Python nested loop where the second loop is always one object ahead of the first loop

Question

I am trying to optimize a node data structure of a graph/network by avoiding duplicate edges. The graph is undirected and has no loops (in the sense that no node has edges to itself). Each node data structure contains a list of edges. To this end, I have the following function for adding an edge between two nodes of the graph:

"""
Creates an edge between `node1` and `node2`.
If i < j and there is an edge between node i and node j, then 
the edge between node i and node j will be stored in node i.
"""
def create_edge(self, node1, node2):
    if node1.id < node2.id:
        node1.edges.append(node2.id)
    else:
        node2.edges.append(node1.id)

And then, as part of the instantiation of the graph/network object, the following code is executed to create edges between nodes, based on some probability:

    for node1 in self.green_network:
        for node2 in range(node1 + 1, self.green_network):
            if random.random() < self.probability_of_an_edge:
                node1.create_edge(node2)

Now, I realise this second segment of code is incorrect, but I think it roughly illustrates what I am trying to do. Given my stated goal, it seems to me that, for every node in the first loop, for the second loop, I want to loop from (every node + 1) to the last node. So, for instance, we first loop between node.id = 1 to node.id = 2, node.id = 3, node.id = 4, …, and then between node.id = 2 to node.id = 3, node.id = 4, node.id = 5, …, and then between node.id = 3 to node.id = 4, node.id = 5, node.id = 6, …, and so on. That way, we avoid duplicates, given how the create_edge function behaves.

This is related, but it doesn’t seem to answer the question of iterating over objects, which is what I’m trying to do here.

Asked By: The Pointer

||

Source

Answer 1

As suggested in the comments, you could/should generate all the combos and consider them with equal probability.

from itertools import combinations
from random import random

class Node:
    count = 0

    def __init__(self):
        self.edges = []
        self.idx = Node.count
        Node.count += 1

    def add_edge(self, other: "Node"):
        # this adds an edge to BOTH nodes
        # if you *really* want, you could compare indices here and only add one edge... Buyer be ware! :)
        self.edges.append(other.idx)
        other.edges.append(self.idx)

# make 5 nodes
nodes = [Node() for _ in range(5)]

# this will produce a GENERATOR for all 10 possible combinations (5 choose 2 = 10)
all_pairs = combinations(nodes,2)

for n1, n2 in all_pairs:
    if random() > 0.25:  # 75% chance of addition...
        n1.add_edge(n2)

# check
for node in nodes:
    print(f'Node {node.idx} connects to: {node.edges}')

Output:

Node 0 connects to: [3]
Node 1 connects to: [2, 4]
Node 2 connects to: [1]
Node 3 connects to: [0, 4]
Node 4 connects to: [1, 3]

Answered By: AirSquid

Python nested loop where the second loop is always one object ahead of the first loop

Question:

Answers:

Output: