Retrieving residual capacities without networkx

[paredesroger]

Not sure if the issue tracker is the right place to address my question. Basically, I need to run a maximum flow algorithm several times but I am interested not only in the value of the maximum flow (or size of the cut) but also in the flow graph (residual capacities would suffice). PyMaxflow is very fast when it comes to compute the maximum flow but using get_nx_graph(self) to retrieve the residual capacities becomes a bottleneck in my code. Is there anyway I can access to the residual capacities without using get_nx_graph(self)?

Thank you so much!

[pmneila]

Hi,

It's the right place.

Unfortunately, there is no way to get the residual capacities of the graph other than get_nx_graph. I'm going to try to implement a faster access to these capacities at some point in the next few weeks, but you will have to live with get_nx_graph until then, sorry.

[paredesroger]

For now is ok and actually following this process is still faster than using only networkx so I'm glad that I came across PyMaxflow. However, the implementation to access to the residual capacities would definitely augment the functionality of your library and I bet that I am not the only one that would benefit from it.

Thank you so much for your help!

[davisv7]

Bump.