Implement RaggedArray.from_xarray()

[milancurcic]

As discussed with @selipot on 10/20, it's in scope for clouddrift to allow getting a RaggedArray instance from an xarray.Dataset.

RaggedArray.from_netcdf() already does this internally, i.e.:

clouddrift/clouddrift/dataformat.py

Lines 78 to 97 in 437c0e4

	with xr.open_dataset(filename) as ds:
	nb_traj = ds.dims["traj"]
	nb_obs = ds.dims["obs"]
	attrs_global = ds.attrs
	for var in ds.coords.keys():
	coords[var] = ds[var].data
	attrs_variables[var] = ds[var].attrs
	for var in ds.data_vars.keys():
	if len(ds[var]) == nb_traj:
	metadata[var] = ds[var].data
	elif len(ds[var]) == nb_obs:
	data[var] = ds[var].data
	else:
	print(
	f"Error: variable '{var}' has unknown dimension size of {len(ds[var])}, which is not traj={nb_traj} or obs={nb_obs}."
	)
	attrs_variables[var] = ds[var].attrs

Some assumptions are currently needed for the dimension names (currently assumed "traj" and "obs"). The method should allow the user to specify dimension names to use, in absence of an established convention.