nkoub/multinetx

multiNetX is a python package for the manipulation and visualization of multilayer networks. It is build on NetworkX

multinetx's Issues

multilayer network using erdos_renyi_graph

Hello Sir,

I am facing problem while running the below given code of first example given for creating multilayer network. I have successfully installed the library.
import multinetx as mx
import numpy as np
g1 = mx.erdos_renyi_graph(100,0.5,seed=218)
g2 = mx.erdos_renyi_graph(100,0.6,seed=211)
g3 = mx.erdos_renyi_graph(100,0.7,seed=208)
adj_block = mx.lil_matrix(np.zeros((N3,N3)))

The error while runnnng above code is
g1 = mx.generators.erdos_renyi_graph(100,0.5,seed=218)

AttributeError: module 'multinetx' has no attribute 'generators'

Please help.

Support for networks with different number of nodes in each layer

Hi ,
Using this code for MultilayerGraph from " https://github.com/InesCarrasco/multinetx/blob/master/classes/multilayer.py " is giving the error "TypeError: 'list' object cannot be interpreted as an integer" at pos = mx.get_position( ... )
Please advice.

Upgrade dependencies

I see multinetx requires networkx==2.3 but now 3.2.1 is available. Due to changes in some dependencies of networkx too, new python versions >3.9 have issues with old versions of networkx, which requires updating and that causes conflict with multinetx.

https://stackoverflow.com/questions/66174862/import-error-cant-import-name-gcd-from-fractions

Issue with "projection" in utilities.py

I am using the projection function on a MultilayerGraph object so that I can use the NetworkX modularity function in my code. In Python 3, has_key has been removed. As a result, line 179 (if multinet.edges.get(e).has_key("weight") and edge_weight:) throws an attribute error even though multinet.edges.get(e) does in fact yield a dictionary object, which in my case has weight as a key. Perhaps this part could be rewritten so that it is up to date with Python 3 conventions?

edit error in get_position

in base/draw.py line 71:
it should be base_pos = nx.layout.circular_layout(G.get_layer(0))
not base_pos = nx.layout.circylar_layout(G.get_layer(0))

keep original node labels after Multilayergraph

Hi Nikos,

Thanks for the awesome package.
I am trying to learn this tool using my own edgelists as input layers.
Here's my test:

cluster1 = nx.Graph()
cluster1.add_edges_from([
    ("A", "B"),
    ("A", "C"),
    ("B", "C")])
cluster2 = nx.Graph()
cluster2.add_edges_from([
    ("D", "E"),
    ("F", "E"),
    ("D", "F")])
clusts = mx.MultilayerGraph(list_of_layers=[cluster1,cluster1])

fig = plt.figure(figsize=(15,5))
ax1 = fig.add_subplot(121)
ax1.axis('off')
mx.draw_networkx(clusts,
                # pos=pos1,
                 ax=ax1,
                # labels={},
                 with_labels=True)
plt.show()

However I noticed the node labels converted into integers after mx.MultilayerGraph()

How can I label the nodes with original names as "A", "B", "C", et al.

Thank you very much!

Best, Tao

Support for networks with different number of nodes in each layer

Seems like there have to be equally many nodes in each layer. A generalization, aloowing different number of nodes would be nice.

indentations

There may exist some indentations problems

I/O mechanisms?

Loading and Saving (outside of .pkl) and to maintain the class structure of multinetx vs. fall back on networkx? @nkoub

Compatibility with networkx 2.4+ would be helpful if possible!

Thanks for maintaining this excellent tool.

What is meaning of this parameter?

I want to draw a Multi layer network by using Python. The expected graph likes this:

This is my command:

mst_pearson,mst_kendall,mst_tail are my original networks, I want to use them to establish a three-layers multilayer network. But the error is that:

What is that error mean? How could I fix it?