Stream: IG creation
Topic: Errors and warnings in IG
Bryn Rhodes (Jul 30 2019 at 05:07):
So, I _think_ that the rest of the errors in the CPG IG are either publishing or template related. I'm using templates that originally came from the build IGs, where can I find the latest greatest templates to see if some of the errors I'm seeing are related to my templates?
Grahame Grieve (Jul 30 2019 at 05:09):
I think, at this point, we don't really have the concept of latest and greatest templates
Bryn Rhodes (Jul 30 2019 at 05:14):
A lot of the broken links feel like I just have a bad template somewhere, but I can't find anywhere that looks like it would do what's happening in the links, so I'm not sure where to go from here.
Grahame Grieve (Jul 30 2019 at 05:14):
which ig is this?
Bryn Rhodes (Jul 30 2019 at 05:15):
Clinical Guidelines. This error, for example: http://build.fhir.org/ig/HL7/cqf-recommendations/branches/master/StructureDefinition-cpg-activitydefinition-definitions.html#75d6950f-da07-4882-8da4-bdd5cd42a270
Bryn Rhodes (Jul 30 2019 at 05:15):
Basically, every extension defined in the IG resolves incorrectly
Bryn Rhodes (Jul 30 2019 at 05:16):
It has a '/' instead of a '-'
Grahame Grieve (Jul 30 2019 at 05:17):
that means that you haven't told the IG publisher where you've organised for the resource to generate to in the ig.json
Bryn Rhodes (Jul 30 2019 at 05:17):
Ah.
Bryn Rhodes (Jul 30 2019 at 05:17):
I can fix that. Thanks!
Bryn Rhodes (Jul 30 2019 at 16:54):
Okay, got past those, the next problem I have is that I have lots of examples in this IG that are also conformance resources, but I don't want to use the IG's canonical, they came from different IGs and I want to use the canonical for where the examples come from. The publisher complains that my canonicals don't match, but I've indicated they are examples in the ImplementationGuide. Am I okay to suppress these errors and warnings?
Grahame Grieve (Jul 30 2019 at 19:27):
is there any reason why a conformance resource labelled as an example should cause an error if it doesn't match the canonical URL?
Bryn Rhodes (Jul 30 2019 at 19:35):
I don't think so
Grahame Grieve (Jul 30 2019 at 21:07):
fixed next release
Bryn Rhodes (Aug 01 2019 at 20:35):
So I'm trying to resolve several warnings about not being able to resolve a profile definition that's defined in my IG. For example, I'm getting this: StructureDefinition reference "http://hl7.org/fhir/uv/cpg/StructureDefinition/cpg-codesystem" could not be resolved, but that's the right canonical URL and as far as I can tell, the resource is correctly identified in the ig.json and the cpg.xml. Anyone seen this?
Grahame Grieve (Aug 01 2019 at 22:01):
yes. What version IG publisher are you using?
Bryn Rhodes (Aug 01 2019 at 22:52):
Just pulled it off the build site today: https://oss.sonatype.org/service/local/repositories/snapshots/content/org/hl7/fhir/publisher/org.hl7.fhir.publisher.cli/0.9.43-SNAPSHOT/org.hl7.fhir.publisher.cli-0.9.43-20190723.110618-1.jar
Grahame Grieve (Aug 01 2019 at 23:03):
wrong site. there's some issue - which they don't think of as a bug - that the latest is not actually the latest. use this link:
Grahame Grieve (Aug 01 2019 at 23:04):
https://fhir.github.io/latest-ig-publisher/org.hl7.fhir.publisher.jar
Bryn Rhodes (Aug 01 2019 at 23:05):
Is that what the auto-publisher is using?
Grahame Grieve (Aug 01 2019 at 23:05):
yes
Bryn Rhodes (Aug 01 2019 at 23:06):
Okay, it's still in the QA errors on the build site. But yes, I'll use the publisher from that location going forward.
Grahame Grieve (Aug 01 2019 at 23:06):
I'll have a look then
Grahame Grieve (Aug 02 2019 at 01:12):
so the first error is http://hl7.org/fhir/uv/cpg/StructureDefinition/cpg-ordersetdefinition - I can't find that.
Grahame Grieve (Aug 02 2019 at 01:33):
in fact I can't find anything that the IG publisher can't
Bryn Rhodes (Aug 02 2019 at 03:18):
Yeah, that's what's confusing me. It's in the ig.json resources list (StructureDefinition/cpg-ordersetdefinition), it's in the cpg.xml (by the same relative id), and it builds to the profiles list here: http://build.fhir.org/ig/HL7/cqf-recommendations/StructureDefinition-cpg-ordersetdefinition.html.
Bryn Rhodes (Aug 02 2019 at 03:18):
I've verified the canonical is correct: http://hl7.org/fhir/uv/cpg/StructureDefinition/cpg-ordersetdefinition and the StructureDefinition resource has the correct id: cpg-ordersetdefinition
Bryn Rhodes (Aug 02 2019 at 03:19):
All the ones that give the warning are like that.
Bryn Rhodes (Aug 02 2019 at 03:19):
The weird thing is, there are several that don't give the warning, they work just fine, but I can't find a pattern.
Bryn Rhodes (Aug 02 2019 at 03:20):
I thought it might be that it's a derived profile, but it's happening with cpg-codesystem, which is straight from CodeSystem.
Grahame Grieve (Aug 02 2019 at 03:23):
the problem for me is that I can't find any resource claiming to be cpg-codesystem
Grahame Grieve (Aug 02 2019 at 03:23):
so it doesn't surprise me that the IG publisher can't resolve it
Bryn Rhodes (Aug 02 2019 at 03:27):
It's a profile created from a spreadsheet.
Grahame Grieve (Aug 02 2019 at 03:33):
spreadsheet... oh.
Grahame Grieve (Aug 02 2019 at 13:13):
new version released - you'll get a different set of errors now
David Pyke (Aug 02 2019 at 13:16):
Definitely do:
[java] Internal error in location for message: 'Error in ?? at 1, 1: Found / expecting a token name', loc = '/home/loftwyr/FHIR/eLTSS/output/CarePlan-eltss.html#/html/body/div/div/div/div/div/div/div/div/div/t', err = 'The image source 'icon_fixed.gif' cannot be resolved'
David Pyke (Aug 02 2019 at 13:19):
Whatever you did, please undo it. You broke a lot more than you fixed
David Pyke (Aug 02 2019 at 13:19):
(deleted)
David Pyke (Aug 02 2019 at 13:20):
All my extensions broke
David Pyke (Aug 02 2019 at 13:21):
us/eLTSS/ImplementationGuide/eltss error Conformance resource /home/loftwyr/FHIR/eLTSS/src-generated/resources/eltss.xml - the canonical URL (http://hl7.org/fhir/us/eLTSS/ImplementationGuide/hl7.fhir.us.eltss-0.1.0) does not match the URL (http://hl7.org/fhir/us/eLTSS/ImplementationGuide/eltss)
ImplementationGuide.url error Resource id/url mismatch: hl7.fhir.us.eltss-0.1.0/http://hl7.org/fhir/us/eLTSS/ImplementationGuide/eltss
ImplementationGuide.id error Resource id/id mismatch: eltss/http://hl7.org/fhir/us/eLTSS/ImplementationGuide/eltss
Lloyd McKenzie (Aug 02 2019 at 14:16):
Are you using my template David?
David Pyke (Aug 02 2019 at 14:38):
Yep
Lloyd McKenzie (Aug 02 2019 at 15:32):
I'm working on an update to address most of those. Some of them are wrong. (e.g. version should never be in canonical URL)
Grahame Grieve (Aug 02 2019 at 21:44):
ok, well, I'm just issuing a new IG publisher that fixes the icon_fixed issue, and the issue with the id/url of the implementation guide itself
Bryn Rhodes (Aug 02 2019 at 21:58):
Thank you Grahame
Lloyd McKenzie (Aug 03 2019 at 03:28):
@David Pyke (and others). davinci-pas and davinci-pcde are now happy with the appropriate identifiers. For those using my framework, you'll need to do the following:
- grab the new framework/build.xml
- edit your src/properties.txt and add a new item that looks like this:
guideid=hl7.fhir.us.davinci-pcde
- edit your src/[igcode].xml file and change the id and url to be the id rather than the code (e.g. hl7.fhir.us.davinci-pcde instead of davinci-pcde). Note: do not change the name of your ig
- delete the src-generated/resources/[igcode].xml file
If you run into any issues, holler.
David Pyke (Aug 03 2019 at 17:20):
Just to ask, is there a central location where all the framework files live or should I just DL it from one of yours?
Lloyd McKenzie (Aug 03 2019 at 17:22):
No central place (yet). But working on that today :)
Bryn Rhodes (Aug 03 2019 at 17:29):
I'm now getting an error in QI-Core that there are some bad links in the generated definitions page for a complex extension, specifically the sub-elements of the extension are being generated as targeting valueCodeableConcept, instead of value[x]. Is this something I am doing wrong in the extension definition? (in the qicore-encounter profile)
Mark Kramer (Aug 03 2019 at 17:30):
@Bryn Rhodes , you should have used CIMPL.
Bryn Rhodes (Aug 03 2019 at 17:32):
That may well be, but as you well know, QI-Core predates CIMPL, and I've been unable to marshal the required capacity to support the switch.
Mark Kramer (Aug 03 2019 at 17:33):
In fact, QI-Core is the reason CIMPL exists. Que sera, que sera.
Lloyd McKenzie (Aug 03 2019 at 17:39):
@Bryn Rhodes The rules for handling polymorphic types have changed a bit. Asserting valueCodeableConcept no longer constrains the type. To constrain the type you have to do it using value[x]. Not sure if that's your problem or not
Bryn Rhodes (Aug 03 2019 at 17:52):
In the extension definition in the spreadsheet, I just list the element codes (i.e. procedure.type, procedure.rank) which map to extension elements, so I'm not specifying the underlying element. I've verified that the generated extension StructureDefinition is using value[x], so it feels like something in the generation of the definitions page.
Lloyd McKenzie (Aug 03 2019 at 17:58):
If it's a link error in the definitions page, then something is going wonky in the snapshot definition. I.e. there's a path in the differential that isn't matched by an identical path in the snapshot.
David Pyke (Aug 03 2019 at 18:03):
I was getting that. I had to unslice the extensions in their definition and in the resource that was using them. That made the error go away. I never asked for them to be sliced, but there you go.
Grahame Grieve (Aug 03 2019 at 21:08):
I'm still scrambling to catch up to the consequences of the changes to slicing by type that we agreed to in Seattle in the generated html. It's definitely caused by tooling, and we'll go to ballot with the broken links if I don't have time to fix them
Eric Haas (Aug 09 2019 at 00:29):
The rules for handling polymorphic types have changed a bit. Asserting valueCodeableConcept no longer constrains the type. To constrain the type you have to do it using value[x]. Not sure if that's your problem or not
Just catching up after vacation. Brand new US Core QA errors...
error The link 'StructureDefinition-us-core-pulse-oximetry-definitions.html#Observation.component:flow-rate.valueQuantity' for "valueQuantity" cannot be resolved (valid targets: [page-content-wrapper, Observation...
Do these need to be value[x] with type Quantity now?
Grahame Grieve (Aug 09 2019 at 01:44):
no. I just have a few dangling issues there
Grahame Grieve (Aug 10 2019 at 22:30):
Formal Announcement: FHIR IG Creation Training, Sept 11/12 2019, Atlanta
In the lead-in to the Atlanta WGM, HL7 will be running several tutorials
aimed at increasing our capacity to create and publish high quality FHIR
Implementation Guides (IGs).
One of those tutorials is focused on creating IGs. It covers:
- how to build community consensus
- how to author an Implementation Guide - the technical process and tools
- testing your IG, and quality works
- understanding the HL7 approval process
Our purpose in running this tutorial is to build the FHIR community's
capacity to produce IGs - demand currently exceeds supply by quite a bit.
The total number of attendees is limited. We are inviting expressions of
interest to attend the tutorial. Priority will be given to attendees who
will build our capacity. There is no cost for the tutorial (ONC is covering
the cost - thanks); instead, we are looking for in-kind contributions to
our capacity. Note that we are not covering travel costs for this tutorial.
To register your interest, send an email to wkubick@hl7.org by Aug 17
(COB US eastern) and explain, in 3 paragraphs or less:
- your previous experience with FHIR, implementations, and IG creation
- which organizations in the FHIR community you have worked with
- how your participation will increase our capacity to build IGs
We expect to record the tutorial and make it available, though there
will be strong hands on aspects to the tutorial. We also plan to make
additional follow up tutorials subsequently.
Grahame Grieve (Aug 10 2019 at 22:36):
p.s watch out for additional tutorial notices appearing shortly
Grahame Grieve (Aug 10 2019 at 22:37):
Faculty, for those interested, is @Sarah Gaunt @David Pyke @Nagesh Bashyam, with support from @Lloyd McKenzie @Melva Peters and myself
David Hay (Aug 11 2019 at 00:21):
Sounds awesome! Hope it goes well and is repeated in Aus next year...
David Pyke (Aug 11 2019 at 16:14):
As this is ONC sponsored, that would require an AUS sponsor for it to happen. But I'm game to teach it whereever.
Grahame Grieve (Aug 11 2019 at 21:02):
ONC might be interested in supporting running it in Au, but they might not be able to either. We'll see how it goes and then re-assess after Sept
Sarah Gaunt (Aug 12 2019 at 21:41):
Might be good to put a "proper" announcement about this out sooner rather than later - people I've talked to about this have already booked flights for Atlanta...
Grahame Grieve (Aug 13 2019 at 21:21):
oh absolutely. The announcement was too late. I'm sorry about that
Last updated: Apr 12 2022 at 19:14 UTC
