@jshaw35 You're a rock star thanks a lot. A few questions for the assembled multitudes:

• Do we want to ask for the RTTOV and swathing bits to be submitted as separate PRs? This PR touches a LOT of files
• Are we comfortable adding changes that aren't covered by the continuous integration? That's not ideal practice... maybe @jshaw35 could use help figuring out how to install RTTOV as part of the CI?

Jonah, This is truly great work! Much thanks for all of your efforts!

@jshaw35 You're a rock star thanks a lot. A few questions for the assembled multitudes:

* Do we want to ask for the RTTOV and swathing bits to be submitted as separate PRs? This PR touches a LOT of files

* Are we comfortable adding changes that aren't covered by the continuous integration? That's not ideal practice... maybe @jshaw35 could use help figuring out how to install RTTOV as part of the CI?

Great points @RobertPincus.

re: separate PRs. @dustinswales had also raised this but the swathing and COSP-RTTOV changes are difficult to separate at this point. We decided to submit this as a single PR. I'd rather not separate the changes but also want to make this easy to review. Perhaps there is a middle ground.

re: CI for RTTOV, I think this is a great idea and could try to support it. I have experience building RTTOV on HPC platforms but not much familiarity with the CI workflow. Dustin, what do you think?

@jshaw35 Thanks for adding the swath test to the CI scripts. I just updated the CI script to include the swathed case in the artifact.
As for RTTOV in the CI, to get started we need to have the RTTOV codebase (tarball) staged somewhere so we can grab it. @alejandrobodas Maybe we could add this to the Google Drive folder where you currently keep the KGOs? Or @jshaw35 maybe you have access to a FTP site or something similar? (You might have other files to store for RTTOV config?)
Once that is in place, I could update the CI script to download/build/install RTTOV. Then @jshaw35 could add a new RTTOV enabled COSP test.
Thoughts?

@jshaw35 @dustinswales thank a lot for this, it's great!
I agree that it would be good to have a CI test that covers this new simulator.
@dustinswales I'm hesitant to get RTTOV from the GDrive folder unless we are completely sure that that doesn't break the terms of the license.

@alejandrobodas thanks for these comments!
re: CI for swathing, I think adding this to CI can be accomplished now by adding kgos for the new swathed test.

re: RTTOV, I totally agree that we should follow the license terms. Is there a location we can host the source code that can be accessed by the CI only?

re: RTTOV, I totally agree that we should follow the license terms. Is there a location we can host the source code that can be accessed by the CI only?

I suppose we could use a Github 'secret' to download code from a password-protected location, but I've never used secrets so I don't know how it's done.

All, I'm not sure if we're going to make too much progress on solving the problem of running CI tests with an RTTOV installation. What do people think about progress with this PR with a stub interface and investigate the RTTOV installation in a separate discussion?

All, I'm not sure if we're going to make too much progress on solving the problem of running CI tests with an RTTOV installation. What do people think about progress with this PR with a stub interface and investigate the RTTOV installation in a separate discussion?

@alejandrobodas / @mo-abodas, a few comments/updates regarding this question.

I had been working with @dustinswales to create set up an RTTOV test in the CI, but that work is currently tabled and will likely be further delayed. Furthermore, in December we should hear if the RTTOV license is granted an exception and made open source, which would help ease the creation of the CI/CD workflow. I think that RTTOV functionality should be brought into the CI and plan to support that work, but am also eager to see this PR added to COSP2 (for disclosure, COSP-RTTOV cannot be added to CESM without being a part of the official CFMIP COSP repository).

The current COSP-RTTOV code maintains the stub functionality as in previous COSP versions. CI tests are currently failing because we have added an additional test of the swathing functionality but the kgos with masked outputs are not updated yet. If a new kgo is created for the swathed output then the swathing functionality will be brought into CI and RTTOV will remain an optional stub.