_build_skeleton_path_graph() incorrectly handles degree-0 nodes in calculation of n_points

[ssavary]

skan/src/skan/csr.py

Lines 395 to 400 in 1f71a5b

	endpoints = (degrees != 2)
	endpoint_degrees = degrees[endpoints]
	num_paths = np.sum(endpoint_degrees)
	path_indptr = np.zeros(num_paths + buffer_size_offset, dtype=int)
	# the number of points that we need to save to store all skeleton
	# paths is equal to the number of pixels plus the sum of endpoint

skan/src/skan/csr.py

Lines 406 to 409 in 1f71a5b

	n_points = (
	graph.indices.size + np.sum(endpoint_degrees - 1)
	+ buffer_size_offset
	)

Degree 0 nodes are mistakenly removed from the set. The correction calculation should be:

n_points = (
        graph.indices.size + np.sum(np.max(endpoint_degrees - 1, 0))
        + _buffer_size_offset
        )

With very degenerate cases, I get n_points = 0 and memory dumps occur (on linux). The real issue is the n_points computation, not the memory dump, which is only a consequence of bad n_points).

Here is a matrix that triggers the bug (for some reason the true/except block also triggers the core dump immediately, otherwise it takes a few calls to car.Skeleton):

import numpy as np
from skan import csr

X = np.array([1, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 1]).astype(bool)
try:
     S = csr.Skeleton(X)
except Exception as E:
     print(E)

[jni]

Yikes, good catch. Thank you for the thorough report!

the correction calculation should be

Do you mean np.maximum rather than np.max?

In [1]: np.max([0, 9, -2, 0, 9], 0)
Out[1]: 9

In [2]: np.maximum([0, 9, -2, 0, 9], 0)
Out[2]: array([0, 9, 0, 0, 9])

Anyway, would you be able to/like guidance creating a pull request with a fix and a test? 🙏

[ssavary]

Do you mean np.maximum rather than np.max?

Yes you are right! np.maximum is the right one.