chimera_graph mxn lattice is flattened

[JoelPasvolsky]

Description I expected that chimera_graph(m) or chimera_graph(m,m) would have be a m-by-m lattice but instead am getting a row of length mxm.

To Reproduce

>>> c4 = dnx.chimera_graph(4)
>>> c4.nodes
NodeView((0, 4, 5, 6, 7, 1, 2, 3, 32, 36, 37, 38, 39, 33, 34, 35, 64, 68, 69, 70, 71, 65, 66, 67, 96, 100, 101, 102, 103, 97, 98, 99, 8, 12, 13, 14, 15, 9, 10, 11, 40, 44, 45, 46, 47, 41, 42, 43, 72, 76, 77, 78, 79, 73, 74, 75, 104, 108, 109, 110, 111, 105, 106, 107, 16, 20, 21, 22, 23, 17, 18, 19, 48, 52, 53, 54, 55, 49, 50, 51, 80, 84, 85, 86, 87, 81, 82, 83, 112, 116, 117, 118, 119, 113, 114, 115, 24, 28, 29, 30, 31, 25, 26, 27, 56, 60, 61, 62, 63, 57, 58, 59, 88, 92, 93, 94, 95, 89, 90, 91, 120, 124, 125, 126, 127, 121, 122, 123))

This gives:

Expected behavior I expected:

Environment:

• OS: WIN10
• Python version: 3.7.0

Additional context Add any other context about the problem here.

[boothby]

I think this is a problem with the inspector and not the chimera graph generator.

import dwave_networkx as dnx
from matplotlib import pyplot as plt

c4 = dnx.chimera_graph(4, 4)
plt.figure(figsize=(5,5))
dnx.draw_chimera(c4)
plt.savefig('c4.png')

[JoelPasvolsky]

@boothby, the output of dnx.chimera_graph(4) includes node 40, 44, 45, 46, 47, 41, 42, 43. I had expected C4 nodes to be [0, 1 .. 8x4, 8x16, 8x16+1, ... 8x20, 8x20+1, ...] But I guess that assumes C4 as a subgraph of C16. Thanks!

[boothby]

You can get that subgraph as follows:

g = dnx.chimera_graph(16)
c4_nodes = [v for v, d in g.nodes(data=True) if all(x < 4 for x in d['chimera_index'][:2])]
c4 = g.subgraph(c4_nodes)

[JoelPasvolsky]

Thanks, @boothby, I just changed the first line to g = dnx.chimera_graph(16, node_list=qpu.nodelist, edge_list=qpu.edgelist) to account for available qubits & couplers, and that works perfectly: