def simulate():
N = 100
R = 5
b1 = 1.5
b0 = 0.5
a = 1.0
alpha0 = a * np.ones(R)
SBM_1 = StochasticBlockModel(R, b0, b1, alpha0)
(A, f, theta) = sample_network(SBM_1, N)
(B, pi) = theta
p = 0.3
q = 0.7
N = 100
R = 5
####run once on mac for picked SBM data
####start-----------pickle string won't keep the graph
graph_1 = Gen_SBM(p, q, N, R)
print graph_1
Graph.write_pickle(graph_1, "SBM_03_07_1000_5.p")
####end-------------
SBM_1 = Graph.Read_Pickle("SBM_03_07_1000_5.p")
SBM_2 = sparsify.spars_combi(SBM_1, epsilon=0.5)
def sim(eps, GraphName):
SBM_expanded = Graph.Read_Ncol(GraphName + "_expanded.txt", directed=False)
w_sum1 = sum(SBM_expanded.es["weight"])
print w_sum1
e_num1 = SBM_expanded.ecount()
SBM_2 = sparsify.spars_combi(SBM_expanded, epsilon=eps)
# through away weights for 3-hop recalculation, write into a temparary file
SBM_2.write_ncol(GraphName + "_sparse_list.txt", weights=None)
Expand(GraphName + "_sparse")
SBM_recalc = Graph.Read_Ncol(GraphName + "_sparse_expanded.txt", directed=False)
w_sum2 = sum(SBM_recalc.es["weight"])
# need to modify! calculate real weight instead
print w_sum2
e_num2 = SBM_2.ecount()
return (w_sum1 - w_sum2), (e_num1 - e_num2)
