catch colormap issue and raise a more descriptive error message

[vincentsarago]

Titiler/rio-tiler let the user to do a lot of things when it comes to image rendering. One issue that seems to come ofter is when users want to apply a colormap after some band maths, user would need to rescale the values to match the data type of the colormap (usually uint8, 0->255). Because we cannot automagically rescale the data (e.g using the min/max of the local array, which would cause each tile to look different) we should at least make sure the error message says something a bit more useful than "arrays used as indices must be of integer (or boolean) type" 😅

either in

titiler/src/titiler/core/titiler/core/utils.py

Lines 73 to 77 in 6d1519e

	if colormap:
	data, alpha_from_cmap = apply_cmap(data, colormap)
	# Combine both Mask from dataset and Alpha band from Colormap
	mask = numpy.bitwise_and(alpha_from_cmap, mask)
	datatype_range = (dtype_ranges[str(data.dtype)],)

or https://github.com/cogeotiff/rio-tiler/blob/ce62e555562fd7a38bde4671ab680dfb2b5bdd15/rio_tiler/colormap.py#L90-L135

ref: EOEPCA/data-access#142

[vincentsarago]

in https://github.com/cogeotiff/rio-tiler/blob/main/CHANGES.md#721-2024-11-14 we've added a catch to make sure the issue mentioned in ☝ doesn't happen.

Most of colormap related issue are already caught in rio-tiler, and the most common one (data type issue) is handled so I'm going to closes this for now.

if anyone feels we should re-open this please post a reproducible example 🙏