FHIR Chat · IG History page links cannot be resolved

Stream: IG creation

Topic: IG History page links cannot be resolved

David deRoode (Jul 16 2020 at 22:32):

IG began getting these four build errors regarding the History pages (xml, json, ttyl, Narrative Content), any thoughts on what could be causing these?

ImplementationGuide-hl7.fhir.us.dentalreferral.history.html#/html/body/div/div/div/div/div/ul/li/a at Line 123, column 6|The link 'ImplementationGuide-hl7.fhir.us.dentalreferral.html' for "Narrative Content" cannot be resolved
ImplementationGuide-hl7.fhir.us.dentalreferral.history.html#/html/body/div/div/div/div/div/ul/li/a at Line 129, column 8|The link 'ImplementationGuide-hl7.fhir.us.dentalreferral.xml.html' for "XML" cannot be resolved
ImplementationGuide-hl7.fhir.us.dentalreferral.history.html#/html/body/div/div/div/div/div/ul/li/a at Line 136, column 8|The link 'ImplementationGuide-hl7.fhir.us.dentalreferral.json.html' for "JSON" cannot be resolved
ImplementationGuide-hl7.fhir.us.dentalreferral.history.html#/html/body/div/div/div/div/div/ul/li/a at Line 143, column 8|The link 'ImplementationGuide-hl7.fhir.us.dentalreferral.ttl.html' for "TTL" cannot be resolved

Eric Haas (Jul 17 2020 at 00:43):

They just popped for me as well...

Eric Haas (Jul 17 2020 at 00:57):

looks like a template error.. I'll check it out

Eric Haas (Jul 17 2020 at 02:25):

This PR hides the IG resource's HX file but does not fix the extension which are still causing build errors. The changes to the config.json file to basically consolidate the extension hx into the profile hx template may only be implemented by the ig-publisher tooling or maybe script? I don't know... can you check it out @Lloyd and or @Grahame

Eric Haas (Jul 17 2020 at 02:28):

Example of build errors for DEQM: https://build.fhir.org/ig/HL7/davinci-deqm/qa.html

Jose Costa Teixeira (Jul 17 2020 at 06:01):

When I build this locally with the base template (including latest changes) I don't get those errors

David Hay (Jul 17 2020 at 06:28):

I'm getting a lot of these errors as well - there seem to be history fragments generated for all extensions - and they don't have any entries so they all get these errors...

using the latest publisher...

Any way to turn the history off? Or do there need to be provenance entries for each extension...

Jose Costa Teixeira (Jul 17 2020 at 06:34):

can you point to the errors you're getting, @David Hay ?

Jose Costa Teixeira (Jul 17 2020 at 06:34):

you do not need to do anything - if something fails it's the template that is not OK

David Hay (Jul 17 2020 at 07:01):

here you go: http://build.fhir.org/ig/HL7NZ/nhi/branches/master/index.html

David Hay (Jul 17 2020 at 07:03):

Thanks for looking at this!

Jose Costa Teixeira (Jul 17 2020 at 08:04):

in a few hours I'll check this. thanks

John Moehrke (Jul 17 2020 at 11:24):

@Jose Costa Teixeira I get those too with my IHE template using IG.

John Moehrke (Jul 17 2020 at 11:25):

I figure I must have accidentally turned on something for history but not fully turned it on. I don't know what I did wrong, don't know what I should have done. https://github.com/IHE/FormatCode

David Pyke (Jul 17 2020 at 12:23):

I"m getting it to, using a derivative of fhir.base.template.

Jose Costa Teixeira (Jul 17 2020 at 13:52):

All IGs will suffer from the same. I think i found the issue

Jose Costa Teixeira (Jul 17 2020 at 13:53):

It's in the base template so all templates will suffer from the same

John Moehrke (Jul 17 2020 at 14:07):

suffering in a group is far less painful than suffering alone. I really appreciate your expertise, heal us doctor.

John Moehrke (Jul 17 2020 at 14:09):

what Medication will you prescribe (MedicationRequest)? -- a little humor for a friday --

Jose Costa Teixeira (Jul 17 2020 at 14:21):

i'd tried to describe that but Dosage does not have enough fields to describe that.

Jose Costa Teixeira (Jul 17 2020 at 14:23):

John Moehrke said:

suffering in a group is far less painful than suffering alone

I know, that is why I let the group suffer :smiling_devil:

Jose Costa Teixeira (Jul 17 2020 at 14:25):

I'm doing the "therapeutic proof", not sure the term in English. making a fix to see if i got the diagnostic right

David Pyke (Jul 17 2020 at 14:40):

I'll be ready for the Procedure to address the Condition

John Moehrke (Jul 17 2020 at 14:47):

Jose Costa Teixeira said:

I'm doing the "therapeutic proof", not sure the term in English. making a fix to see if i got the diagnostic right

is it a double -blind study?

Jose Costa Teixeira (Jul 17 2020 at 14:49):

not really

John Moehrke (Jul 17 2020 at 14:50):

well include @David Pyke and I and you will have two blind (ignorant) guys helping out

Jose Costa Teixeira (Jul 17 2020 at 14:50):

if you are looking for the differential:
I think the nav-tabs is causing an issue because xxx.history.html is the name of the fragment AND the name of the output file

Jose Costa Teixeira (Jul 17 2020 at 14:51):

the nav-tabs was introduced by @Eric Haas I think, and he started the debugging, so I am proceeding and seeing what is happening based on his issue report

Jose Costa Teixeira (Jul 17 2020 at 14:54):

i'm changing the config json to use xxx.base-history and not xxx.history

Jose Costa Teixeira (Jul 17 2020 at 15:08):

I'm definitely not sure this s the issue, but it allows me to look for the issue

Eric Haas (Jul 17 2020 at 15:10):

@Jose did you look at my PR: https://github.com/HL7/ig-template-base/pull/110

Jose Costa Teixeira (Jul 17 2020 at 15:10):

yes

Jose Costa Teixeira (Jul 17 2020 at 15:11):

i am now tracing back my steps to see which step is the one that fixed the issue (if I fixed it)

Eric Haas (Jul 17 2020 at 15:16):

history frags are being automatically generated for all the artificacts by default and the only issue remaining with the tabs is that extensions are creating using the wrong template and I don't know why.

Jose Costa Teixeira (Jul 17 2020 at 15:19):

yes, but the fragment has the same name as the final file - i don't know if that breaks it, but it sure doesn't help debugging. in a few minutes I shold be able to diff the things I have (i got it to 0 errors in NZ profile)

Jose Costa Teixeira (Jul 17 2020 at 15:19):

also there was a missing line in the config.json

Jose Costa Teixeira (Jul 17 2020 at 15:20):

  "template-format": "",

inside StructureDefinition (the profile, not the extension). I think this gave me a few issues with the xxx-.json.html and xxx.xml.html...

David Pyke (Jul 17 2020 at 15:26):

I just did a build and the implementation guide history error went away but the extension structure def history error remains

John Moehrke (Jul 17 2020 at 15:27):

certainly see the history on conformance resource generated pages where there are Provenance. But isn't there still a need for history of changes at the IG level? If not, then I am fine with that going away.

John Moehrke (Jul 17 2020 at 15:29):

or has history of changes at the IG level moved to a Provenance based method?

Eric Haas (Jul 17 2020 at 15:43):

I think the templates are correct and the generated fragments all exist. I'm not sure about the magic behind matching what is in the config.json to the proper template and that is where I think the issue is. @Jose Costa Teixeira @Grahame Grieve

Jose Costa Teixeira (Jul 17 2020 at 15:45):

i have a config.json that fixed my issue, but it also includes some other changes, i will try to pinpoint which one helps.

Jose Costa Teixeira (Jul 17 2020 at 15:46):

@Eric Haas I think your issue in the PR is because of the missing line in the config.json

Jose Costa Teixeira (Jul 17 2020 at 15:48):

without that line, it still tries to create another template-format.html thing, but it shouldn't because it already creates the xml and the json explicitly

Jose Costa Teixeira (Jul 17 2020 at 15:48):

this is the change:
image.png

Eric Haas (Jul 17 2020 at 15:49):

thats odd because that is a new line and does not involve the templates for extensions

Jose Costa Teixeira (Jul 17 2020 at 15:51):

that is what I am trying to see if my changes help fix this or at least understand.

Jose Costa Teixeira (Jul 17 2020 at 15:52):

want to check a new config.json? It includes some other stuff but then we can trim down

Jose Costa Teixeira (Jul 17 2020 at 15:52):

or @John Moehrke @David Pyke does the offer for help debugging still stand? i'd appreciate a confirmation that this may be a step forward in fixing this

Jose Costa Teixeira (Jul 17 2020 at 15:54):

config.json

Jose Costa Teixeira (Jul 17 2020 at 15:54):

this is what I have now

Eric Haas (Jul 17 2020 at 15:58):

@Jose chrome blocked it I'll just add that line and try it now.

Jose Costa Teixeira (Jul 17 2020 at 15:59):

it's several lines, but yes please start with that

Jose Costa Teixeira (Jul 17 2020 at 15:59):

in chrome you can sill get the file:

Jose Costa Teixeira (Jul 17 2020 at 16:00):

instead of clicking on the file on the bottom left notification, on the right part of that notification there is a "show All" (or CTRL +J)

Jose Costa Teixeira (Jul 17 2020 at 16:00):

there you have "keep unsafe file"

Jose Costa Teixeira (Jul 17 2020 at 16:04):

but I am just realizing that this may not be it - i think this config did not have the lines that create the other history pages (Which is where the issue was )

John Moehrke (Jul 17 2020 at 16:14):

I will note that the problem has gone away... so something magical has happened.

Jose Costa Teixeira (Jul 17 2020 at 16:16):

which is actually more scary

Jose Costa Teixeira (Jul 17 2020 at 16:16):

because I don't see any changes in the code since 12 hours ago

John Moehrke (Jul 17 2020 at 16:21):

yup, I can't confirm that it was failing 12 hours ago

Eric Haas (Jul 17 2020 at 16:25):

adding in the line

"template-format": "", to the config.json made all the hx page tab links for profiles break as well so will remove from both SD and SD:extension and see..

Eric Haas (Jul 17 2020 at 16:31):

@Jose Costa Teixeira removing that line from both SD and SD:Extension in config.json fixed the tab link error issue.. BUT I did not check for any other side effects and I gotta go...