Spanning tree algorithms as statistics?

[phreakocious]

I would like to add the ability to run spanning tree/forest algorithms against a Gephi graph.. In particular, I'd like to implement:

Chu-Liu/Edmonds (directed graphs)
Dijkstra
Kruskal

My question relates to what to do with the output of the function? It seems that adding new attributes for each edge and node would be appropriate, and distinction between the output of the different algorithms would be needed. Would attribute names like these work?

CLE_NodeRoot Boolean // Is this the root node of the tree
CLE_NodeTree Boolean // Is this node part of the tree
CLE_EdgeWeight
CLE_EdgeTree
etc.

Thanks!

[halim]

Hi
did you manage to add spanning tree algo to Gephi. if so i would love to know how!
thank you
halim

[phreakocious]

I was hoping to get some feedback from the development community before putting the time into it.

After poking around through the source, it doesn't look like it would be very difficult to implement this. There is already code for it in the source, but the results are not made visible to the user, nor are they saved in projects.

I'm a bit stuck with Gephi development at the moment as I want to work on the beta version, but am using the Neo4j options from the alpha and don't want to reconcile the differences manually.

[admin]

huh, with a long delay but...

Yes these attribute names would work.

The Dijkstra algorithm implemented is used internally, so it's normal not to produce an output. However you can duplicate and adapt it inside a Metric Plugin, as I think you already did.

You can ask the Neo4j developer, Martin Skurla, to synchronize the source code with the trunk so the merge will be easy. Please ask development question on the gephi-plugin mailing list.

[mbastian]

Actually, the Neo4j branch is already synced with the 0.7 beta code. I recently did the dirty work on merging these two things You will get the most recent version of the development, both for Gephi and the Neo4j plug-in.