add Dataset plot accessor.

[dcherian]

to wrap _DatasetPlotMethods.

This should be easy → follow what is done with _CFPlotMethods

[jukent]

@dcherian Where does _DatasetPlotMethods exist? I can't find it in this repository's code. And I only see one reference to _CFPlotMethods --

cf-xarray/cf_xarray/accessor.py

Line 482 in 19d80cf

An extra decorator, if necessary. This is used by _CFPlotMethods to set default

Are these upstream in Xarray?

[dcherian]

I may have changed design since this was written. I think now we just need to modify _CFPlotMethods appropriately to work.

We want the following to work:

import cf_xarray
import xarray as xr

ds = xr.tutorial.load_dataset("air_temperature").isel(time=slice(2))
ds.air.attrs["standard_name"] = "air_temperature"
ds.plot.scatter(x="air", y="lat")  # works
ds.cf.plot.scatter(x="air_temperature", y="latitude")  # doesn't work

cf-xarray/cf_xarray/accessor.py

Line 624 in 19d80cf

key_mappers=dict.fromkeys(self._keys, (_get_axis_coord_single,)),

We should add the _get_standard_name mapper to the above dictionary. I think that might be all that's needed to get scatter to work.

[jukent]

_get_standard_name

I'm a little iffy on using the dict.fromkeys method with more than one value. Would it look something like:

key_mappers=dict.fromkeys(self._keys, (_get_axis_coord_single, _get_standard_name)), ?

[dcherian]

Yes