add CHELSA V2

[tiemvanderdeure]

This package currently downloads CHELSA data from V1.

CHELSA V2 has a bunch of additional variables, so I want to make those accessible from this package.

I think version 2 should become the default, but we need to still allow for version 1. The easiest way to do that would be to have version as a keyword argument for the CHELSA type

To-do list

• Version argument?
• Bioclim+ variables
• Version 2 for future climate

[rafaqz]

Sure Chelsa version type parameter makes sense, 2 is assumed if not specified

[rafaqz]

Re test failures: the Aqua.jl keywords have changed. We'll need to remove the toml formatting part and also probably specify the Aqua.jl version in Project.toml

[tiemvanderdeure]

Just need to add some tests and this is good to go.

For future climate, I haven't added a version keyword because CMIP6 is always version 2.

The bioclim variables are in the exact same path as the bioclim+ variables, so we could consider making it so that getraster(CHELSA{BioClim}) calls getraster(CHELSA{BioClimPlus}), and that will save us some lines of code.

[tiemvanderdeure]

@rafaqz could you review this and help me troubleshoot these test failures?

[rafaqz]

Ok so Aqua.jl dosen't do toml formatting anymore, yoy can delete that line from the tests.

[rafaqz]

Aqua is in the runtest.jl file, near the top. While youre at it maybe add a line there to include("chelsa_climate.jl") ;)

[tiemvanderdeure]

@rafaqz It now fails because of codecov. Could you take a look?

[rafaqz]

This looks good to go. Do you want to also update the readme table so CHELSA includes plus and climate?

(We can ignore that error its just codecov)

[tiemvanderdeure]

FYI @rafaqz - precompilation gives this output. I can't figure out which line triggers this (nothing to do with this PR I think - it's the same on master)

Precompiling RasterDataSources
        Info Given RasterDataSources was explicitly requested, output will be shown live
┌ Warning: Replacing docs for `RasterDataSources.getraster :: Union{}` in module `RasterDataSources`   
└ @ Base.Docs docs\Docs.jl:243

[rafaqz]

Yeah would be nice to get a line number for that...

Fixed in #65

[rafaqz]

How can WorldClim downloads be this unreliable

[tiemvanderdeure]

Are you kidding me, worldclim is down again!! I think this PR is done - can you re-trigger tests manually when worldclim is up again @rafaqz?

[rafaqz]

Yeah, maybe we need to cut any testing dependency on them or make it optional based on if it loads at all... Or maybe we can use GitHub actions cache somehow to cache the downloads now

It's so weird for such a widely used dataset.

[tiemvanderdeure]

@rafaqz I think this PR is good to go - the tests just failed because worldclim was down. Can you manually re-run the tests? And maybe give the code a final glance?

[rafaqz]

Thanks, great to have this :)