it turned it into a massive code set -- one term every every term in the ont.
… They were working on HPO ont.
eric: The OBO Foundry has formal rules about modeling things. FMA is Foundational Model of Anatomy -- about what is a part of what.
… CPT is from AMA.
… There was a press release saying that AMA is partnering w HL7.
gaurav: They're also looking at how they can put the terminology into terminology.hl7.org, but might have a licensing issue.
eric: what use cases?
gaurav: If the terms exist in HTO then presumably they could validate
FHIR ShEx
eric: it has about 1500-2k shapes, and you need a dedicated shape for every list, because it needs to be able to point to itself, so it needs a label.
… Is there willpower and energy to add templates to ShEx to allow you, instead of creating ListOfX, to say you have a llist template w parameter X.
… Purpose would be to get that done before finishing shex "extends", to get it into the IEEE standard.
dbooth: Like a macro expansion?
eric: That's one of the big questions. Could be a syntax trick. Or it could be built into the semantics of the language, and gives more power and allows round tripping.
… To do it right would require someone think hard about it.
dbooth: Sounds worthwhile.
ADJOURNED