Added code to handle multi-graph in mst

[Aditya-Shandilya1182]

No description provided.

[Aditya-Shandilya1182]

This code is written to address issue #7451.

[Aditya-Shandilya1182]

Code used to test the feature:

for g in nx.SpanningTreeIterator(nx.cycle_graph(3,create_using=nx.MultiGraph)):
    print(g.edges())

Result before changes:

Result after changes:

[mjschwenne]

The code itself looks pretty good, but you should definitely add some test cases using multi-graphs to the iterator. Things that are worth testing would be that the number of generated trees is correct, and then that the trees themself are presented in a non-decreasing (and non-increasing for the reverse mode). For a small tests, enumerating the trees explicitly can be a good sanity check. The current tests for the iterator should provide a good roadmap.

[Aditya-Shandilya1182]

I will try to add the test cases.

[eriknw]

I like this pattern! What do you think about using it above when setting d[partition_key]? This way, we wouldn't need separate for loops for if G.is_multigraph() and else.

[Aditya-Shandilya1182]

It looked neat this way.
Will it affect the performance of the code? I would love to hear your suggestions. I will make sure to change the code accordingly.