Stream: Cancer Interoperability
Topic: Breast Radiology IG
May Terry (Jul 25 2019 at 19:34):
Topic for enhancements to get the Breast Radiology IG ready for ballot. cc: @Mark Kramer , @Kurt Allen , @Richard Esmond
Kurt Allen (Jul 25 2019 at 20:00):
Hey @May Terry and @Mark Kramer I sent out git invitations to both of you. Let me know if you didnt get them...
Kurt A
May Terry (Jul 25 2019 at 20:02):
I received the email. thanks @Kurt Allen ! qq... looks like you have 2 config versions, config.v3.json and config.v4.json. To confirm, the one which generated the IG on build.fhir.org is config.v3.json (which is to STU3). were there any gotchas on your config.v4.json (which specifies FHIR_R4)? i.e...were you able to get the R4 version compiling with shr-cli?
Kurt Allen (Jul 25 2019 at 20:10):
Hey @May Terry - V4 worked except for logical models which are only supported in the shr-cli V3. We felt that logical models gave a more concise and simple display of the structure, for people who don't live for for fhir'isms and structure definition details.
We are currently using V3.
Mark Kramer (Jul 25 2019 at 20:14):
That might be something we can get going in the next few days. Maybe....
Kurt Allen (Jul 25 2019 at 21:40):
Hello @May Terry and @Mark Kramer I talked with Richard and he is good for meeting at 9:30 EST / 10:30 CST . I suggest we use the cancer goto meeting for the venue.
https://app.gotomeeting.com/?meetingId=116193445
I assume we are starting on Monday?
'ttfn
Kurt A.
Mark Kramer (Jul 25 2019 at 21:41):
:ok:
May Terry (Jul 29 2019 at 14:55):
@Kurt Allen - to have value set local codes show up in the IG, within the CIMPL value set files, just remove the reference to CodeSystem reference each of the value set terms. e.g.: in the AbnormalityDensityVs.txt file, change the line from TEST#EqualDensity "EqualDensity" to #EqualDensity "EqualDensity. Then remove the markdown in the description so it looks cleaner. The output will look like this... Screen-Shot-2019-07-29-at-10.54.51-AM.png . cc: @Mark Kramer
May Terry (Jul 29 2019 at 15:01):
Note that that in the screenshot, the displayed Defining URL and system url is defaulting to the ProjectURL parameter that you specified in the CIMPL config file (http://build.fhir.org/ig/HL7/fhir-breast-radiology-ig). This will also update when you change the URL to the production link (fhir.org...).
Kurt Allen (Jul 29 2019 at 17:13):
@May Terry Thanks, made the changes and pushed it to the fhir build. Looks good, also got the graphic on the landing page to show up properly. Working on the Observation component codes to make them 'local' codes as well.
cc:@Mark Kramer @Richard Esmond
Mark Kramer (Jul 29 2019 at 17:18):
I am bringing you up to date with latest OBF. (On my own branch)
Mark Kramer (Jul 29 2019 at 17:28):
@Kurt W. Allen Question on binding strength -- do you intend your bindings to be required? I see things like this: Value: concept from BreastDensityVS -- this means the code MUST come from this VS, period. No extensibility. Intended?
May Terry (Jul 29 2019 at 17:31):
Good point @Mark Kramer - I suspect @Kurt Allen and @Richard Esmond might want to make some extensible and it will vary by value set. I suppose we should create a content profile file for them as a starter since I don't see it in their current build.
Kurt Allen (Jul 29 2019 at 22:48):
Hey @Mark Kramer Richard wanted the binding strength to be be required (no extensibility). We can talk about that decision tomorrow morning. Let me know what you want me to merge your stuff. You can do a pull request, or just let me know and I will merge your branch. Whatever is easiest for you.
Kurt Allen (Jul 30 2019 at 16:18):
@Mark Kramer - I executed your pull request. There were a lot of merge conflicts, so you will want to pull the developmentmk branch before you continue work (I didn't delete that branch after he merge 'cause I assumed you wanted to do future work there..)
Mark Kramer (Jul 30 2019 at 17:04):
(deleted)
Kurt Allen (Jul 30 2019 at 20:47):
@Mark Kramer I just got an e-mail from Lynn Laakso indicting that our hl7-logo.png graphic was out of date. I tried adding their (attached) new logo by putting it in a CIMPL\Pages\assets\images folder but the 'old' one was still in the build output. Is this something that you package into your build tools?
Mark Kramer (Jul 30 2019 at 20:48):
Ours is old, too.
Mark Kramer (Jul 30 2019 at 20:49):
I will have to locate where it comes from and replace it
Mark Kramer (Jul 30 2019 at 20:49):
I will let you know
Mark Kramer (Jul 30 2019 at 21:09):
@Kurt Allen replace the old logo located here: shr-cli\node_modules\shr-fhir-export\lib\ig_files\pages\assets\images\hl7-logo.png. The new one is http://hl7.org/fhir/R4/assets/images/hl7-logo.png (smaller scale than the one you uploaded)
Kurt Allen (Jul 30 2019 at 22:06):
Done and pushed. Thx,
Mark Kramer (Jul 31 2019 at 15:08):
@Kurt Allen @Richard Esmond The model doc and logical model are generating a huge number of publication errors that I can't fix. How would you feel about going to R4 and dropping the model doc and logical models? Almost all the IGs are coming out under R4 now.
Kurt Allen (Jul 31 2019 at 15:24):
@Mark Kramer and @Richard Esmond ;
I like the logical model, but I think it is more important to get rid of the errors. Not much use if have the logical model if it cant get published because of too many errors :-) (i.e. half a loaf etc)
Mark Kramer (Jul 31 2019 at 15:26):
Can you confirm this with Richard?
Richard Esmond (Jul 31 2019 at 18:33):
That would be fine by me.
Kurt Allen (Jul 31 2019 at 18:41):
@Mark Kramer @Richard Esmond I will push a r4 version. I added batch files to create r4 locally (git latest development checkin)
Kurt Allen (Jul 31 2019 at 19:07):
@Mark Kramer
There is a value set defined (Fhir-BreastRadiologyReport\CIMPL\breastrad\ValueSets\BreastAbnormalityCodesVS.txt) and used that does not appear to generate either a Fhir CodeSystem or a Fhir ValueSet record. This appears to be causing all but one of the remaining errors. I dont know what is different about this value set definition...
Kurt Allen (Jul 31 2019 at 19:59):
@Chris Moesel and @Mark Kramer
I modified the code so that BreastAbnormalityCodesVS is now output.
I added the section below, which has the 'Code from BreastAbnormalityCodesVS' line it it, and used that as the parent where I had previously used CodedObservationComponent.
*//ConceptComponent-Entry
Group: BreastCodedObservationComponent
Parent: CodedObservationComponent
Code from BreastAbnormalityCodesVS
DataValue only concept*
*What I expect is going on is that there is code in shr-cli to only output a value set if it appears top have been used. Adding it to a
*CodeSystem: BA = http://hl7.org/fhir/us/breast-radiology/breastrad/vs/breastabnormalitycodesvs*
line doesnt trigger that, nor does using it as
Code = BA#AbnormalityType
but placing it in a Code from BreastAbnormalityCodesVS does mark it as being needed and flags it for output to fhir.
Bug?
Kurt A.
Kurt Allen (Jul 31 2019 at 20:51):
Hey @Mark Kramer
Looks like we are down to two errors, and I think they are in CIMPL base classes.
C:\Development\MITRE\cimpl\Fhir-BreastRadiologyReport\out.v4\fhir\guide\resources\structuredefinition-breastrad-BreastRadiologyDocument.json
Path Severity Message
StructureDefinition/breastrad-BreastRadiologyDocument: StructureDefinition.mapping[3].uri error URL value 'http://hl7.org/fhir/documentreference' does not resolve
and
C:\Development\MITRE\cimpl\Fhir-BreastRadiologyReport\out.v4\fhir\guide\resources\valueset-obf-StudyArmTypeVS.json
Path Severity Message
obf-StudyArmTypeVS error Error from server: Unable to provide support for code system https://ncit.nci.nih.gov
Mark Kramer (Aug 01 2019 at 19:03):
@Kurt Allen I n the PR that I issued earlier today, there are two errors that are probably easy to fix. One is a regression that I caused. Can you address these?
package-list.json: canonical is wrong - is 'http://hl7.org/fhir/us/breast-radiology' should be 'http://hl7.org/fhir/us/breast-radiology/fhir'
package-list.json: No entry found for version 1.0.0
May Terry (Aug 02 2019 at 13:20):
@Kurt Allen and @Mark Kramer - FYI: I pulled the development branch and tried to publish the IG. It fails midway with the following error: Exception in thread "main" java.lang.Error: Case mismatch of file /Users/mayt/cimpl/shr-cli-6.5.0/brr/fhir/guide/qa/ValueSet-breastrad-AbnormalityDensityVS-valdiation.html: found ValueSet-breastrad-AbnormalityDensityVs-valdiation.html. This happens running config.v3.json and config.v4.json. Apparently, there is a misspelling with "valdiation".
May Terry (Aug 02 2019 at 13:40):
Notes from this morning's sync-up. The error is likely more related to the mixed case of "VS" between AbnormalityDensityVS and AbnormalityDensityVs. This error doesn't appear on Windows systems (potentially because Windows is looser on checking case sensitivity) however this is checked on MacOS systems. Kurt to look into resolving the case discrepancy to support both platforms.
Mark Kramer (Aug 02 2019 at 14:08):
@Kurt Allen are you using shr-cli-6.5.0?
Mark Kramer (Aug 02 2019 at 14:25):
@Kurt Allen All the header and footer errors are from the reference model pages. When modeldoc="false" all those errors should disappear. No code changes will be required.
Mark Kramer (Aug 02 2019 at 15:01):
Down to three build errors when includeModelDoc = false. All three are broken links on the index page:
index.html#/html/head/link/body/div/div/div/div/div/ul/li/a at Line 152, column 72 error The link 'modeldoc/index.html' for "reference model" cannot be resolved
index.html#/html/head/link/body/div/div/div/div/div/ul/li/a at Line 152, column 740 error The link 'StructureDefinition-cimi-statement-ConditionPresenceStatement.html' for "ConditionPresenceStatement profile" cannot be resolved
index.html#/html/head/link/body/div/div/div/div/div/ol/li/a at Line 184, column 6 error The link 'modeldoc/index.html' for "Reference Model" cannot be resolved
Kurt Allen (Aug 02 2019 at 15:15):
Yep, version 6.5.
Kurt A.
Mark Kramer (Aug 02 2019 at 15:29):
I can't get to the bottom of the documentreference error (URL value 'http://hl7.org/fhir/documentreference' does not resolve). The link is in the StructureDefinition, and it is not the correct URL for document reference (http://hl7.org/fhir/StructureDefinition/DocumentReference). But where this is coming from, I don't know. I've seen this before. @May Terry do you remember this error?
May Terry (Aug 02 2019 at 15:37):
@Mark Kramer - yep! I get this error in my local builds of the IG in the obf-Composition resource. There is an equivalent one pointing to Composition not resolving on obf-DocumentReference. The error goes seems to disappear when you run it through the noprofilebuild.
Kurt Allen (Aug 02 2019 at 15:44):
@Mark Kramer @May Terry
I pushed new ig with modeldoc off, and examples on.
All is merged into development branch.
Kurt Allen (Aug 02 2019 at 15:47):
@Mark Kramer
I created a new branch 'SpliceErrors' and pushed it to the Applicadia github. THis branch has the changes you suggested this morning, and the build fails. The fail appears in the 'Finding Left Breast' and 'Finding Right Breast' sections, which are nested under 'Findings'. To see the error, run the R4.BuildProfiles.bat batch file.
Kurt Allen (Aug 02 2019 at 15:51):
@Mark Kramer
I just pushed another mod to 'SpliceErrors'. The last one had a failed experiment and was not exactly what you had suggested. New branch is (I think) what you suggested.
Mark Kramer (Aug 02 2019 at 16:01):
What's a splice? I'm lost.
Kurt Allen (Aug 02 2019 at 16:14):
Sorry, Splice -> Slice. I guess my fingers have a lisp.
Prob. appropriate for a programmer...
Kurt Allen (Aug 02 2019 at 16:23):
@May Terry @Mark Kramer
I have corrected all the {Name}Vs.txt errors in breastrad. Should now be {Name}VS.txt
May Terry (Aug 02 2019 at 16:47):
@Kurt Allen - I just did a pull of the latest development branch and the IG Publisher still fails with the same error. I'm using shr-cli 6.5.0 on a MacOS.
Kurt Allen (Aug 02 2019 at 17:34):
I don't know what might be the problem... Mark may have thoughts during our meeting..
Grahame Grieve (Aug 02 2019 at 21:51):
a couple of comments:
- the IG publisher checks for case consistency whereever it can, since the build happens on linux, which requires the case be correct.
- I don't have a clue about the documentreference thing - is that sorted? you can ask me to investigate these things - I don't do much but support the IG process in the week before publishing deadlines
Mark Kramer (Aug 03 2019 at 12:11):
I just looked at http://build.fhir.org/ig/HL7/fhir-breast-radiology-ig/ and there are problems with links on the home page. Whole paragraphs have been turned into links. There must be one or more missing end-anchor tag. This should be fixed.
Alex Goel (Sep 14 2019 at 14:37):
https://github.com/goelalex/SDC-on-FHIR
Last updated: Apr 12 2022 at 19:14 UTC
