Stream: genomics/committers
Topic: MolecularSequence for R5
Kevin Power (Feb 22 2022 at 17:35):
Starting up a topic here to discuss the progress.
CC: @ Bob Milius / @John David Larkin Nolen
Kevin Power (Feb 22 2022 at 17:36):
Here is the motion that was made and passed today:
Motion to remove at least two elements: quality, structureVariant and change the name of the variant element to “Edits” for R5. Add a strong narrative describing IM work and point to the IG. Remove, update, change or edit examples as appropriate.
Bob Milius (Feb 22 2022 at 18:54):
@Kevin Power pointed me to this: https://github.com/hl7/fhir/wiki/Get-Started-with-FHIR-on-GitHub which also points to https://confluence.hl7.org/display/FHIR/FHIR+Build+Process
Can anyone verify if this is the current recommended workflow for submitting a change to build?
Bob Milius (Feb 22 2022 at 19:02):
Should I be cloning master or the 5.0.0-snapshot1 branch, or something else for the R5 ballot build?
Kevin Power (Feb 22 2022 at 19:21):
Looks like master is the right one, considering the commits I am seeing to it - but might defer to @John David Larkin Nolen
Bob Milius (Feb 22 2022 at 19:28):
@John David Larkin Nolen replied in email. His process is:
- Download the master
- Make a branch with my username and the JIRA number
- Make changes and run the publisher locally.
- Expect some errors and make fixes. Usually a change in the resource will break either an example or two and/or the search parameters.
- Fix that and get the build to work locally.
- Commit and push
- Fire off a pull request and cross your fingers. Sometimes the CI build publisher finds things that are not found locally.
Bob Milius (Feb 22 2022 at 19:48):
As far as I can tell, we don't have a jira to modify MolecularSequence. Correct?
Kevin Power (Feb 22 2022 at 19:51):
We have several JIRA (I linked to them from your document @ Bob Milius ) - but we might need a new one that matches up to what we voted on during the meeting today.
Bob Milius (Feb 22 2022 at 20:00):
ah... okay. I used the link for the search https://jira.hl7.org/issues/?filter=12310 and didn't see them on the first page. I didn't realize there are many pages of results.
Kevin Power (Feb 22 2022 at 20:02):
You want to review the filter I linked in the google doc you started:
Kevin: This list should be compared with previous logged requests:
https://jira.hl7.org/issues/?filter=13632
Bob Milius (Feb 23 2022 at 16:35):
I made a local branch and started a build of FHIR core, and ran out of memory. Then I noticed this in the FHIR Build Process document
NOTE: As of February 21, 2022 the build appears to require at least 16GB of ram memory. Your results may vary.
I increased memory and restarted the build... took 45 minutes.
BUILD SUCCESSFUL in 45m 7s
1 actionable task: 1 executed
Next step to make some edits, rebuild. I haven't pushed the branch yet.
Bret H (Feb 23 2022 at 16:43):
fyi. I'm done with a TA correction. all pushed up. : ^ ) Thanks for working on the Resource Bob!
Kevin Power (Feb 23 2022 at 16:45):
@Bret H - This thread is just talking about changes to FHIR Core and the MolSeq resource, not our IG. And I think you only made a change to the IG, is that right?
Bret H (Feb 23 2022 at 16:54):
Yes. You are correct. Sorry to cause confusion.
Thanks for working on the Resource Bob.
Bob Milius (Feb 23 2022 at 17:54):
Created FHIR-36085 to be used in commit descriptions (pull requests need to have a Jira tracker number)
Bob Milius (Feb 23 2022 at 21:55):
I created a branch... commented out structureVariant and quality, removed fda examples. Took about 30min to build.
https://github.com/HL7/fhir/tree/bmilius_molseq_r5
Bob Milius (Feb 23 2022 at 22:42):
Looks like we'll have more time...
see R5 NOT going to ballot in May
Kevin Power (Feb 23 2022 at 22:55):
Perhaps need to rethink our approach and perhaps think more seriously about 'Align with IM in R5' ?
Bob Milius (Feb 24 2022 at 02:07):
I updated the branch. I had to edit the genomics.html to comment out the fda examples it references (It was breaking the build). So one thing I learned, if I edit something in the molecularsequence folder, only that resource gets built (about 5 minutes). But then if I run the publish.sh again, the whole fhir core gets rebuilt, including everything that depends on the resource I just edited. So no errors in the partial build, but errors found in the whole build. Don't count on a successful build if you just edit the resource. We need to do complete rebuild to see if it broke anything else (about 45 min).
Bob Milius (Feb 24 2022 at 02:08):
Kevin Power said:
Perhaps need to rethink our approach and perhaps think more seriously about 'Align with IM in R5' ?
we should keep it in mind. Let's see what the new timeline is.
Bob Milius (Feb 24 2022 at 02:49):
I'll start renaming the variant element tomorrow (Thu).
Bob Milius (Feb 24 2022 at 16:25):
I'm using vscode as my editor. It creates a .vscode directory. I assume I should add it to .gitignore. Correct?
Bret H (Feb 24 2022 at 18:05):
Yep. The whole vs directory is ignored
Visual Studio
.vs/
If useful, here's my gitignore for the IG work, different but similar. .gitignore
Bret H (Feb 24 2022 at 18:06):
(deleted)
Kevin Power (Feb 24 2022 at 18:11):
This makes sense to me, but since we are talking about FHIR Core here, perhaps there are some constraints? But yea, I would suggest adding it.
Bret H (Feb 24 2022 at 18:13):
*adding an ignore for the visual studio files folder, right?
Kevin Power (Feb 24 2022 at 18:14):
Correct.
Kevin Power (Mar 23 2022 at 14:14):
I have pushed changes to the branch @ Bob Milius - I merged from main to our branch to keep it in sync, and I started making updates to change 'variant' to 'edit' - I hope it should build in the next 20ish minutes.
Kevin Power (Mar 23 2022 at 20:49):
After my changes today, I have created a draft PR, which I think is what triggers the build to happen, so that will give us a published branch to review.
https://github.com/HL7/fhir/pull/1811
Kevin Power (Mar 23 2022 at 22:51):
And we have a branch. Feel free to review the current state of the MolSeq resource and the genomics guidance page. It is certainly a work in process.
https://build.fhir.org/branches/bmilius_molseq_r5/
Bob Milius (Mar 24 2022 at 14:04):
fyi... R5 Roadmap announcement
Kevin Power (Mar 24 2022 at 14:12):
Thanks Bob. Still no decision on R5 content cutoff (June or August). :frown:
Kevin Power (Mar 24 2022 at 19:36):
I have made some significant changes to the 'genomics guidance' and 'MolecularSequence' pages the last couple of days. Can't say it is all good yet, but I welcome review and feedback:
https://build.fhir.org/branches/bmilius_molseq_r5/
Bret H (Mar 28 2022 at 01:23):
https://build.fhir.org/branches/bmilius_molseq_r5/genomics.html looks like it will be excellent, best I have seen it in years.
Last updated: Apr 12 2022 at 19:14 UTC
