mhenderson / listcolouring

List colouring in Python with NetworkX.

listcolouring

List colouring in Python with NetworkX.

import networkx as nx
import matplotlib.pyplot as plt

import listcolouring
from listcolouring import list_init, greedy_list_edge_colouring, print_list_edge_colouring

G = nx.petersen_graph()

G = list_init(G, range(0, 10), 3, 0)
G = greedy_list_edge_colouring(G)
    
options = {'with_labels': True, 'node_color': "white"}
colors = nx.get_edge_attributes(G,'colour').values()
nx.draw_shell(G, nlist = [range(5, 10), range(5)], edge_color = colors, **options)

plt.savefig("img/petersen-shell.png", format = "PNG"

listcolouring's Issues

Create a release

In other words, add a NEWS.md

old code

import networkx as nx

G = nx.petersen_graph()

d = nx.coloring.greedy_color(G, strategy="largest_first")

print(G.nodes)
print(G.edges)

G[0][1]["colour"] = 4

print(G[0][1])
print(G[0])

print(d)

G[0][1]["permissible"] = {1, 2, 3}
print(G[0][1])

print(G[0][1]["colour"] in G[0][1]["permissible"])

GL = nx.line_graph(G)

print(GL)

print(nx.inverse_line_graph(GL))

print(GL.nodes)
print(GL.edges)`

Add docstrings.

Turn existing comments into docstrings and add docstring for other functions.

Publish package on PyPi

This would allow me to use Colab to create my notebook, I think.

Lists have a fixed size of 3.

There's two aspects to this issue.

One is just that 3 is hard-coded when generating a random list assignment.

But we would also like to have lists of various sizes.

Write a function to initialise a list-colouring instance

Split the code for generating list colouring instances from the implementation of greedy list colouring.

Create Sphinx documentation.

There should be a docs folder. See the psssodls project for an example.

Write a list-edge-colouring script

The input is a list edge colouring instance (assuming that some file format supports attributes).

Remove example from README

Create a notebook link instead.

Add README badges.

PyPi and packaging workflow.

Allow partial colourings.

At least it probably should be the case that the main loop iterates over only those edges with a colour of None. That way we can attempt to greedily finish colouring a partially coloured graph.

Create a Jupyter notebook demo.

Setup Read the Docs.

See psssodls for an eggsample.

Write a function to apply a greedy list-edge-colouring algorithm to a list-edge-colouring instance