Does google's pagerank algorithm on a directed graph. Requires networkx for the graphs.

pagerank.py

try:
    import networkx as nx
except ImportError:
    raise ImportError('Requires networkx')
    
def pageranker(g, damp, iterations):
    #Does this for a set amount of iterations. 
    #In the future, may add convergence conditions.
    #Not optimal for large graphs. Should probably store the connected_to list somewhere instead of calling it each time.
    dict_of_ranks = {}
    prior_values_of_rank = {}
    for i in g.nodes():
        dict_of_ranks[i] = 1
        prior_values_of_rank[i] = 1
    for k in range(iterations):
        for i in g.nodes():
            pagerank_of_node = 0
            #PR(x) = (1-d) + d*(PR(T1)/C(T1) + ... + PR(TN)/C(TN)) for T1,...,TN connected to x. 
            #PR = pagerank
            #C(Ti) = outdegree of Ti
            if len(connected_to(g, i)) == 0:
                pagerank_of_node = (1-damp)
            else:
                for j in connected_to(g, i):
                    pagerank_of_node += (dict_of_ranks[j]/g.out_degree(j))
                    pagerank_of_node = pagerank_of_node * damp
                    pagerank_of_node += (1-damp)
            prior_values_of_rank[i] = pagerank_of_node
        for i in prior_values_of_rank:
            dict_of_ranks[i] = prior_values_of_rank[i]
    for i in dict_of_ranks:
        dict_of_ranks[i] = dict_of_ranks[i]/len(g.nodes())
    return dict_of_ranks

def connected_to(g, node):
    list_of_connected = []
    for i in g.nodes():
        if node in g[i]:
            list_of_connected.append(i)
    return list_of_connected

Note this gives an estimate, it does not give an accurate value. Leaving room to further update to make it more accurate.