expose example datasets through `xdggs.tutorial`

[keewis]

xarray.tutorial is really useful when trying to demonstrate functionality. To provide something similar but with DGGS datasets, I've interpolated the air_temperature dataset to H3 and healpix and uploaded it to xdggs-data.

Here's an example:

import xdggs

ds = xdggs.tutorial.open_dataset("air_temperature", "h3")

I'm not sure if this is the best way to choose the grid, or if something like

ds = xdggs.tutorial.open_dataset("air_temperature/h3")

would be better?

[keewis]

I think the current state allows us to also support the other suggestion, so I'm going to go ahead and merge.

The tests will do network requests (to a github repo), if that turns out to be an issue / flaky we can always run them in a separate CI job.