FHIR Chat · Build failure - duplicate resources in IG?

Stream: genomics/committers

Topic: Build failure - duplicate resources in IG?

Kevin Power (Jul 07 2020 at 14:37):

I just upgraded my device, and as a test I tried to get my FHIR IG builds going again. When I tried to build the latest version of genomics-reporting, I am getting the following errors:

     [java] Load Package hl7.fhir.pubpack#0.0.6
     [java] Load Package hl7.fhir.xver-extensions#0.0.4
     [java] Load Content                                                                     (00:17.0500)
     [java] Duplicate Resource in IG: ValueSet/hgnc
     [java] Duplicate Resource in IG: ValueSet/dna-change-type
     [java] Duplicate Resource in IG: ValueSet/variant-inheritance
     [java] Duplicate Resource in IG: ValueSet/high-low-codes
     [java] Duplicate Resource in IG: ValueSet/functional-annotation
     [java] Duplicate Resource in IG: ValueSet/hgvs
     [java] Duplicate Resource in IG: ValueSet/seq-phase-relationship
     [java] Exception in thread "main" java.lang.Error: Unable to continue because duplicate input resources were identified
     [java]     at org.hl7.fhir.igtools.publisher.Publisher.load(Publisher.java:2966)
     [java]     at org.hl7.fhir.igtools.publisher.Publisher.createIg(Publisher.java:841)
     [java]     at org.hl7.fhir.igtools.publisher.Publisher.execute(Publisher.java:707)
     [java]     at org.hl7.fhir.igtools.publisher.Publisher.main(Publisher.java:7756)

BUILD FAILED
/Users/cerkyp/Documents/GitHub/genomics-reporting/framework/build.xml:161: Java returned: 1

Kevin Power (Jul 07 2020 at 14:38):

Anyone else get this, or just me? @Patrick Werner / @Jamie Jones ?

Jamie Jones (Jul 07 2020 at 14:39):

local build? I'll take a look

Kevin Power (Jul 07 2020 at 14:40):

Yes, local. Was afraid to trigger the cloud build in case it bombs there too.

Jamie Jones (Jul 07 2020 at 15:33):

error confirmed. looks like our valueset declarations need some thought after recent updates

Kevin Power (Jul 07 2020 at 15:34):

'Recent updates' ? You mean IG Publisher updates, or IG updates?

Jamie Jones (Jul 07 2020 at 16:04):

publisher/framework, no IG content changes since May

Kevin Power (Jul 07 2020 at 16:06):

I can't seem to find what is causing the problem. :frown:

Kevin Power (Jul 07 2020 at 18:24):

FYI - Found someone who had a similar issue, and added our issue on this topic: https://chat.fhir.org/#narrow/stream/179252-IG-creation/topic/New.20release.201.2E101.20fails

Jamie Jones (Jul 07 2020 at 20:09):

all 7 value sets we define as bindings in the observation spreadsheet are the ones generating errors. TBD codes is the only one not generating an error, and it is the only one not referenced on the "Bindings" tab in Obs.

Grahame Grieve (Jul 07 2020 at 22:14):

this is a genuine error in the IG as far as I am concerned. I added additional error checking after it turned out that other IGs were having unpredictable outcomes because of duplicate resources

Kevin Power (Jul 07 2020 at 23:11):

Ok we can try removing the ‘extra’ references on the bindings tab of the spreadsheet.

Jamie Jones (Jul 08 2020 at 01:25):

Tried a few things without luck

Patrick Werner (Jul 08 2020 at 12:50):

Grahame Grieve said:

this is a genuine error in the IG as far as I am concerned. I added additional error checking after it turned out that other IGs were having unpredictable outcomes because of duplicate resources

these VS resources weren't duplicates. They were just referenced via their filename (without endings). They are only defined as xml resources

Patrick Werner (Jul 08 2020 at 12:53):

i just tried to switch to their urls, no luck. Now getting:

  [java] Exception in thread "main" java.lang.Error: Access single value, but value count is 0
     [java]     at org.hl7.fhir.r5.renderers.utils.ElementWrappers$PropertyWrapperMetaElement.value(ElementWrappers.java:303)
     [java]     at org.hl7.fhir.r5.renderers.DiagnosticReportRenderer.render(DiagnosticReportRenderer.java:67)
     [java]     at org.hl7.fhir.r5.renderers.ResourceRenderer.render(ResourceRenderer.java:76)
     [java]     at org.hl7.fhir.igtools.publisher.Publisher.generateNarratives(Publisher.java:1043)
     [java]     at org.hl7.fhir.igtools.publisher.Publisher.loadConformance(Publisher.java:3632)
     [java]     at org.hl7.fhir.igtools.publisher.Publisher.createIg(Publisher.java:845)
     [java]     at org.hl7.fhir.igtools.publisher.Publisher.execute(Publisher.java:707)
     [java]     at org.hl7.fhir.igtools.publisher.Publisher.main(Publisher.java:7756)

Kevin Power (Jul 08 2020 at 13:18):

So, in reading the spreadsheet authoring guidance, it says that when we indicate a binding, it need to point to a binding name on the Bindings tab:

Binding: Conditional. This must be specified for any data element that allows a type of code, Coding or CodeableConcept (including "*"). It should not be present for other elements. It must refer to a Binding name from either the #Bindings Tab in the current spreadsheet or one of the other resource spreadsheets. (Note - the order in which resources are first loaded into memory is controlled by the build/source/fhir.ini file. Resources that define bindings must be listed before other resources that use them.

We have XML files for these ValueSets, and we give them a 'binding name' on the Bindings tab of the spreadsheet. So I am not sure why they are now flagged as duplicates?

Jamie Jones (Jul 08 2020 at 13:19):

i'm not sure what is happening behind the xslt curtain in the generation steps

Kevin Power (Jul 08 2020 at 13:21):

Yet another reason to get off these spreadsheets :frown:

Patrick Werner (Jul 08 2020 at 13:48):

Kevin Power said:

So, in reading the spreadsheet authoring guidance, it says that when we indicate a binding, it need to point to a binding name on the Bindings tab:

Binding: Conditional. This must be specified for any data element that allows a type of code, Coding or CodeableConcept (including "*"). It should not be present for other elements. It must refer to a Binding name from either the #Bindings Tab in the current spreadsheet or one of the other resource spreadsheets. (Note - the order in which resources are first loaded into memory is controlled by the build/source/fhir.ini file. Resources that define bindings must be listed before other resources that use them.

We have XML files for these ValueSets, and we give them a 'binding name' on the Bindings tab of the spreadsheet. So I am not sure why they are now flagged as duplicates?

exactly. we only are defining the bindings in the spreadsheets, referring to the external (xml StrucDefs in src/resources/vocabulary) defined VSs

Lloyd McKenzie (Jul 08 2020 at 14:01):

That definitely shouldn't be treated as a duplicate

Kevin Power (Jul 08 2020 at 15:05):

So, where does that leave us? We haven't found a way to get around this (yet)?

Kevin Power (Jul 10 2020 at 13:36):

After some conversations with @Patrick Werner and @Jamie Jones , it seems this might be the time to move off the spreadsheets for authoring. I logged J#27942 for the workgroup to consider.