Stream: IG creation
Topic: Publisher history issue
Ted Klein (Jul 13 2020 at 23:03):
Latest version of the Publisher integrated with the ci Build (v1.1.2) is no longer processing history records linked through the Provenance bundles in UTG to NamingSystem resources - seems to be ok for CodeSystem resources. All NamingSystem resources now triggering errors in the form of: NamingSystem-AS4E.html#/html/body/div/div/div/div/div/ul/li/a at Line 143, column 4 error The link 'NamingSystem-AS4E.history.html' for "History" cannot be resolved
NamingSystem-AS4E.json.html#/html/body/div/div/div/div/div/ul/li/a at Line 144, column 4 error The link 'NamingSystem-AS4E.history.html' for "History" cannot be resolved
NamingSystem-AS4E.ttl.html#/html/body/div/div/div/div/div/ul/li/a at Line 144, column 4 error The link 'NamingSystem-AS4E.history.html' for "History" cannot be resolved
NamingSystem-AS4E.xml.html#/html/body/div/div/div/div/div/ul/li/a at Line 144, column 4 error The link 'NamingSystem-AS4E.history.html' for "History" cannot be resolved
Grahame Grieve (Jul 14 2020 at 00:14):
@Lloyd McKenzie this looks like a template issue to me
Lloyd McKenzie (Jul 14 2020 at 01:57):
@Jose Costa Teixeira can you investigate?
Jose Costa Teixeira (Jul 14 2020 at 08:38):
I will check this later today
Jose Costa Teixeira (Jul 14 2020 at 08:41):
Thanks @Ted Klein for pointing this is only on NamingSystem - that helps
Etienne Cantineau (Jul 14 2020 at 10:06):
@Jose Costa Teixeira I have the same error with an instance target. It also breaks the display of the tab menu (the one with narrative content/xml/json/ttl)
Jose Costa Teixeira (Jul 14 2020 at 10:45):
@Etienne Cantineau source?
Etienne Cantineau (Jul 14 2020 at 10:48):
@Jose Costa Teixeira https://build.fhir.org/ig/ec-ehealth/Build-source/Practitioner-practitioner1.html
Jose Costa Teixeira (Jul 14 2020 at 17:17):
Good news: I think these are the same issue
Jose Costa Teixeira (Jul 14 2020 at 17:17):
Good news #2: I think I already fixed something similar
Jose Costa Teixeira (Jul 14 2020 at 17:18):
Bad news: I really don't remember how ...
I'm checking and will get back when i have something
Jose Costa Teixeira (Jul 14 2020 at 17:36):
@Lloyd McKenzie PR created. Tested with Etienne's IG, it works. UTG is still building but I have no reason to think it won't
Rick Geimer (Jul 15 2020 at 04:21):
I'm having the same issue:
http://build.fhir.org/ig/HL7/ccda-on-fhir-r4/qa.html
ImplementationGuide-hl7.fhir.us.ccda.history.html#/html/body/div/div/div/div/div/ul/li/a at Line 123, column 6 error The link 'ImplementationGuide-hl7.fhir.us.ccda.html' for "Narrative Content" cannot be resolved
These ".history.html" files are showing up, but are only linked to from qa.html, and have bad links in each of their XML, JSON. and TTL tabs.
Jose Costa Teixeira (Jul 15 2020 at 06:08):
this is a different issue. the other was that history was not being generated. this is that history files contain bad links.
Jose Costa Teixeira (Jul 15 2020 at 06:25):
just checked and I need to add some debugging tags to the template, because this is harder to track.
Jose Costa Teixeira (Jul 15 2020 at 06:27):
I'm checking this. these pages are not supposed to exist, so we need to see why they are being generated.
Jose Costa Teixeira (Jul 15 2020 at 10:56):
@Etienne Cantineau does this work now?
Jose Costa Teixeira (Jul 15 2020 at 11:48):
@Ted Klein I ran into an Out of Memory error while building UTG. Do you want to see if it is fixed please?
Etienne Cantineau (Jul 15 2020 at 11:55):
@Jose Costa Teixeira No i still have the problem
Jose Costa Teixeira (Jul 15 2020 at 12:56):
@Etienne Cantineau i have it working ok.
image.png
One thing I noticed in your setup: your Provenance resource - is that supposed to be an example of Provenance, or is it supposed to convey history of changes to the example Practitioner1?
Etienne Cantineau (Jul 15 2020 at 13:04):
@Jose Costa Teixeira It's an example of Provenance, not an history element.
Jose Costa Teixeira (Jul 15 2020 at 13:10):
Ah ok. So that is one open issue. We need to do some changes, but your IG is correct. And I do not know if there is a workaround.
Jose Costa Teixeira (Jul 15 2020 at 13:10):
@Lloyd McKenzie here's one example of a Provenance instance that is not history
Jose Costa Teixeira (Jul 15 2020 at 13:18):
Grahame, I think we can only consider "history": true if the Provenance resource is effectively in a Bundle of type History, which is not this case.
Ted Klein (Jul 15 2020 at 19:02):
I have to run my ci build locally with max memory settings in the java command or it runs out of memory. I use: java -Xmx16G -Xss1G -jar /Users/tedklein/documents/tedhl7/vocabulary/projects/utg/FHIR_tool_stuff/org.hl7.fhir.publisher.jar -ig ig.ini for my commend; anything smaller than 16G memory or 1G stack and it blows up.
Lloyd McKenzie (Jul 15 2020 at 19:03):
So the issue is "how do we allow overriding the memory settings for a single IG without overriding them generally and still using the standard .bat files. @Jose Costa Teixeira thoughts?
Jose Costa Teixeira (Jul 15 2020 at 19:24):
Only this:
https://stackoverflow.com/questions/25608260/call-java-from-batch-file-how-to-split-jvm-and-main-method-parameters
have a jvm_args parameter in the batch file which is normally empty, but if the user does genonce.bat -jvm"XYZ" the batch would intercept the text between quotes and put this as options instead of parameters.
Jose Costa Teixeira (Jul 15 2020 at 19:26):
(hope this is not urgent because does not sound like easy or fun)
Jose Costa Teixeira (Jul 15 2020 at 20:02):
I have not encountered this need, but is this a common thing?
Lloyd McKenzie (Jul 15 2020 at 21:35):
Primarily just UTG, I think
Grahame Grieve (Jul 16 2020 at 05:21):
I think we can only consider "history": true if the Provenance resource is effectively in a Bundle of type History
well.. that wasn't a rule that we made. Why should it be a rule?
Lloyd McKenzie (Jul 16 2020 at 05:29):
I think the rule should be that it's history if the Provenance refers to one of the artifacts in the spec and isn't flagged by the IG as being an example. (We could have examples in the spec that also have example provenances.)
Jose Costa Teixeira (Jul 16 2020 at 06:03):
I think this could work. if it's about one of the artifacts in the spec AND is not example (exampleBoolean OR exampleCanonical == false).
Grahame Grieve (Jul 16 2020 at 06:12):
I think that's the rules I already use
Jose Costa Teixeira (Jul 16 2020 at 06:25):
Etienne's repo has exampleCanonical=true
https://github.com/ec-ehealth/Build-source/blob/master/input/ImplementationGuide-BeCoreImplementationGuide.json#L447
and the resources.json still says history=true for the provenance's target resource
Jose Costa Teixeira (Jul 16 2020 at 06:25):
should I create an issue?
Grahame Grieve (Jul 16 2020 at 06:51):
yes
Jose Costa Teixeira (Jul 16 2020 at 07:59):
https://github.com/HL7/fhir-ig-publisher/issues/134
(i went through the code to see if I could do something but I don't think I can, sorry)
Eric Haas (Jul 16 2020 at 22:17):
it looks like this in config.json:
109 "ext-history": "{{[type]}}-{{[id]}}.ext.history.html"
should be :
109 "ext-history": "{{[type]}}-{{[id]}}.profile.history.html"
I have not tried out the history option and all my igs have it false. how
do you turn it on?
Eric M Haas, DVM, MS
Health eData Inc
35 Crescent Avenue, Sausalito, CA 94965
Eric Haas (Jul 17 2020 at 02:52):
Jose Costa Teixeira (Jul 17 2020 at 06:09):
everything is now profile-history (no separate layouts for extension history)
Jose Costa Teixeira (Jul 17 2020 at 06:12):
@Eric Haas to enable history, you just have to have a history of changes. Add a bundle of Provenances that point to the resource for which you want to enable changes:
Jose Costa Teixeira (Jul 17 2020 at 06:13):
see the sample IG, input/history/h1.json, change that (or add a new one) with references to your profile/extension in the target:
https://github.com/FHIR/sample-ig/blob/master/input/history/h1.json#L11
Last updated: Apr 12 2022 at 19:14 UTC
