CCPP Framework Meeting Minutes 2024 10 24

Agenda

• GitHub issues/PRs
• New items
• Updates
• Other business
• notes

---

Attendees: Michael Kavulich, Jesse Nusbaumer, Dom Heinzeller, Michael Waxmonsky, Dustin Swales, RhaeSung Kim, Soren Rasmussen, Courtney Peverley, Grant Firl, Matt Dawson

GitHub issues/PRs

CCPP Framework (issues, PRs, discussions)

• Add CCPP register phase PR#582

  • Some reviews waiting to be addressed

• Add constituent tendency capability PR#584

  • Has two approvals, needs re-review from @mwaxmonsky

• capgen host model DDT problem Issue#589

  • Added test using a DDT host object to pass information #591
    • Mike gave a review, brought up question about xmllint (see below)

• Amend offline_check_fortran_vs_metadata.py to log what it has checked Issue#592

• Fortran parsing error Issue#593

• Unit conversion bug when variable is used by two different schemes (with different units) Issue#594

• Make full constituents object available via metadata Issue#597

• [xmllint missing from CI jobs Issue#601](xmllint missing from CI jobs)

  • Do we want xmllint to be opt-in or opt-out?
    • Draft PR: https://github.com/NCAR/ccpp-framework/pull/602

• Bugfix for auto transforms PR#604

  • Some CI failures, they seem familiar...

Drafts

• DRAFT TO DISCUSS: Add support and testing for equivalent units #571

Standard names (issues, PRs, discussions)

• Reconsideration of Add four wind CF variables #73? #77

New items for discussion

• How to handle schemes that are called multiple times, but with different inputs?
  • One possible solution: syntax like <scheme name_of_fruit="orange">eat_fruit</scheme> and <scheme name_of_fruit="apple">eat_fruit</scheme>
• Capgen documentation!
• Python support
  • https://github.com/NCAR/ccpp-framework/discussions/605
• Standard Names governance
• Supported units
  • The question has come up again: what units should we support in CCPP framework?
  • https://github.com/NCAR/ccpp-framework/issues/603

Updates from last time

Previous notes

Other business

Meeting notes

CCPP Framework

• Register phase
  • Courtney to respond to comments
• Constituent tendency
  • Michael W to re-review
• DDT test host object test
  • Waiting for Steve G to respond to review comments
  • Led to xmllint issue (#601)
    • Michael W added xmllint to the CI python libraries to fix this, but…
    • Dom’s idea of having a flag to toggle the xmllint check because of operational concerns
      • Everyone is on board with that
    • Michael K - we maintain a schema file in UFS and SCM but don’t ever validate against it
    • Courtney to open issue to get xmllint-ing into offline script
• Bugfix for auto transforms (#604) - Dustin
  • Can’t let group control transforms for variable (get applied every time rather than just the schemes that need the transform
  • Dustin to look into doctest failures
• Fortran parsing error - Steve has a fix, but Michael W has volunteered to look at regex in the meantime
• Equivalent units PR draft
  • Dustin to change the logic to return the variable as is
• Dom - UFS physics changes for optional variables?
  • Dom & Dustin to chat
  • Until the physics are updated, Dom has small change to CCPP framework (NCAR/main) to get things to work

Standard Names

• Reconsideration of add four wind CF variables (#77)
  • Difference between “of” and “in”
  • Related to “Update wind derivative/integral names” (#78)
    • Dustin - we need to update the rules to reflect what we’re doing
    • Dustin - confusion about why “horizontal” is part of the base name instead of a “component”
      • Jesse votes removing the “Generic Names” table in the rules - causes confusion
    • Jesse - perhaps the XML file should just be base names and then a list of all the possible prefixes and suffixes
    • We’ll merge the PR and then have a separate PR for further rule discussion
• PR reviews
  • Not ALL code owners have to review every PR
• Standard names rules and governance discussion 10/31
• What to do if there are two identical quantities from different schemes/processes that need to be kept apart? (#79)
  • Add “due_to” suffix?
  • Everyone seems ok with this

Discussion

• Same suite called multiple times with different inputs
  • Dustin - Would be useful for testing, but UFS doesn’t have use cases for this
    • Dom - could be useful for calling different convection schemes
  • Would need limitations - only scalars
• Python version discussion
  • People should add their requirements to the discussion