Stream: IG creation
Topic: Rendering Argo extensions in custom IG
Chris Moesel (Apr 01 2019 at 15:05):
We have a DSTU2 IG with some profiles based on Argonaut profiles. Our profile on ArgoPatient uses the argo-defined extensions, but when we use the publisher, it renders the extensions in a strange way. For example:
pasted image
Here is the snapshot element for it:
pasted image
Is this a known issue? Do I potentially have something wrong in my config or something?
Lloyd McKenzie (Apr 01 2019 at 15:08):
The argonaut extensions are defined in the IG and have the right canonical URL in their defining structure definitions and in the ImplementationGuide (and in the referencing structure definition)
Chris Moesel (Apr 01 2019 at 15:21):
Thanks, @Lloyd McKenzie -- I'm using the right canonical URL but I just realized that a while back I had commented out the dependency on the Argonaut IG in my ig.json file. I suspect maybe that's it. That said, I think I commented it out because it wasn't working. I guess we'll see!
Chris Moesel (Apr 01 2019 at 15:28):
Now I see why I commented it out. The IG Publisher tries to find a package-list.json, but there is none for Argonaut. I already reported this a while back, but as best I can tell, no progress has been made:
https://chat.fhir.org/#narrow/stream/179252-IG-creation/topic/IG.20Dependency.20on.20Argonaut
Chris Moesel (Apr 01 2019 at 15:28):
@Brett Marquard or @Eric Haas -- is there a package-list.json for Argonaut somewhere?
Eric Haas (Apr 01 2019 at 15:57):
not yet
Grahame Grieve (Apr 02 2019 at 19:00):
it's on my task list for tomorrow
Grahame Grieve (Apr 02 2019 at 19:00):
as part of publishing argonaut IGs
Chris Moesel (Apr 02 2019 at 19:28):
I've noticed that the Argonaut IG also violates several DSTU2 constraints, most notably when it specifiesids that use :, [, and ] characters (as is now the recommended practice in R3/R4). In DSTU2, the id cannot be > 64 characters and it only supports A-Z, a-z, 0-9, -, and .. The IG publisher spits out tons of errors if you try to use :/[/] in a DSTU2 IG. @Grahame Grieve -- I'm wondering if you're doing anything to address that when re-publishing the Argonaut IGs? I'm still trying to figure out the best approach to handling it myself -- as I find those characters very useful in ids.
Grahame Grieve (Apr 02 2019 at 19:31):
you know that in later versions, id has a fixed value?
Grahame Grieve (Apr 02 2019 at 19:31):
otherwise I'll add it to my list
Chris Moesel (Apr 02 2019 at 19:37):
By "fixed value", do you mean the specific algorithm regarding how the id is formulated? (e.g., how slices and choices get encoded into the path to create the id)
Chris Moesel (Apr 02 2019 at 19:38):
I actually want to follow that in DSTU2 as well, but that darned definition of id won't allow it!
Grahame Grieve (Apr 02 2019 at 19:39):
yes and yes
Chris Moesel (Apr 02 2019 at 19:43):
But for better or worse, I just discovered a work-around that may or may not be intentional. If I set the name to the value I really want for the id, then the publisher actually overwrites the id with the value of name -- and doesn't complain when that violates the spec!
Grahame Grieve (Apr 02 2019 at 19:44):
well at least the second outcome is not intended
Chris Moesel (Apr 02 2019 at 19:44):
OK. Then I guess I shouldn't leverage that little quirk of the publisher. Oh well...
Grahame Grieve (Apr 02 2019 at 19:45):
do you really need to work in R2 ?
Chris Moesel (Apr 02 2019 at 19:46):
Since name is dual-purpose in DSTU2 and also used for the slicename, however -- that little feature (copying name to id) makes it hard for me to implement any consistent pattern in my id.
Chris Moesel (Apr 02 2019 at 19:46):
The mCODE initiative is DSTU2.
Grahame Grieve (Apr 02 2019 at 19:46):
oh? ok. I didn't realise mCode was R2. Why?
Chris Moesel (Apr 02 2019 at 19:49):
Because that's where the vendors were when the decision was made -- with no published release dates for updates -- and a goal of _mCODE_ is to work today rather than tomorrow. It's hard to pilot R3 or R4 profiles on systems that only support Argonaut.
Chris Moesel (Apr 02 2019 at 19:50):
At least that's my understanding of it. Essentially the same reason I'm guessing for why that new ONC-proposed rule specifies DSTU2.
Chris Moesel (Apr 02 2019 at 19:51):
(Although I understand that there is some pushback on that)
Grahame Grieve (Apr 02 2019 at 19:51):
yes pushback on the rule for the future.
Grahame Grieve (Apr 02 2019 at 19:51):
I guess I can't fault that logic
Chris Moesel (Apr 02 2019 at 19:52):
Trust me, I'd much rather be working with R4. ;-)
Last updated: Apr 12 2022 at 19:14 UTC
