hypergraphx.dynamics.transition_matrix not working when edges are not ordered from 0 to N

[larryzhang95]

I was trying to use the random walk function which depends on the transition_matrix matrix function in the dynamics submodule.

I got the error:

IndexError: index 492 is out of bounds for axis 0 with size 92

when I ran adjacency_matrix(H) where H is the hypergraph from 'test_data/workplace

I then dug further and noticed that in the hedge_list list in transition_matrix, the first edge printed out is (492, 938).

For the code in transition_matrix:

    for l in hedge_list:
        for i in range(len(l)):
            for j in range(i+1, len(l)):
                T[l[i], l[j]] += len(l) - 1
                T[l[j], l[i]] += len(l) - 1

T[l[i],l[j]] , would index T[492, 938]. T is only size 92 x 92, and thus this would be invalid.

My guess is that an easy fix for this is to find the index of the value l[i] and l[j] would solve this. I will try it on my end and report back

[larryzhang95]

    for l in hedge_list:
        for i in range(len(l)):
            for j in range(i+1, len(l)):
                T[nodes.index(l[i]), nodes.index(l[j])] += len(l) - 1
                T[nodes.index(l[j]), nodes.index(l[i])] += len(l) - 1

Changing to this did work, though I would test it further just in case

[larryzhang95]

I also added a nodes variable :

nodes = HG.get_nodes()

[FraLotito]

Thanks for reporting this! Yes, it seems that we are missing a mapping from nodes to integers ids in [0, N) in that functions, I'll fix that soon.