Stream: genomics
Topic: IG FHIR Version
Patrick Werner (Nov 20 2019 at 14:15):
I just tried to derive a profile from our IG with Forge, unfortunately Forge (and CIMPL) don't support FHIR version 4.0.1 .
So downstream usage of our profiles is restricted at the moment (i heard that the Forge Update will take some time). So i was wondering if we should address this issue, we could:
- Switch back to 4.0.0 (i think its to late for that now as publishing is currently being done)
- build the IG in a 4.0.0 version and put it in the ci build to be downloaded as a zip as an alternative for now?
Jamie Jones (Nov 20 2019 at 14:25):
What are the changes in 4.0.1?
Jamie Jones (Nov 20 2019 at 14:26):
If 4.0.0 allows to use better tooling it may be worth briefly downgrading the target version until it is supported.
That said, I'm not sure Forge is a one-stop shop solution for us?
Kevin Power (Nov 20 2019 at 14:31):
- Switch back to 4.0.0 (i think its to late for that now as publishing is currently being done)
- build the IG in a 4.0.0 version and put it in the ci build to be downloaded as a zip as an alternative for now?
If those are the options, I vote #2 - but not sure that is an option. When I try to build with 4.0.0 locally, I get this error:
[java] java.lang.Exception: Error: the IG declares that is based on version 4.0.0 but this IG publisher only supports publishing the following versions: 1.0.2, 1.4.0, 3.0.2, 4.0.1, 4.1.0
Bret H (Nov 21 2019 at 11:27):
drat
Lloyd McKenzie (Nov 21 2019 at 15:03):
You don't want to be on 4.0.0. 4.0.1 fixes a bunch of stuff that was completely broken. Forge should have a release that supports 4.0.1 "soon". Don't create a broken version just to cover a 1-2 month gap in coverage.
Patrick Werner (Nov 21 2019 at 15:17):
i heard the Forge "soon" will be quite long this time
Patrick Werner (Nov 21 2019 at 15:19):
But i'll try to work around it. Shouldn't i be fine to just replace 4.0.1 with 4.0.0 in the StrucDefs ? Or did 4.0.1 introduce breaking changes which will blow up my idea?
May Terry (Nov 21 2019 at 16:52):
We're working on CIMPL tooling support for 4.0.1, but it won't be for 2-3 weeks and it's also contingent on a fix for an IG Publisher 4.0.1 handling bug that Chris Moesel found and reported to Grahame.
Lloyd McKenzie (Nov 21 2019 at 20:01):
4.0.1 introduces things that impact tools, but shouldn't impact designed profiles. So authoring against 4.0.0 should be fine
Mullai Murugan (Dec 11 2019 at 20:48):
Great news! Newbie question - We are currently loading the draft profiles into our HAPI server, do we now manually load the new STU release? Or would it automatically be available as part of the 4.0.1 release?
Jamie Jones (Dec 11 2019 at 20:49):
As far as I know, HAPI shouldn't have any IG profiles by default, unless it's @Patrick Werner 's HAPI server at MOLIT ;)
Mullai Murugan (Dec 11 2019 at 20:56):
@Jamie Jones thanks. To confirm, we would still need to load the CG profiles to the HAPI server.
Mullai Murugan (Dec 13 2019 at 16:12):
@Jamie Jones is this the right url for CG profiles - http://build.fhir.org/ig/HL7/genomics-reporting/package.tgz?
Jamie Jones (Dec 13 2019 at 16:17):
That looks like it would have all the definitions, yes. We also included a 'downloads' tab on the IG, which (for the published stu 1.0 version) is available here http://hl7.org/fhir/uv/genomics-reporting/downloads.html
Just the artifact definitions should be all you need to add to a server, let us know if you run into any difficulty--some implementations throw errors on a few of our profiles as the "narrative" section of the structureDefinition is very large.
Jamie Jones (Dec 13 2019 at 16:19):
you can also use the java validator as we describe it on http://hl7.org/fhir/uv/genomics-reporting/index.html#understanding-fhir
Mullai Murugan (Dec 13 2019 at 17:30):
Thank you, @Jamie Jones !
Last updated: Apr 12 2022 at 19:14 UTC
