networkx 2.5 breaks

[lamelizard]

I tried to run the code in the readme:

import networkx as nx
import graphwfc
GI = nx.Graph([(1,2),(2,3),(3,4)])
GI.add_nodes_from([(1,{'c':1}),(2,{'c':1}),(3,{'c':2}),(4,{'c':3})])
GL = nx.DiGraph([(1,2)])
GO = nx.random_tree(1000)

But it raises the following networkx error: AttributeError: 'Graph' object has no attribute 'pred'

I'm using networkx==2.5

Originally posted by @asher-pembroke in https://github.com/lamelizard/GraphWaveFunctionCollapse/issues/1#issuecomment-735341081

[lamelizard]

Additionally: python -m graphwfc -v value-examples break in 2.5 but work in 2.4 (it breaks in nx.read_graphml) the example code does not even work in 2.1...

[lamelizard]

I uploaded a new version; while the example code should work now with networkx 2.5, somehow loading of files breaks on 2.5 but works on 2.4. So for the time being we will require <=2.4, I hope that's not an issue for anybody(?). However, this needs a fix eventually.

@asher-pembroke, would you mind trying out the new version to see if everything works now? It works on my machine(TM), but feedback would be appreciated nonetheless.

[asher-pembroke]

@lamelizard 2.4 works for me! I'm on windows, btw. I'm fascinated by WFC and am excited to see how it could be used in the graph context. Once I've post my examples once played around with it a little.

Thanks for building this!

[lamelizard]

Thank you, that is great to hear. I'll keep this issue open until I can fix the loading issue.