Stream: IG creation
Topic: Template visual glitches
Mark Kramer (Jan 07 2020 at 20:29):
pasted image
Column too narrow to show cardinality
Mark Kramer (Jan 07 2020 at 20:30):
pasted image
column too narrow
Mark Kramer (Jan 07 2020 at 20:32):
pasted image
No ordering
Mark Kramer (Jan 07 2020 at 20:34):
Very difficult to read; can't tell where the second column aligns with first column pasted image
Mark Kramer (Jan 07 2020 at 20:38):
Footer size variation
pasted image
Mark Kramer (Jan 07 2020 at 20:39):
Logo distortion
pasted image
Mark Kramer (Jan 07 2020 at 20:42):
Most important part -- the definition of the element -- is getting lost
pasted image
Mark Kramer (Jan 07 2020 at 20:43):
Spacing issue before 4-th level heading
pasted image
Mark Kramer (Jan 07 2020 at 20:44):
valueQuantity:valueQuantity?
pasted image
Chris Moesel (Jan 07 2020 at 20:56):
pasted image
Column too narrow to show cardinality
Actually, I think the missing cardinality is because when we generate the diff element, we're only putting in things that are truly different... So if only the min changed or only the max changed, you might get what looks like an unbalanced cardinality.
The question is whether we should be intentionally putting non-diff properties into the diff in order to make the doc more readable, or if the doc generator should be doing something about this.
Jose Costa Teixeira (Jan 07 2020 at 21:03):
Most important part -- the definition of the element -- is getting lost
@Lloyd McKenzie PR9 created
Mark Kramer (Jan 07 2020 at 23:08):
@Chris Moesel I understand the missing cardinality now. Thanks for the explanation.
Mark Kramer (Jan 07 2020 at 23:12):
Sidebar menu too narrow.
pasted image
pasted image
Sarah Gaunt (Jan 07 2020 at 23:21):
Is this the "official" (aka Lloyd's) template you are talking about? Because I'm not getting missing cardinalities (this is with the latest template - which also appears to have a purple(!), rather than red, header and footer)...
pasted image
pasted image
pasted image
Sarah Gaunt (Jan 07 2020 at 23:23):
And if it is an issue, I definitely think it should be taken care of by the generator and not by us intentionally putting non-diff properties into the diff...
Jose Costa Teixeira (Jan 07 2020 at 23:38):
@Lloyd McKenzie the purple was the base template, not the hl7 template. I just checked and it seems to be changed to purple. is this intended?
Jose Costa Teixeira (Jan 07 2020 at 23:39):
@Mark Kramer also here cardinalities are fine. which ig are you building to see this?
Mark Kramer (Jan 08 2020 at 15:03):
The template is hl7.fhir.template#current. The cardinality "glitch" is only in the Differential view. The theory is that is only one side of the cardinality has changed, then only that change is shown in the diff. In retrospect, this seems reasonable to me.
Lloyd McKenzie (Jan 08 2020 at 16:45):
Wow. Ok, from the top @Mark Kramer:
- I believe this is driven by what's changed in the differential. However, the other cardinality should still be showing up greyed out. Please submit a change request if you're seeing this using the most recent publisher and template - you should be using )
- Overly narrow column - where is this showing up? Artifact list? I'll see if I can improve
- There is ordering. It's alphabetic by name. If you want to control the order, then you need to manually list the resources in your IG. If you do that, the resources will list in the order you define.
- Also presumably artifact list? I'll try to improve.
- Footer size variation - not understanding the issue?
- For the logo distortion, which template are you using? Because the current template shouldn't have the FHIR logo on the left-hand-side
- Can you better explain the issue?
- Will try to fix the spacing
- @Grahame Grieve ?
- (sidebar menu) Is there a place I can see your IG? the indentation seems more than I'd have expected
- @Sarah Gaunt How are you getting the purple? When I do a local build using template = hl7.fhir.template, I get red as usual. (And the CI build doesn't work for hl7.fhir.template due to trust issues.)
Mark Kramer (Jan 08 2020 at 18:05):
@Lloyd McKenzie
- I can confirm that I'm using #current templates.
- Will file an issue
- Narrow column is on Table of Contents -> Artifact Summary
- The list of artifacts is an alphabetic list of profiles, followed by an alphabetic list of extension, as one big list.
- Correct, also the the Artifact Summary. * TABLE GRID LINES WOULD HELP TREMENDOUSLY everywhere! *
- The footer size on this page is not like the footer size on other pages. It is about 3x bigger (24 pt instead of 9 pt font)
- The logo is still shown on the History page. That's where the screen shot comes from.
- The documentation has a bunch of information, including the official URL, status, maintainer, content of use, etc. that are given much more visual prominence than the definition of the element. The definition (which most people would consider the most import thing) is the lost in the noise. Important stuff first, less important things later.
- (sidebar menu) You could see it, if autobuild were working, but it is not. This screenshot is from: https://build.fhir.org/ig/HL7/fhir-shorthand/
Lloyd McKenzie (Jan 08 2020 at 18:34):
- So the issue is that extensions aren't being split from profiles? When I look at fhir-shorthand, I'm not seeing anything in the artifacts.html at all. Is everything checked in?
- So footer on artifact summary page looks different for you? (It's looking the same for me)
- History page is not in the control of the template at all. It's generated by @Grahame Grieve as part of the publication process.
- Understood.
- The menu shouldn't be going that deep. I'll try to figure out why that's happening.
Chris Moesel (Jan 08 2020 at 19:44):
@Lloyd McKenzie -- I think some of @Mark Kramer's comments are based on another IG (not the fhir-shorthand one), so that's why some of those issues are reproduced there.
Sean McIlvenna (Jan 08 2020 at 19:56):
@Lloyd McKenzie , for the purple, I pointed this out to you yesterday and you said you thought it was a bug that you were going to investigate. Is that no longer the case?
Sean McIlvenna (Jan 08 2020 at 19:59):
TestSeanJames.zip Here is an example IG that I just ran today that demonstrates the purple issue.
Sean McIlvenna (Jan 08 2020 at 20:24):
also, perhaps there should be a separate stream under #IG creation just for the new template
Sean McIlvenna (Jan 08 2020 at 20:24):
given that ToF is close to putting the new template into production, I'm sure there will be plenty of feedback very quickly
Sean McIlvenna (Jan 08 2020 at 20:25):
we should think how to handle the triage of that feedback
Sarah Gaunt (Jan 08 2020 at 23:39):
@Lloyd McKenzie for issue 3 around ordering - confused about this statement: "There is ordering. It's alphabetic by name. If you want to control the order, then you need to manually list the resources in your IG. If you do that, the resources will list in the order you define."
If you don't manually list the resources in your IG, how are they part of the IG? Surely you always need to manually list the resources in your IG?
Lloyd McKenzie (Jan 08 2020 at 23:55):
With the new template so long as you stick your resources in folders with the right names, the publisher will find the files figure out what their metadata is and populate the IG accordingly (including grouping things based on whether it's a profile, logical model, extension, data type or some other type of resource).
Sarah Gaunt (Jan 09 2020 at 00:02):
What about setting exampleCanonical for examples? Do you lose the ability to have the profiles auto-alphabetically ordered as soon as you set an example's profile?
And speaking of examples, is there a trick to get them to show up on the Examples tab of the StructureDefinition page? I have the following in the ImplementationGuide and this example isn't showing up (that profile just has a "No examples are currently available for the Profile." message...
<resource>
<reference>
<reference value="Observation/observation-BSeR-ArthritisFeedbackObservation-1-eve-everywoman"/>
<display value="observation-BSeR-ArthritisFeedbackObservation-1-eve-everywoman"/>
</reference>
<name value="observation-BSeR-ArthritisFeedbackObservation-1-eve-everywoman"/>
<exampleCanonical value="http://hl7.org/fhir/us/bser/StructureDefinition/BSeR-ArthritisFeedbackObservation"/>
</resource>
Lloyd McKenzie (Jan 09 2020 at 00:30):
What's supposed to happen when resources are found automatically is that if an instance declares a profile that's defined in the IG, it'll be treated as an example. (And if it's a resource that isn't a conformance resource, it'll be treated as an IG not tied to any particular profile.) I believe that's working in the sample-ig.
Sarah Gaunt (Jan 09 2020 at 00:48):
Well, the sample IG doesn't have an instance that declares a profile that's defined in the IG, so hard to tell if it's working. It only has the one Patient example...
Sarah Gaunt (Jan 09 2020 at 00:49):
Can I create an example for the sample IG and test? i.e. will I have permissions to push to that repo?
(PS. Thanks for getting that stupid John Jacob song in my head with your Patient name... UGH...) :joy:
Sarah Gaunt (Jan 09 2020 at 00:57):
NM, I'll just test locally.
Sarah Gaunt (Jan 09 2020 at 01:15):
So I tested locally and on the artifacts page the two different examples are nicely grouped:
pasted image
But there is still nothing on the profile page:
pasted image
Lloyd McKenzie (Jan 09 2020 at 02:09):
Purple is fixed. It was an issue introduced when we added the active tables
Sarah Gaunt (Jan 09 2020 at 02:43):
Still showing up as purple in the sample ig both locally and ci build
Lloyd McKenzie (Jan 09 2020 at 04:41):
SampleIG is supposed to be purple. It builds against the base template. However, if you build against the HL7 template, you'll now get white, CDA should give you green and FHIR will give you red.
Sarah Gaunt (Jan 09 2020 at 05:25):
Ok, thanks for the info
Lloyd McKenzie (Jan 10 2020 at 20:04):
@Mark Kramer can you indicate an IG where I can reproduce the issues that aren't found in the shorthand one?
Grahame Grieve (Jan 13 2020 at 06:28):
I've tried to follow this discussion, but I don't know what the numbers are talking about. is there any open issues still?
Lloyd McKenzie (Jan 13 2020 at 07:53):
Not that I could see in current IGs. Waiting for @Mark Kramer to indicate where he was encountering them.
Jose Costa Teixeira (Jan 13 2020 at 09:14):
There is one issue I cannot find out where it comes from (template or publisher): the examples do not show up in the example list for a profile.
I note that every site.data.pages[xxx].examples is an empty list, so the template will of course not show them. Is this created in the Publisher? If so, what are the conditions for this list to be populated?
Grahame Grieve (Jan 13 2020 at 11:17):
where's an example I can reproduce?
Grahame Grieve (Jan 13 2020 at 11:32):
you have to populate it in the ig as far as I know
Jose Costa Teixeira (Jan 13 2020 at 12:50):
they are in the BE-core ig
Jose Costa Teixeira (Jan 13 2020 at 12:54):
<resource> <reference> <reference value="Communication/be.comm1"/> </reference> <fhirVersion value="4.0.1"/> <name value="Communication example 1"/> <description value="New message from DZOP to nurse"/> <exampleCanonical value="https://www.ehealth.fgov.be/standards/fhir/StructureDefinition/be-communication"/> </resource>
This is the IG part
Jose Costa Teixeira (Jan 13 2020 at 12:57):
and the profile is there:
<?xml version="1.0" encoding="utf-8"?> <StructureDefinition xmlns="http://hl7.org/fhir"> <id value="be-communication" /> <url value="https://www.ehealth.fgov.be/standards/fhir/StructureDefinition/be-communication" /> <version value="0.1.0" /> <name value="BECommunication" /> ...
Lloyd McKenzie (Jan 13 2020 at 16:57):
Examples not showing up is an issue with the publisher. I expect to submit a pull request today on it.
Jose Costa Teixeira (Jan 13 2020 at 17:42):
Ah, thanks
Jose Costa Teixeira (Jan 13 2020 at 17:43):
Sorry for making the same request twice
Grahame Grieve (Jan 13 2020 at 22:30):
is be-core in github?
Jose Costa Teixeira (Jan 13 2020 at 22:35):
yes, master branch
Grahame Grieve (Jan 13 2020 at 22:57):
it's not in the ci-build, so what's the URL?
Jose Costa Teixeira (Jan 13 2020 at 22:58):
https://github.com/hl7-be/be-core
Jose Costa Teixeira (Jan 13 2020 at 23:09):
(Lloyd mentions above that he found the issue)
Grahame Grieve (Jan 13 2020 at 23:10):
hmm so I don't have to do anything? great
Jose Costa Teixeira (Jan 13 2020 at 23:14):
not on this topic, I guess...
Mark Kramer (Jan 14 2020 at 19:54):
@Lloyd McKenzie I have the latest publisher, and I've re-run the IGs to check the status of the issues. They come from two different IGs.
- The FHIR Shorthand IG (repo.
- The mCODE IG has not been published to build.fhir.org, but I have attached a zip of the build files to this message so you can try it yourself.
I'm going to use letters because I can't follow the numbering anymore...
Issue A (mCODE IG): Profile differentials can show only one side of the cardinality, if only one of the min or max has changed. Not claiming this is wrong, but it merits consideration.
Issue B (mCODE IG): Left-hand column too narrow on Artifacts Summary page -- no change, column is still too narrow.
Issue C (mCODE IG): Ordering of artifacts in TOC -- still not fully alphabetical -- it goes A-Z with profiles, then A-Z with extensions, all as one big list.
Issue D (mCODE IG): Still no gridlines in table on Artifacts Summary page, but the cells are now top-justified, and that makes it easier to read. Would prefer to see gridlines, though.
Issue E (FSH IG): Right-side sidebar nav menu too narrow (index page)
Fixed Issue F (mCODE IG): Footer on the Examples page is now well-behaved (was wrong size)
Grahame Grieve (Jan 14 2020 at 19:57):
Issue A is a bug somewhere. That presentation should never happen. can you point to an example so I can debug?
Jose Costa Teixeira (Jan 15 2020 at 06:56):
Suggestion : to keep track of these issues, we should create and track
- issues in github (if we know where the problem is) or
- tasks in Jira
Jose Costa Teixeira (Jan 15 2020 at 06:57):
If we do not keep track of our requests and we just need a fix here and there, we can hadrly claim we need change logs
Jose Costa Teixeira (Jan 15 2020 at 06:58):
if this is not a good proposal, and knowing that we already have letters and numbers for issues,
I'm calling dibs on roman numerals
Lloyd McKenzie (Jan 15 2020 at 14:24):
If it's something I don't expect to get to in the next few days, I'll probably capture something in the Git tracker for the template or wherever the change needs to go to fix. But I do hope to get to these soon-ish, so not necessarily a need
Last updated: Apr 12 2022 at 19:14 UTC
