dfnGraph (v2.8) fails to identify the backbones with bipartite approach considering weight (length and hydraulic resistence)

[yingtaohu]

Hello all,

When i run the example "pruning" to generate bipartite graphs, there is an error:

So i change the code in "pruning.py":
"path_nodes = sorted(list(k_shortest))--------> path_nodes =list(k_shortest)"
Then the running is OK.

However, if i try to use "greedy_edge_disjoint" method to identifying the backbones of DFN, here is another issue:

So i change the code in "pruning.py" again:
"if G.graph['representation'] != "intersection":--------> if G.graph['representation'] != "bipartite":"
Then the running is OK. Is it right?

These results are good for me. However,how can i identify the backbones with bipartite graph-based approach considering weight (length and hydraulic resistence) ? Is there any open source codes ? I cann't find it. So i cann't repeat the results in:
Hyman et al. 2018 "Identifying Backbones in Three-Dimensional Discrete Fracture Networks: A Bipartite Graph-Based Approach" SIAM Multiscale Modeling and Simulation.

Thanks in advance!

[hymanjd]

Hi, the greedy disjoint path algorithm requires that you use the intersection graph, not the bipartite.

All source code is in the pydfnworks/dfnGraph subdirectory

[yingtaohu]

Thanks, Jeffrey.