Steiner tree heuristic using networkx

[Brian]

Hi everyone - 

Just wanted to see if anyone had developed anything that can create steiner trees using networkx. Is there any plan to add this functionality to networkx? 

I'm aware that some other options exist such as GOBLIN and GeoSteiner, but I was hoping there was something available within the networkx ecosystem, as it's very easy to use!

Thanks

[Moritz Beber]

Hi Brian,

Google turns up this: https://github.com/xlu29466/PyGibbCAMP/blob/master/SteinerTree.py but I'm not aware of any such implementation directly in networkx. However, steiner trees are based on the concepts of minimum spanning trees (https://networkx.github.io/documentation/latest/reference/algorithms.mst.html), minimal cuts and maximum flows (https://networkx.github.io/documentation/latest/reference/algorithms.flow.html).

I know this is extremely vague but I hope it can help you get started.

Best,

Moritz

[Brian]

Thanks, Moritz. I was able to find that script as well, but had some trouble getting what I needed out of it. I only have training for electrical engineering, so I'm really new to all this graph/network/computer science stuff. I really like networkx, though! If I'm able to get any Steiner-stuff working with networkx, I will be sure to share it with the community! I was considering trying to create my own implementation of the methods used for GeoSteiner, but that might prove to be too ambitious for me...  We'll see. 

Thanks again for the suggestions!