Documentation still refers to degrees for skeleton_to_csgraph

[SimonPeatman]

The documentation here gives the example:

pixel_graph0, coordinates0 = skeleton_to_csgraph(skeleton0, spacing=spacing_nm)

But the text immediately beneath it says:

Finally, degrees is an image of the skeleton, with each skeleton pixel containing the number of neighbouring pixels. This enables us to distinguish between junctions (where three or more skeleton branches meet), endpoints (where a skeleton ends), and paths (pixels on the inside of a skeleton branch.

It looks like this information is out of date. How do I get an array of degrees as described?

[jni]

Oh interesting, looks like the current attribute/API is not at all well documented. Apologies! There's a few options to get you unstuck:

• use skan.csr.make_degree_image: like it says on the box.
• Use the (undocumented 😬) degrees attribute of the Skeleton class (docs). degrees now is just an array of degrees of the same length as the nonzero pixels of the image. You can use this as follows:

  skel = Skeleton(skeleton_image)
  degree_image = np.zeros_like(skeleton_image, dtype=np.uint8)
  degree_image[tuple(skel.coordinates.T)] = skel.degrees

• Same as above, but using the new PathGraph API (using .node_coordinates instead of .coordinates)

I would be super grateful if you submitted a PR to improve the docs. Also, I notice you work in meteorology? If you have a meteorology use case for skan I would love to hear about it and maybe add it to our usage examples??? 🙏