FHIR Chat · SUSHI 0.14.0 · shorthand

Stream: shorthand

Topic: SUSHI 0.14.0

Nick Freiter (Jul 07 2020 at 19:23):

Announcing SUSHI 0.14.0

SUSHI 0.14.0 contains the following enhancements and bug fixes:

• Deprecates usage of , when listing items
• Deprecates usage of | when listing choices of items
• Adds insert rules
• Adds support for specifying more complex dependencies in config.yaml
• Adds optional include keyword for Value Set rules
• Deprecates usage of the optional units keyword
• Adds support for inline resources in the contained array of a StructureDefinition
• Adds support for escape characters in strings
• Fixes bug allowing choice types to be constrained to specializations of specific choices
• Fixes bug which causes extra extensions to appear on Instances
• Fixes bug with checking the type of a Reference
• Fixes bug with references to inline resources
• Aligns _genonce, _gencontinuous and _updatePublisher scripts with those in the sample-ig

For more details, see the full release notes: https://github.com/FHIR/sushi/releases/tag/v0.14.0

Chris Moesel (Jul 07 2020 at 19:31):

For those who don't read the detailed release notes, note that this release begins to align with the upcoming FSH 1.0 and contains some syntax changes that came out of the ballot. So... expect some warnings.

David Hay (Jul 07 2020 at 21:16):

and to show I do read the detailed notes - they have a typo: 'lsiting'. :)

May Terry (Jul 07 2020 at 21:18):

David Hay said:

and to show I do read the detailed notes - they have a typo: 'lsiting'. :slight_smile:

And to show that I could be better about reading release notes, I missed the part that deprecating some delimiters will not error out but give passive warnings instead. ;-)

Shovan Roy (Jul 08 2020 at 01:04):

I am not sure if anyone else is experiencing the same. I took version 0.14.0 and applied the necessary adjustments as per the new rules and was able to run sushi successfully without any warning.
However '_updatePublisher.bat' is not working. Each time I'm trying to execute it, I am getting the following error: "Exception calling "DownloadFile" with "2" argument(s): "The request was aborted: Could not create SSL/TLS secure
channel. ..." and giving me the option :
Updating scripts
Update scripts? (Y/N).

If I select 'Y' then all the scripts are getting updated but when I am trying to run the '_updatePublisher.bat' post the update, I'm getting the same error. did I miss any step?

Grahame Grieve (Jul 08 2020 at 01:05):

that's nothing to do with sushi and means that you need to update the batch file

Shovan Roy (Jul 08 2020 at 01:09):

thanks @Grahame Grieve since the batch files are getting generated as a part of 'sushi .' command, I thought sushi is already applying the necessary update on the batch files before it gets generated. I'll have a look now on what to update on the batch files.

Chris Moesel (Jul 08 2020 at 02:41):

Yeah, SUSHI provides the batch files for convenience. @Shovan Roy -- we did update the batch file in the 0.14.0 release, but we just used the exact batch file that is in the sample-ig project. We didn't change anything. One of our devs with a Windows system tested it and it worked for him... @Jose Costa Teixeira -- any idea what might be happening here?

Chris Moesel (Jul 08 2020 at 02:44):

Until that gets sorted out, you can manually download the publisher jar here: https://github.com/HL7/fhir-ig-publisher/releases/latest/download/publisher.jar

If you copy it into build/input-cache (assuming you have SUSHI output to a build folder), then you _should_ be able to run _genonce.bat after that (I think).

Shovan Roy (Jul 08 2020 at 03:08):

thanks @Chris Moesel I downloaded the latest 'publisher.jar' and kept it in the input-cache folder and then run _genonce.bat file. Everything went well,
fyi..in qa.html, I see the following:
image.png

Vassil Peytchev (Jul 08 2020 at 03:10):

Seems that this is a known issue, and it is being worked on

Grahame Grieve (Jul 08 2020 at 03:37):

I mean, it 's my fault, because I changed the release process. This is the second time in as many months but I promise it's all stable now

Jose Costa Teixeira (Jul 08 2020 at 04:24):

Chris Moesel said:

Yeah, SUSHI provides the batch files for convenience. Shovan Roy -- we did update the batch file in the 0.14.0 release, but we just used the exact batch file that is in the sample-ig project. We didn't change anything. One of our devs with a Windows system tested it and it worked for him... Jose Costa Teixeira -- any idea what might be happening here?

the batch files were updated to accommodate for the jar file name change.
In this case, to make sure you have the latest stuff, the formula is :

1. run the _updatePublisher (just to make sure it updates itself, you can say No to updating the jar) and
2. run _updatePublisher again to get it to download the right publisher jar.

Jose Costa Teixeira (Jul 08 2020 at 04:26):

minor annoyance with the language updates: It seems I can no longer do

• identifier, subject, status MS

Jean Duteau (Jul 08 2020 at 04:32):

Jose Costa Teixeira said:

minor annoyance with the language updates: It seems I can no longer do

• identifier, subject, status MS

Yeah, that was the result of the first change listed above:
Deprecates usage of , when listing items

Nick Freiter (Jul 08 2020 at 12:32):

@Jose Costa Teixeira you should be able to do

* identifier and subject and status MS

For now the , will still work (but with a warning), but in a future version, the , will be altogether replaced by and. We received feedback on the FHIR Shorthand ballot that it was confusing for us to sometimes use , and sometimes use and when listing items, so the FHIR Shorthand spec is being changed to only use and to list items. So this change in SUSHI is to keep us aligned with the FSH spec.

Jose Costa Teixeira (Jul 08 2020 at 12:34):

Sorry, i did not mean to indicate I'm stuck - it's just a minor annoyance - not with sushi, but the language change.
It's really minor, but it's an annoyance :)

Chris Moesel (Jul 08 2020 at 12:46):

One of the costs of being an early adopter, I guess. Sometimes being hip comes with a price. ;-)

Jose Costa Teixeira (Jul 08 2020 at 17:43):

true (except the hip part) - I did not mean this as a suggestion to change - annoyance for one is better than ambiguity for many. :)

Shovan Roy (Jul 09 2020 at 03:43):

Hi All,

I've drafted profile on R4 Consent, here is my code:

• status and dateTime MS
• dateTime 1..1
• patient only Reference(AUBasePatient)
• provision.actor.reference only Reference(Device or AUBasePatient or AUBaseOrganisation or AUBasePractitionerRole or AUBasePractitioner)

I haven't added any example as yet. However, after running the publisher, I can see the following error reported, any suggestion?
Path: StructureDefinition.snapshot.element[9].example[0].value.ofType(Identifier)
Message: if identifier.system is ''urn:ietf:rfc:3986'', then the identifier.value must be a full URI (e.g. start with a scheme)

Lloyd McKenzie (Jul 09 2020 at 05:10):

Do you have an exampleIdentifier with the specified system and a value that isn't a full URI?

Shovan Roy (Jul 09 2020 at 06:08):

@Lloyd McKenzie , No, I haven't added any example in identifier in my profile. But I can see the generated resource has the following example:
General:<valueIdentifier xmlns="http://hl7.org/fhir">
<system value="urn:ietf:rfc:3986"/>
<value value="Local eCMS identifier"/>
</valueIdentifier>

Trying to find out where it is coming from

Igor Sirkovich (Jul 09 2020 at 07:18):

@Shovan Roy, do you derive your profile from AU Base? Possibly, this error comes from the underlying AU Base profile.

Etienne Cantineau (Jul 09 2020 at 08:19):

Hello, i have problems with the inline reference correction. if I use Reference(OrganizationEx01), instead of having Organization/OrganizationEx01, now i get #OrganizationEx01 which is not resolvable by the publisher

Shovan Roy (Jul 09 2020 at 08:48):

Igor Sirkovich said:

Shovan Roy, do you derive your profile from AU Base? Possibly, this error comes from the underlying AU Base profile.

Hi @Igor Sirkovich that's what I thought initially, I'm using AU Base but for Consent, my code is extending the HL7 R4 "Consent". there is no AU Consent ( at least I couldn't find any). I looked a the sushi code (I'm not expert here) I coudldn't find anything there as well.

Lloyd McKenzie (Jul 09 2020 at 13:51):

@Shovan Roy You'll have to check each snapshot in the derivation hierarchy until you get to the base standard - it's coming from somewhere...

Shovan Roy (Jul 10 2020 at 00:11):

@Lloyd McKenzie in my profile, I am extending the HL7 standard Consent Profile. To make it simple, I removed all the AU related references fromthe sushi code. The new Sushi file looks like the followings:

• status and dateTime MS

I don't see any anything related to the identifier example in the sushi generated StructureDefinition. However, the publisher StructureDefinition-ncdhc-consent.json generated snapshot has "Example General: {"system":"urn:ietf:rfc:3986","value":"Local eCMS identifier"}". I have attached the SD generated by sushi for your reference.
@Grahame Grieve , just wondering if this is something IG publisher is injecting while generating the snapshot.

Problem Summary: IG publisher is showing the following error while generated Consent profile derived from standard R4 Consent. image.png

Grahame Grieve (Jul 10 2020 at 00:27):

I don't inject anything like that - it has to come from one of the profiles

Shovan Roy (Jul 10 2020 at 00:49):

thanks @Grahame Grieve . you are right.. I was able to trace this in the core R4 Consent profile.

Lloyd McKenzie (Jul 10 2020 at 01:14):

@David Pyke Can you fix this in core? ("Local eCMS identifier" isn't a super-intuitive name anyhow - and the system should be changed to an example.org base rather than one that presumes the value will be a URL.

David Pyke (Jul 10 2020 at 01:31):

Yep, I'll fix it and push it in.

David Pyke (Jul 10 2020 at 01:39):

The only place that exists is in the Consent.identifier example which is the following URL: http://example.org/Local/eCMS/identifier Not sure how it's being dragged into his code.

David Pyke (Jul 10 2020 at 01:39):

I'm removing it anyway but I"m not sure how that is breaking things.

David Pyke (Jul 10 2020 at 01:40):

But that's a holdover from the original Consent resource so it's been broken a long time

Lloyd McKenzie (Jul 10 2020 at 02:15):

If it's in R4, may want to make a change request for that for consideration if we ever do another R4 patch

Shovan Roy (Jul 10 2020 at 02:27):

I have faced it in R4, happy to raise a CR.

Lloyd McKenzie (Jul 10 2020 at 03:13):

Please do @Shovan Roy

Shovan Roy (Jul 10 2020 at 04:14):

CR raised: https://jira.hl7.org/browse/FHIR-27940

Last updated: Apr 12 2022 at 19:14 UTC