Stream: conformance
Topic: Strange terminology 'system' QA warning messages
Rob Hausam (Aug 08 2019 at 02:33):
I'm doing some slicing testing of AllergyIntolerance.code discriminating on a value set binding (I think that's actually "legal", or at least it's implied that it could be). Anyway, it seems to be working and is behaving as expected in the IG Publisher/Validator. But when testing with an instance that contains additional valid SNOMED CT and LOINC Coding instances beyond the required binding of the slice to absent-or-unknown-allergies-uv-ips, the Validator is giving this warning:
AllergyIntolerance/eumfh-39-07-1: AllergyIntolerance.code warning [Unable to resolve system http://snomed.info/sct - system is not complete, Unsupported system http://loinc.org - system is not specified or implicit]
Since both of those urls are the correct canonical url's for those code systems, there isn't any way that this particular warning should ever occur.
Rob Hausam (Aug 08 2019 at 03:14):
Actually, the warning didn't occur in the CI build - only in my local build. I don't know why that would be the case, but I still don't think the warning should ever occur (unless somehow the local build is failing to recognize these code systems at all, which is still would obviously be an issue).
Grahame Grieve (Aug 08 2019 at 05:38):
somehow you have the snomed definition code system resource in scope
Rob Hausam (Aug 08 2019 at 06:33):
Hmm. I haven't done anything to reference that in what I'm testing or otherwise in the IG.
Rob Hausam (Aug 08 2019 at 13:28):
@Grahame Grieve Can you provide further details on why you think this warning means that I have "the snomed definition code system resource in scope"?
Grahame Grieve (Aug 08 2019 at 18:02):
because that is the only way you could get 'snomed is not complete'
Rob Hausam (Aug 08 2019 at 18:18):
Are you sure that it's the only way? In the IPS IG I just modified the slicing in the AllergyIntolerance-uv-ips profile to slice AllergyIntolerance.code by value set rather than by coding.system, and I added a SNOMED and a LOINC Coding to the CodeableConcept in the example instance (along with the absent-unknown-uv-ips Coding that was already there) - that's it. I'm not thinking of anything in the IG (before or after the modifications) that references the definitional SNOMED CT CodeSystem resource - there's just SNOMED value sets.
Grahame Grieve (Aug 08 2019 at 18:24):
no it's somewhere in the core space. This is in the IG?
Rob Hausam (Aug 08 2019 at 18:26):
right - yes, I see it in the qa.html when I run the IG build, or when I use the validator on the example file while specifying the ips IG
Rob Hausam (Aug 08 2019 at 18:38):
I'm running this command locally: java -jar ~/Downloads/org.hl7.fhir.validator.jar ~/git-repo/fhir-ips/examples/allergyIntolerance-39-07-1.xml -ig hl7.fhir.uv.ips
against the CI build (not the ballot version) of ips with this modified example:
allergyIntolerance-39-07-1.xml
(I'm still experimenting, and this is just the latest iteration)
Last updated: Apr 12 2022 at 19:14 UTC
