cgearhart

def sample(weights, size=1):
    beta = 0
    samples = []
    w_max = max(weights)
    index = random.randint(1, N)
    while len(samples) < size:
        beta += random.uniform(0, 2 * w_max)
        while w[index] < beta:
            beta -= weights[index]
            index += 1

cgearhart

def mst(graph):
    '''
    mst retuns the minimum spanning tree of `graph`, which must be a symmetric NxN array, where 
    element (i,j) is the weight of the connection from node i to node j. The diagonal should be all 
    zeros (because i and j are the same point), and unconnected nodes are `None`.
    
    The nth position of the array returned contains the parent node of node n. The order always
    starts from node 0 as the root of the tree, so order[0] is always None.
    '''
    order = [None] * len(graph)