Question

I’m building a graph class represented as a dict. The graph class is used to perform path finding on a large 2D grid matrix. The nodes are stored as keys and the neighbour nodes are stored as values.

This works nice for fast path searches but I also need to get specific nodes out of the dict determined by their x and y coordinates.

class Node(object):
    def __init__(self, x, y):
        self.x = x
        self.y = y

class Graph(dict):
    def get_node_at(self, x, y):
        for node in self:
            if node.x == x and node.y == y:
                return node

    def set_neighbours(self):
        pass

graph = Graph()

# build a 2d matrix of nodes
for x in range(10):
    for y in range(10):
        graph[Node(x,y)] = []

# set the neighbour nodes for each node
graph.set_neighbours()
mynode = graph.get_node_at(6,6) # will later be executed quite often

Is there a way to optimize the graph class so the get_node_at method performs better on a large grid?

Answer 1

add an index to Graph which would look like {(x,y):node}. this would require implementing __setitem__ to add the Node to the index, and remove it if another Node already exists and __delitem__ would also remove the Node from the index. this would allow get_node_at(self, x, y) to just do return self.index[(x,y)].