Semi-transparent cloud mask

[jverrydt]

I now have the following custom algorithm defined to mask clouds in my Sentinel 2 TCI imagery. This is working but we want to switch to something that is adding a semi-transparant layer to also show the borders of the swots. Can I also achieve this with a custom algorithm?

`cloud_mask.py

class CloudMaskAlgorithm(BaseAlgorithm):
    value: int = 21000

    def __call__(self, img: ImageData) -> ImageData:
        # Create a mask where values in the first band are within the specified range
        original_mask = np.ma.getmaskarray(img.array)
        cloud_mask = (img.array == self.value)
        combined_mask = original_mask | cloud_mask

        masked_array = np.ma.MaskedArray(img.data, mask=combined_mask)

        return ImageData(masked_array, assets=img.assets,
            crs=img.crs,
            bounds=img.bounds,
            band_names=img.band_names,
            metadata=img.metadata,
            cutline_mask=img.cutline_mask)`

[vincentsarago]

@jverrydt we removed support for partial alpha mask when we switched to numpy.ma.MaskedArray in rio-tiler 5.0

MaskedArray's mask are bools (True or False) so we cannot keep information about partial values. We could maybe introduce a alpha_mask attribute in ImageData class and then make use of it in the render function

ref: #689