Descripency in number of edges between "iterEdges", "forEdges", and "NumberOfEdges" after sparsification

[m-eghbali]

Hi,

I recently started using Networkit. I am working on an example in which I try to sparsify an undirected graph using the following code.

G = nk.GraphFromCoo((edge_coo[0], edge_coo[1]), directed= bool(is_graph_directed) , weighted=False, n=graph.num_nodes)
G.indexEdges()
scansp = nk.sparsification.SCANSparsifier()
scanG = scansp.getSparsifiedGraphOfSize(G, 0.7)

After running this code, the number of edges from scanG.numberOfEdges(), scanG.iterEdges(), and scanG.forEdges() is significantly different.
I was wondering if I'm missing something or if there is a known issue.

Spec:
networkit 11.0
torch 2.0.1
torch-geometric 2.3.1

Thanks

[fabratu]

Hi and sry for the late reply. So far, I wasn't able to reproduce your issues. Since your testcase doesn't specify the underlying dataset, I tried the above code with simple graphs, graphs with self loops, self loops + multi edges and directed graphs.

For all these cases (based on generated instances), the output of scanG.numberOfEdges(), scanG.iterEdges(), and scanG.forEdges() is accounting for the same edges.