infer dimension ordering within nc_specs using cf-xarray

[huard]

Pull Request Checklist:

• This PR addresses an already opened issue (for bug fixes / features)
  • This PR fixes #xyz
• (If applicable) Documentation has been added / updated (for bug fixes / features).
• (If applicable) Tests have been added.
• CHANGELOG.rst has been updated (with summary of main changes).
  • Link to issue (:issue:number) and pull request (:pull:number) has been added.

What kind of change does this PR introduce?

• ...

Does this PR introduce a breaking change?

Other information:

[RondeauG]

LGTM!

So if I get this correctly, this will only help the generation of the rvt file so that the dimensions are in the expected order, right?

From my tests, Raven is also quite sensitive about the order of the dimensions within the dataset itself (t, y, x). Do you think that it is within the scope of RavenPy to warn the user if their data isn't ordered correctly?

[huard]

Correct.

Huh. I'm surprised by this. I thought Raven covered all possible combinations. I would recommend documenting such issues and opening a ticket on the Raven repo.

[RondeauG]

Actually, while I was talking about this with @richardarsenault, I realized that it might be a partial issue. With data in the X,Y,T order, I had this issue with missing meteorological data that proves that Raven still understood the NetCDF file correctly:

However, this was with fake elevation data that covered the entire grid. When I moved on to real elevation data (and after fixing the issue of missing grid cells), I had errors that make me think that elevation specifically is still hard-coded Y,X.

[huard]

I see explicit logic around the dimension order of lat, lon and elev.
https://github.com/CSHS-CWRA/RavenHydroFramework/blob/f9c77a2491ce271493ea9629859d45f84ed181cd/src/ForcingGrid.cpp#L1016

[RondeauG]

True, but I have errors when data isn't in the right order... This might be an xHydro issue? This is fine here, then.