absolute difference of NxN square matrix diagonals

matrix.py

#! /usr/bin/env python

"""
abs_diag_sum
a function to find the absolute difference of the sum of diagonals of an NxN square matrix

where the square matrix is represented as list of lists

such that the square matrix

2 5 6
7 5 8
9 6 7

would be represented as [[2, 5, 6], [7, 5, 8], [9, 6, 7]]

and the difference of sum of diagonals then would be
(2 + 5 + 7) - (6 + 5 + 9) = 6

@Alloys Mila
"""

def abs_diag_sum(matrix):
    if type(matrix) is list:

        # initialize the sum of each diagonal to zero
        primary_diagonal = 0
        secondary_diagonal = 0

        for i in range(len(matrix)):
            row = matrix[i]

            # check that the length of the row is equal to the lenth of the matrix
            # because for an NxN square matrix rows = columns
            if len(row) is len(matrix):
                # bump the sum
                primary_diagonal = primary_diagonal + int(row[i])

                secondary_marker = (len(row) - i) - 1
                secondary_diagonal = secondary_diagonal + int(row[secondary_marker])
            else:
                print('make sure to provide a valid square matrix')

        return abs(primary_diagonal - secondary_diagonal)
    else:
        print('expecting the input to be a list got ' + type(c) + ' instead')

# test
sum = abs_diag_sum([[2, 5, 6], [7, 5, 8], [9, 6, 7]])
print(sum)