Understanding the output of raw_pattern in rawpy

Question

I have the following two examples when loading 2 different raw images (one with Nikon and another with Sony Both having Bayer filters):

Can someone please explain the meaning of the output of the raw_pattern here along with difference in the outputs of the examples provided.

Answer 1

The Nikon (NEF) pattern means "GBRG" and the Sony (ARW) pattern means "RGGB", assuming that a.color_desc == "RGBG". So the integers in raw_pattern are indices to color_desc.

Answer 2

Code to get bayer partten string

import rawpy, numpy as np
raw_obj = rawpy.imread("xxx.dng")
bayer_partten = "".join([chr(raw_obj.color_desc[i]) for i in raw_obj.raw_pattern.flatten()])