Support passing in cmap and norm for the odc.geo.xr.ODCExtension.explore function

[alexgleith]

Hey folks

Not sure if this is expected to work, but I have a nice colourmap for a categorical plot, example here.

Code is something like this:

from matplotlib import colors

classes = [
    [1, "bare_land", "#968640"],
    [2, "forest", "#064a00"],
    [3, "crops", "#ffce33"],
    [4, "grassland", "#d7ffa0"],
    [5, "settlements", "#b3b2ae"],
    [6, "mangroves", "#07b28d"],
    [7, "water", "#71a8ff"],
    [8, "quarry", "#b03a2e"]
]

values_list = [c[0] for c in classes]
color_list = [c[2] for c in classes]

# Build a listed colormap.
c_map = colors.ListedColormap(color_list)
bounds = values_list + [9]
norm = colors.BoundaryNorm(bounds, c_map.N)

data["class"].isel(time=0).plot.imshow(cmap=c_map, norm=norm, size=10)

Would it be possible for the amazing explore() function to be able to render with these colourmaps? Currently it fails with TypeError: Object of type BoundaryNorm is not JSON serializable

[Kirill888]

It's actually an issue with mapping function itself, it doesn't know of norm= so passes it on to folium/ipyleaflet, where it tries to push it to JavaScript, hence the error.

So these need to be augmented to understand norm= parameter:

odc-geo/odc/geo/_rgba.py

Line 174 in 1f88854

def _matplotlib_colorize(x, cmap, vmin=None, vmax=None, nodata=None, robust=False):

odc-geo/odc/geo/_rgba.py

Lines 204 to 206 in 1f88854

	def colorize(
	x: Any,
	cmap=None,

odc-geo/odc/geo/_map.py

Lines 54 to 65 in 1f88854

	def add_to(
	xx: Any,
	map: Any,
	*,
	name: Optional[str] = None,
	fmt: str = "png",
	max_size: int = 4096,
	resampling: str = "nearest",
	# jpeg options:
	transparent_pixel: Optional[Tuple[int, int, int]] = None,
	# RGB conversion parameters
	cmap: Optional[Any] = None,