Stream: implementers
Topic: GraphDefinition
Grahame Grieve (Feb 09 2017 at 12:22):
I've added a new extremely draft resource to the build. you can find it here: http://build.fhir.org/graphdefinition.html
Grahame Grieve (Feb 09 2017 at 12:22):
This arose out of the Wednesday morning dsicussion in San Antonio
Grahame Grieve (Feb 09 2017 at 12:25):
This is a follow on to this thread: https://chat.fhir.org/#narrow/stream/implementers/subject/Cross-Resource.20Referecnces
Grahame Grieve (Feb 09 2017 at 12:25):
comments are welcome
Grahame Grieve (Feb 09 2017 at 12:26):
procedurally, it seems unlikely that this resource will actually be included in STU3, but I've defined it now because it shows a hole in what we have, and that we intend to do something about it at some stage
nicola (RIO/SS) (Feb 09 2017 at 13:24):
Now we also need GraphBundle to return defined graph or ability replace reference elements with referenced resources bodies :)
Lloyd McKenzie (Feb 09 2017 at 14:55):
A regular Bundle should be fine. We might define operations that allow you to specify a GraphDefinition and the id(s) of the root resources and return a Bundle containing the graph(s).
Grahame Grieve (Feb 09 2017 at 18:49):
don't need an operation for that. already says on the page how that would work
Lloyd McKenzie (Feb 09 2017 at 19:04):
Cool. Presume documentation of that will be added to the search page and the additional component will be reflected in $document? Minor nit: Graph.link.description should come after Graph.link.max as a more natural order to read the resource.
Grahame Grieve (Feb 09 2017 at 19:06):
yes we'd add all that. But I'm not going to add that until it's approved to be in the spec
Lloyd McKenzie (Feb 09 2017 at 19:06):
So we'll need a resource proposal, get that approved by FHIR-I and then FMG
Grahame Grieve (Feb 09 2017 at 19:07):
should we do that for R3? I haven't decided
Josh Mandel (Feb 09 2017 at 19:09):
For R3? This seems like the wrong time (given that I can't quite understand what I'm reading yet)
Grahame Grieve (Feb 09 2017 at 19:10):
y that's what I think
Grahame Grieve (Feb 09 2017 at 19:11):
I probably need to build a few good rendered examples in order to help people understand what it does
Josh Mandel (Feb 09 2017 at 19:13):
One option is a boolean flag for including all linked data, but this may be extensive - up to an entire patient record
it's worse than that, because it also includes organizations, practitioners, etc...
Grahame Grieve (Feb 09 2017 at 19:15):
indeed.
Last updated: Apr 12 2022 at 19:14 UTC
