Converting adjacency list to adjacency matrix in python

Question

Trying to write a code to convert a graph representation from adjacency list to adjacency matrix.

The function should input an adjacency list adj_list = [[1,2],[2],[],[0,1]] and output the binary matrix

adj_mat = [0. 1. 1. 0]
          [0. 0. 1. 0.]
          [0. 0. 0. 0.]
          [1. 1. 0. 0.] 

However running the code

 def adj_list_to_matrix(adj_list):
    n = len(adj_list)
    adj_matrix = np.nan * np.ones((n,n))
    np.fill_diagonal(adj_matrix,0)

    for i in range(n):
        for j, w in adj_list[i]:
            adj_matrix[i,j] = w
    return adj_matrix

produces the error message

for j, w in adj_list[i]:
TypeError: cannot unpack non-iterable int object 

Can anyone help resolve this issue to get a working code?

Answer 1

for j, w in adj_list[i]: doesn't work because adj_list[i] is only a list, so you can only unpack one value out of it in a for loop. It looks like you want w to be a weighting factor, but your adjacency list doesn't have any weightings.

You can do this, assuming all the weights are 1 (I think this is what you want based on your expected output in the question).

def adj_list_to_matrix(adj_list):
    n = len(adj_list)
    adj_matrix = np.nan * np.ones((n,n))
    np.fill_diagonal(adj_matrix,0)

    for i in range(n):
        for j in adj_list[i]:
            adj_matrix[i,j] = 1
    return adj_matrix

output

[[ 0.,  1.,  1., nan],
 [nan,  0.,  1., nan],
 [nan, nan,  0., nan],
 [ 1.,  1., nan,  0.]])

Weighted adjacency matrix

If you want a weighted adjacency matrix, you need to put the weights in adj_list like so:

adj_list = [{1:0.2,2:0.5},{2:1},{},{0:0.1,1:0.6}]

Then your code only requires a minor tweak to work

def adj_list_to_matrix(adj_list):
    n = len(adj_list)
    adj_matrix = np.nan * np.ones((n,n))
    np.fill_diagonal(adj_matrix,0)

    for i in range(n):
        for j, w in adj_list[i].items():
            adj_matrix[i,j] = w
    return adj_matrix

output

[[0. , 0.2, 0.5, nan],
 [nan, 0. , 1. , nan],
 [nan, nan, 0. , nan],
 [0.1, 0.6, nan, 0. ]]