Feature request: per band resampling configuration for xr_repoject

[Kirill888]

Currently xr_reproject accepts a resampling: str parameter that configures resampling for all bands of the dataset

odc-tools/libs/algo/odc/algo/_warp.py

Lines 214 to 227 in 3e74605

	def xr_reproject(src: Union[xr.DataArray, xr.Dataset],
	geobox: GeoBox,
	resampling: str = "nearest",
	chunks: Optional[Tuple[int, int]] = None,
	dst_nodata: Optional[NodataType] = None) -> Union[xr.DataArray, xr.Dataset]:
	"""
	Reproject DataArray to a given GeoBox
	:param src : Input src[(time,) y, x]
	:param geobox : GeoBox of the destination
	:param resampling: Resampling strategy as a string: nearest, bilinear, average, mode ...
	:param chunks : In Y,X dimensions only, default is to use input chunk size (ignored if input is not a dask array)
	:param dst_nodata: nodata marker for dst image (default is to use src.nodata)
	"""

Some times it is desirable to resample different bands with different resampling settings, in Datacube.load(..) this is achieved by accepting a dictionary mapping band name to resampling mode to use for the band, with special key '*' to configure resampling for "all other bands".

Would be nice for xr_reproject to accept the same, example

xx_ = xr_rerpoject(xx, geobox,
  resampling={'*': 'average', 'SCL': 'nearest'})