FHIR Chat · ValueSetReference lost on publishing 3.0.1 · IG creation

Stream: IG creation

Topic: ValueSetReference lost on publishing 3.0.1

Travis Stenerson (Jun 27 2018 at 17:13):

@Grahame Grieve

I think I found a bug in the newest publisher - I'm publishing with FHIR version 3.0.1 with differentials that have binding.valueSetReference. On running the publisher this reference is lost. I didn't get an error message that I can see. A slightly older version of the publisher does not lose this field. I couldn't find any mention of this here so thought I would let you know.

Grahame Grieve (Jun 27 2018 at 20:04):

ok I'm at the end of 2 weeks on the road; I'll be prioritising IG issues early next week

Eric Haas (Jun 27 2018 at 20:21):

I always use ValueSetUri with no issue

Travis Stenerson (Jun 27 2018 at 20:26):

I've got an older copy that works fine for my purposes. Just thought I'd let you know.

Grahame Grieve (Jul 06 2018 at 19:26):

@Travis Stenerson I'm trying to catch up on IG publishing issues. I can't see easily from the code what is the problem here. how do I reproduce this?

Travis Stenerson (Jul 06 2018 at 19:51):

Hi @Grahame Grieve

I produced a differential with this:

{
                "id": "Observation.valueCodeableConcept",
                "path": "Observation.valueCodeableConcept",
                "type": [
                    {
                        "code": "CodeableConcept"
                    }
                ],
                "binding": {
                    "strength": "required",
                    "valueSetReference": {
                        "reference": "https://fhir.evinance.net/baseDstu3/ValueSet/cN-category-values-ajcc8-HAN-MSG"
                    }
                }
            }

And what resulted after running the publisher with those differentials was this:

{
                "id": "Observation.valueCodeableConcept",
                "path": "Observation.valueCodeableConcept",
                "type": [
                    {
                        "code": "CodeableConcept"
                    }
                ],
                "binding": {
                    "strength": "required"
                }
            }

The differential is built with python, the valueset was definitely available. I didn't actually get any errors when it built.
the build file has version 3.0.1

Grahame Grieve (Jul 06 2018 at 19:54):

what resulted after running

where?

Travis Stenerson (Jul 06 2018 at 19:55):

That was the produced StructureDefinition, the full snapshot version. In the output folder.

The version that worked was:
FHIR Implementation Guide Publisher (v3.4.0-13801, gen-code v3.4.0-13757)

The version that didn't seem to work had that five digit number at like 13850?

Grahame Grieve (Jul 06 2018 at 19:56):

ah found it

Travis Stenerson (Jul 06 2018 at 19:57):

13866

Grahame Grieve (Jul 06 2018 at 19:57):

k. committed a fix. wait for the next build, you should be good

Travis Stenerson (Jul 06 2018 at 19:58):

Thanks

Grahame Grieve (Jul 06 2018 at 20:37):

@Christopher Schuler also fixed your issue with v3 value set references

Christopher Schuler (Jul 06 2018 at 20:39):

Awesome, thanks @Grahame Grieve !

Christopher Schuler (Jul 06 2018 at 21:55):

@Grahame Grieve , the binding links are much closer now, but still broken. For example, the url for MedicationDispense SubstanceAdminSubstitutionReason in the snapshot table is http://hl7.org/implement/standards/fhir/STU3/v3/SubstanceAdminSubstitutionReason/index.html, but it should be http://hl7.org/implement/standards/fhir/STU3/v3/SubstanceAdminSubstitutionReason/vs.html

Grahame Grieve (Jul 06 2018 at 21:59):

where is this? but why is it even starting with http://hl7.org/implement/standards/fhir/? that URL should never appear anywhere

Christopher Schuler (Jul 06 2018 at 22:04):

Yeah, that's weird - those links weren't pointing there in previous builds. It is in the QiCore IG: http://build.fhir.org/ig/cqframework/qi-core/StructureDefinition-qicore-medicationdispense.html. Those links I posted were from my local build. I committed that about 5 mins ago.

Grahame Grieve (Jul 06 2018 at 22:55):

ok fixed

Last updated: Apr 12 2022 at 19:14 UTC