How to find the maximum/minimum offset in np.diagonal()?

I want to get back all the diagonals of a numpy matrix in the form of a list and np.diagonal() does that but slice notation doesn't work on it.

I tried using slice notation, thinking by just putting a semicolon in there it would return what I wanted, but it did not work.

Example:

This is my matrix

73167176531330624919225119674426574742355349194934
96983520312774506326239578318016984801869478851843
85861560789112949495459501737958331952853208805511
12540698747158523863050715693290963295227443043557
66896648950445244523161731856403098711121722383113
62229893423380308135336276614282806444486645238749
30358907296290491560440772390713810515859307960866
70172427121883998797908792274921901699720888093776
65727333001053367881220235421809751254540594752243
52584907711670556013604839586446706324415722155397
53697817977846174064955149290862569321978468622482
83972241375657056057490261407972968652414535100474
82166370484403199890008895243450658541227588666881
16427171479924442928230863465674813919123162824586
17866458359124566529476545682848912883142607690042
24219022671055626321111109370544217506941658960408
07198403850962455444362981230987879927244284909188
84580156166097919133875499200524063689912560717606
05886116467109405077541002256983155200055935729725
71636269561882670428252483600823257530420752963450

I want it to return a list like this:

[[7],[0,1],[8,5,6],[0,4,8,3]...[3,3],[4]]

Though I can count it, I want to know if there is a way to tell the max/min offset of np.diagonal. The matrix's size may vary..

You can use a list comprehension:

import numpy as np

arr = np.random.randint(10, size=(4, 4)) # Your array here
h, w = arr.shape

out = [ np.diag(arr, k) for k in range(-h+1, w) ]

>>> arr
array([[9, 6, 5, 8],
       [1, 5, 1, 2],
       [7, 0, 6, 5],
       [2, 6, 9, 8]])

>>> out
[array([2]), array([7, 6]), array([1, 0, 9]), array([9, 5, 6, 8]), array([6, 1, 5]), array([5, 2]), array([8])]

</details>