You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on May 1, 2024. It is now read-only.
As the https://pypi.org/project/metis/ project supports networkx multigraphs, it should be possible.
That project however doesn't support vertex cuts and its bitbucket was taken down.
The text was updated successfully, but these errors were encountered:
Both libraries suggest to cut (2,1) and pretend its only one cut.
In the case of the vertex separator however, I think the current approach to throw an exception in the case of a MultiGraph isn't really necessary.
I don't know much about how powerful metis is here, but as the cost of a vertex separator in the normal case is the number of vertices (or the sum of their weights) we should just take the MultiGraph, ignore multi edges and continue.
I removed the @nx.utils.not_implemented_for('multigraph') and that was enough for my current task.
If you know some task where metis needs multi edge information for a vertex separator that would be nice to know.
As the https://pypi.org/project/metis/ project supports networkx multigraphs, it should be possible.
That project however doesn't support vertex cuts and its bitbucket was taken down.
The text was updated successfully, but these errors were encountered: