draw_chimera_embedding does not plot the desired embedded graph

[jlaydevant]

I want to plot an embedded graph on a chimera lattice with the command dnx.draw_chimera_embedding() of a simple system of 5 or more nodes consecutively connected (but with no loop between the first and last node): J={(0, 1): 1, (0, 2): 1, (1, 3): 1, (2, 3): 1}

But when I use the drawing command it eventually plots some graphs where some units have more connections than expected and even some graphs are cycle despite the first and last unit should not be connected.

The embedded graph generated by minorminer.find_embedding() if fine so the issue is about the drawing function.

Here is the code to reproduce the issue:

layersList = [1,1,1,1,1]
h, J = {}, {}

comp0, comp1 = 0, layersList[0]

for i in range(len(layersList)-1):
    for k in range(layersList[i]):
        h.update({k + comp0:-1})
        for j in range(layersList[i+1]):
            J.update({(k + comp0, j + comp1):1})
            
    comp0 += layersList[i]
    comp1 += layersList[i+1]
    
for k in range(layersList[-1]):
    h.update({k + comp0:-1})
   
import dwave_networkx as dnx
# select basic graph on which move our computational graph
connectivity_structure = dnx.chimera_graph(2, 2)
    
from minorminer import find_embedding

import matplotlib.pyplot as plt

for k in range(10):
    plt.figure()
    emb = find_embedding(J, connectivity_structure)
    inter = [edge for edge in J]
    dnx.draw_chimera_embedding(connectivity_structure, emb, show_labels = True, interaction_edges = inter, overlapped_embedding = False)
plt.show()

I eventually found a way to overcome this issue on the file /dwave-networkx/dwave_networkx/drawing/qubit_layout.py line 292 where def show(p, q, u, v): return interactions.has_edge(p, q) should be changed to def show(p, q, u, v): return interactions.has_edge(u, v). Then with this changement the code works very fine.

[arcondello]

Hi @JeremLay , that does appear to be a bug. Would you like to make a PR with your proposed change?