Stream: IG creation
Topic: diff tool to compare FHIR IGs
Richard Stanley (Mar 16 2022 at 17:30):
I'm curious if there is a diff-like tool to compare FHIR IGs or bundles of differently profiled resources. Thanks.
Lloyd McKenzie (Mar 16 2022 at 18:19):
@Grahame Grieve
Gino Canessa (Mar 16 2022 at 18:20):
As a note, I am working on something for this (comparing definitions, not pages if that is relevant). It is still a bit away from usable.
Eric Haas (Mar 16 2022 at 19:20):
mitre/mCode team did some nice work here and hoping they open source it. @Max Masnick ?
Jose Costa Teixeira (Mar 16 2022 at 19:23):
if it's 1:1 comparison - would this help?
https://confluence.hl7.org/display/FHIR/Using+the+FHIR+Validator#UsingtheFHIRValidator-ComparingProfiles
Richard Stanley (Mar 16 2022 at 19:37):
@Jose Costa Teixeira yes I think so, for conformance resources. I’ll give it a go!
Max Masnick (Mar 18 2022 at 16:42):
@Richard Stanley @Eric Haas we are still working on this -- hoping to have it out in the April/May timeframe
Eric Haas (Apr 05 2022 at 15:48):
I proposed in a tracker to compare all the observation in US Core in a table. before writing a script is there a tool that can do this already. ... not just a head to head comparison ( been there and done that and waiting for the MCode version ). but across all profiles of a given type? I am imagining columns would be root level elements, rows profiles and then pack in cells the MS cardinality, value/pattern/binding/invariant/references/types if applicable.
for example.
Observation
| Profile | status | code | ... |
|---|---|---|---|
| US Core Observation Foo | MS 1..1, Observation status Value set | MS 0..1, fixedValue = foo | ... |
| US Core Observation Bar | MS 1..1, Observation status Value set | MS 0..1, fixedValue = bar | ... |
| US Core Observation Baz | MS 1..1, Observation status Value set | MS 0..1, fixedValue = baz | ... |
| ... | ... | ... | ... |
Grahame Grieve (Apr 05 2022 at 22:02):
there is an observation summary already
Grahame Grieve (Apr 05 2022 at 22:04):
check out summary-observations.xhtml in the fragments. It's a work in progress, so we can talk about it
Eric Haas (Apr 05 2022 at 23:10):
Yes that looks good is more succinct and it gives a good overview: here are the columns I'd like to to see:
Columns
profile name
category Slice
code Binding
MS Time Types
MS Value Types
Is DataAbsentReason MS?
Is Component MS ?
Component code Binding
Component MS Value Types
hasMember references
derivedFrom references
Eric Haas (Apr 05 2022 at 23:11):
would need to recurse through all vitals and survey profiles
Eric Haas (Apr 05 2022 at 23:12):
this is specific to US core but could generalize to Observation adding status, binding, subject references etc.
Grahame Grieve (Apr 05 2022 at 23:48):
can't get all of them in there - too many columns
Grahame Grieve (Apr 05 2022 at 23:49):
but I could collapse timetypes to 1-3 icons. And value types too.
Grahame Grieve (Apr 05 2022 at 23:54):
I have icons from questionnaire rendering for
- boolean
- CodeableConcept
- decimal
- quantity
- string
- time
- attachment
- integer
so need 16x16 icons for Range, Ratio, Period, SampledData
Grahame Grieve (Apr 05 2022 at 23:54):
and I have an icon for Must support
Grahame Grieve (Apr 05 2022 at 23:55):
we really don't care about derived from. Members + components, yes
Eric Haas (Apr 07 2022 at 08:10):
I am not worried about the table size since I was thinking of a downloadable csv or excel file not an html table since they are as you point out limited in content.
Grahame Grieve (Apr 07 2022 at 11:28):
I could generate that too
Eric Haas (Apr 07 2022 at 14:11):
I don't see any value in using icons in a csv or excel or limiting the number of columns.
Grahame Grieve (Apr 07 2022 at 14:27):
no not in that context
Last updated: Apr 12 2022 at 19:14 UTC
