[Steven Sandbach (ss421)]

Description

Update required for the API change in: https://github.com/JCSDA-internal/oops/pull/3374

The main changes are:

output:
  date: ...
  state variables: ...
  nemo field file: ...
  output nemo field file: ...

to:

output:
  analysis:
    date: ...
    state variables: ...
    nemo field file: ...
    output nemo field file: ...

and removal of the final section as it is no longer required:

final:
  diagnostics:
    departures: oman

See: see: https://github.com/JCSDA-internal/oops/pull/3374#issuecomment-6177427686 for details.

Testing

Tested via: cylc vip -n "test_oops_feature_outputs" -z TASKS=all_ctest with output here

I did not review the output file and since this file is not compared to anything it would be worth checking there are no issues. Toby Searle (@twsearle), would you mind taking a look at the output files in the updated tests.

The file outputs are located in: /home/users/steven.sandbach/cylc-run/test_oops_feature_outputs/run1/share/make_mo__spice_gnu_debug/orca-jedi/src/tests/testoutput

I nottice that analysis_3dvar_sic_identity.nc and analysis_3dvar_sic.nc are both present but 3dvar_sic_identity.nc and 3dvar_sic.nc are not. Perhaps that is expected?

Issue(s) addressed

None opened

Dependencies

Dependencies

build-group=https://github.com/JCSDA-internal/oops/pull/3374
build-group=https://github.com/MetOffice/jjtests/pull/206
build-group=#204

• https://github.com/MetOffice/jjdocs/pull/700

Impact

Expected impact on downstream repositories or workflows:

Update of workflows that run the variational application following the changes here. Toby Searle (@twsearle) - can you advise if there are configuartions stored elsewhere that need to be tested for this application.

Checklist

• I have updated the unit tests to cover the change
• New functions are documented briefly via Doxygen comments in the code
• I have linted my code using cpplint
• I have run the unit tests
• I have run mo-bundle to check integration with the rest of JEDI and run the unit tests under all environments

[Toby Searle (twsearle)]

Thanks yes, all looks good! The 3DVar stuff is still in an embryonic state, so as long as we are getting the same cost function values and those fields look reasonable (I checked and I think they are right) then we are good.