fix the documentation build for fv3atm, getting the CI to work

[edwardhartnett]

Description

In this PR I change the CMake option ENABLE_DOCS to ENABLE_FV3ATM_DOCS, in order to avoid invoking the documentation build of UPP.

I also add a new CI workflow called "docs", and take the documentation build out of the GCC workflow. This will be more useful to programmers because having the "docs" CI run fail is pretty obvious.

There are no code changes in this PR; this is documentation and CI only.

Issue(s) addressed

Fixes #904

Testing

CI testing. No code changes in this PR.

Dependencies

N/A

Requirements before merging

• All new code in this PR is tested by at least one unit test
• All new code in this PR includes Doxygen documentation
• All new code in this PR does not add new compilation warnings (check CI output)

[DusanJovic-NOAA]

I am not familiar with Doxygen and the changes required in the workflow to support its use, so I can not properly review this PR. I will approve it, assuming the code here does what it's supposed to do.

[edwardhartnett]

Who is the other reviewer required to make this PR mergable?

[DusanJovic-NOAA]

Who is the other reviewer required to make this PR mergable?

@junwang-noaa

[edwardhartnett]

Probably we need to add @LarissaReames-NOAA to the list of those who can approve a PR here.

@junwang-noaa can you please take a look at this PR and approve it?

[junwang-noaa]

@edwardhartnett I approved it already, please see the comments above. After @LarissaReames-NOAA accept the invitation, I will add her to the list. Thanks

Probably we need to add @LarissaReames-NOAA to the list of those who can approve a PR here.

@junwang-noaa can you please take a look at this PR and approve it?

[edwardhartnett]

Thanks!

[edwardhartnett]

Can this be merged please?

[edwardhartnett]

@DusanJovic-NOAA @LarissaReames-NOAA can this be merged please?

[DusanJovic-NOAA]

@LarissaReames-NOAA Should we merge this PR directly, without running standard ufs-weather-model tests and update the FV3 submodule pointer in ufs-wm, as we (currently) do with other fv3atm PRs?

[LarissaReames-NOAA]

@DusanJovic-NOAA If I'm understanding this PR correctly, it presents a unique case where no UWM tests could be affected since it's all GitHub and FV3-specific CI code changes unrelated to UWM tests or forecasts. I've never merged a PR here myself, but if you think that's the best way to handle this, I'm on board.