Stream: IG creation
Topic: Bad validation of IHE format code
Lloyd McKenzie (Feb 10 2022 at 04:50):
I'm getting the following QA message:
The Coding provided (http://ihe.net/fhir/ValueSet/IHE.FormatCode.codesystem#urn:ihe:pcc:apr:handp:2008) is not in the value set http://hl7.org/fhir/ValueSet/formatcodes, and a code is recommended to come from this value set. (error message = Not in value set http://hl7.org/fhir/ValueSet/formatcodes)
But when you look at the R4 expansion for format codes, that code is absolutely in the set...
Grahame Grieve (Feb 10 2022 at 05:05):
how to reproduce?
Lloyd McKenzie (Feb 10 2022 at 05:21):
Sorry. HRex IG
John Moehrke (Feb 10 2022 at 13:53):
so this is part of the thing I have been trying to get fixed for sometime. IHE has now formally published an IG with their vocabulary for FormatCode. What you have is out dated.
John Moehrke (Feb 10 2022 at 13:54):
https://profiles.ihe.net/fhir/ihe.formatcode.fhir/index.html
John Moehrke (Feb 10 2022 at 13:54):
The codesystem no longer has "ValueSet" in it.
http://ihe.net/fhir/ihe.formatcode.fhir/CodeSystem/formatcode
John Moehrke (Feb 10 2022 at 13:55):
I had asked for advice on if I should keep the old canonical URI, or fix it. I got guidance to fix it.
John Moehrke (Feb 10 2022 at 13:56):
the codeSystem OID has not ever been wrong or changed. ;-)
Lloyd McKenzie (Feb 10 2022 at 16:02):
You got the guidance to fix it from whom? Changing canonicals is very problematic...
In any case, my issue is that the codes that are 'valid' according to THO aren't being treated as valid.
John Moehrke (Feb 10 2022 at 19:25):
yes. agreed
John Moehrke (Feb 10 2022 at 19:34):
at some point one could depend on the formatCode IG, and then use the valueSet. but I can't get that working
Grahame Grieve (Feb 15 2022 at 09:48):
well, where is it defined? The reason it's giving this error is because the tx server doesn't know the code system
Grahame Grieve (Feb 15 2022 at 09:58):
http://ihe.net/fhir/ihe.formatcode.fhir/CodeSystem/formatcode should be right...
Grahame Grieve (Feb 15 2022 at 10:02):
and so the server should know it
Grahame Grieve (Feb 15 2022 at 10:27):
yes your canonical is wrong
John Moehrke (Feb 15 2022 at 13:07):
Grahame Grieve said:
yes your canonical is wrong
@Grahame Grieve who has it wrong? Is there something I need to fix?
Grahame Grieve (Feb 15 2022 at 13:08):
no it's @Lloyd McKenzie that has the wrong canonical
John Moehrke (Feb 15 2022 at 13:10):
I am also not able to use the IHE codesystem... so I don't think it is just the canonical. I have not had the time to figure out exactly what combination of settings works vs doesn't.
Grahame Grieve (Feb 15 2022 at 13:16):
you should be able to use the code system. What are you trying?
Lloyd McKenzie (Feb 15 2022 at 15:34):
The canonical I'm using is the one on THO... How is that wrong?
John Moehrke (Feb 15 2022 at 18:55):
Im guessing that THO has the origianl canonical, not the one IHE currently uses
John Moehrke (Feb 15 2022 at 18:55):
which might be the root of many unusual things
Lloyd McKenzie (Feb 15 2022 at 19:57):
The issue we have is that what's on THO is supposed to be source-of-truth for FHIR (and the validator)
John Moehrke (Feb 15 2022 at 20:37):
I have tried to THO to 'copy' the IHE now published codeSystem... I did not know it had a copy of the old stuff. I understood the old stuff was pulled (via FTP) into the FHIR core build.
John Moehrke (Feb 15 2022 at 20:38):
I was told that THO will not take on a shadow copy, given that anyone using an IG will magically find the IHE published work.
Grahame Grieve (Feb 15 2022 at 20:44):
well, THO has http://ihe.net/fhir/ihe.formatcode.fhir/CodeSystem/formatcode
John Moehrke (Feb 15 2022 at 20:45):
oh, well that is the right one. How did it get in there? I understood from you that it would not do that.
Grahame Grieve (Feb 15 2022 at 20:47):
that's in a value set reference to it. Which is correct. I don't know what @Lloyd McKenzie is looking at, because THO doesn't contain a code system or a naming system - as it shouldn't
John Moehrke (Feb 15 2022 at 20:49):
okay
Lloyd McKenzie (Feb 15 2022 at 21:54):
I was incorrect in blaming THO. The issue is in the R4 publication: http://hl7.org/fhir/valueset-formatcodes.html
Grahame Grieve (Feb 15 2022 at 22:01):
it's not fixed in R5 either. @John Moehrke please fix it ASAP
Grahame Grieve (Feb 15 2022 at 22:02):
I do not know how the R5 expansion is generated. Must be some hack under the covers.
John Moehrke (Feb 15 2022 at 22:18):
I have no idea how either... I think the FHIR core pulls it via FTP.. like it did DICOM..
Rob Hausam (Feb 15 2022 at 23:36):
I can try to help sort that out(?).
Grahame Grieve (Feb 15 2022 at 23:37):
don't worry about that, just correct the code system URL in R5
Lloyd McKenzie (Feb 15 2022 at 23:45):
"Fixing" a URL is problematic - as it's undoubtedly out in the wild the wrong way too. Though the fact that the validator chokes on it helps. Certainly the UTG entry needs to include the 'wrong' URL as an alternate.
Grahame Grieve (Feb 15 2022 at 23:47):
what UTG entry?
Lloyd McKenzie (Feb 16 2022 at 02:31):
We would need a NamingSystem in UTG that says "this is the official URL - and this is another one that's been used". We shouldn't have any code systems referenced in core that UTG doesn't either define or indicate what the official URL for them is.
Grahame Grieve (Feb 16 2022 at 02:32):
well, you can propose one. Though I think that we shouldn't have to. And we don't for technical reasons
Lloyd McKenzie (Feb 16 2022 at 02:58):
"We don't for technical reasons" - we don't have to, or don't want to?
Lloyd McKenzie (Feb 16 2022 at 02:59):
My justification for wanting to is that if we're using a terminology, it ought to be discoverable in our source-of-truth for terminology. Plus it ensures that there's an understood authoritative URL that everyone's expected to use.
Grahame Grieve (Feb 16 2022 at 03:13):
don't have to
Grahame Grieve (Feb 16 2022 at 03:13):
and I don't think it has to be definable if we can link to it.
Grahame Grieve (Feb 16 2022 at 03:14):
though we don't link to it right now. I'll add that to my list at some stage
John Moehrke (Feb 16 2022 at 13:02):
I asked if I should have done something with NamingSystem... crickets at the time.
John Moehrke (Feb 16 2022 at 13:06):
Grahame Grieve said:
don't worry about that, just correct the code system URL in R5
I have tried to change the ValueSet in R5, and it won't build. My Branch
https://github.com/hl7/fhir/tree/moehrke-formatCode
John Moehrke (Feb 16 2022 at 13:08):
not sure why i didn't make a pull-request.. but likely didn't because it fails to build.
Grahame Grieve (Feb 16 2022 at 13:13):
what was the error? I can't build R5 right now - some weird OSX thing - I'm gradually getting through them
Last updated: Apr 12 2022 at 19:14 UTC
