def write_expected_deg(G, n, d_mean, D, times):
# for obs_degrees,
# key = degree
# value = freq
print "iterating"
for i in xrange(times):
model.iterate(G)
print "storing degree informations"
obs_degrees = {}
for x in G.nodes_iter():
if G.degree(x) in obs_degrees:
obs_degrees[G.degree(x)] += 1
else:
obs_degrees[G.degree(x)] = 1
outstring = "degree\tfrequency\n"
for key in obs_degrees.iterkeys():
outstring += "%s\t%d\n" % (key, obs_degrees[key])
# save data
print "saving data"
outfile = open("observed_degree_{0}n_{1}k_{2}D.dat".format(n, d_mean, D), "w")
outfile.write(outstring)
outfile.close()
def write_expected_deg(G, n, d_mean, D, times):
#for obs_degrees,
#key = degree
#value = freq
print "iterating"
for i in xrange(times):
model.iterate(G)
print "storing degree informations"
obs_degrees = {}
for x in G.nodes_iter():
if G.degree(x) in obs_degrees:
obs_degrees[G.degree(x)] += 1
else:
obs_degrees[G.degree(x)] = 1
outstring = "degree\tfrequency\n"
for key in obs_degrees.iterkeys():
outstring += "%s\t%d\n" % (key, obs_degrees[key])
#save data
print "saving data"
outfile = open('observed_degree_{0}n_{1}k_{2}D.dat'.format(n, d_mean, D),
'w')
outfile.write(outstring)
outfile.close()
def write_edge_existence(G, iterations, times, n, d_mean, D, sample_rate,
throw_away):
#put this first so that the "Exiting..." occurs before
#any time is spent on lots of computing
if 'c' in G.__dict__:
edge_expected_freq = expected_edge_freq(G)
else:
print "G.c not initalized. Exiting..."
return
#edge_obs_freq keeps track of how often a specific edge appears
#key = edge
#value = frequency
edge_obs_freq = {}
print "Collecting data on edge existence"
for i in xrange(times):
if i % (times / 10) == 0:
print "%d percent" % (100 * i / times)
#Method 1
remake_graph(G)
#Method 2
#model.reconstruct(G, d_mean)
number_of_samples = 0
for j in xrange(iterations + throw_away):
if j >= throw_away and j % sample_rate == 0:
number_of_samples += 1
for edge in G.edges_iter():
if edge in edge_obs_freq:
edge_obs_freq[edge] += 1
else:
edge_obs_freq[edge] = 1
model.iterate(G)
#normalize observed frequencies
for key in edge_obs_freq.iterkeys():
edge_obs_freq[key] = float(
edge_obs_freq[key]) / (number_of_samples * times)
#save the data
outstring = "edge\tdistance\tobs_freq\texp_freq\n"
for key in edge_obs_freq.iterkeys():
outstring += "%s\t%f\t%f\t%f\n" % (key, model.length(
G, key), edge_obs_freq[key], edge_expected_freq[key])
outfile = open(
'edgeExistence_{0}n_{1}k_{2}d_{3}times_{4}iterations.dat'.format(
n, d_mean, D, times, iterations), 'w')
outfile.write(outstring)
outfile.close()
def write_edge_existence(G, iterations, times, n, d_mean, D, sample_rate, throw_away):
#put this first so that the "Exiting..." occurs before
#any time is spent on lots of computing
if 'c' in G.__dict__:
edge_expected_freq = expected_edge_freq(G)
else:
print "G.c not initalized. Exiting..."
return
#edge_obs_freq keeps track of how often a specific edge appears
#key = edge
#value = frequency
edge_obs_freq = {}
print "Collecting data on edge existence"
for i in xrange(times):
if i % (times/10) == 0:
print "%d percent" % (100 * i / times)
#Method 1
remake_graph(G)
#Method 2
#model.reconstruct(G, d_mean)
number_of_samples = 0
for j in xrange(iterations + throw_away):
if j >= throw_away and j % sample_rate == 0:
number_of_samples += 1
for edge in G.edges_iter():
if edge in edge_obs_freq:
edge_obs_freq[edge] += 1
else:
edge_obs_freq[edge] = 1
model.iterate(G)
#normalize observed frequencies
for key in edge_obs_freq.iterkeys():
edge_obs_freq[key] = float(edge_obs_freq[key]) / (number_of_samples*times)
#save the data
outstring = "edge\tdistance\tobs_freq\texp_freq\n"
for key in edge_obs_freq.iterkeys():
outstring += "%s\t%f\t%f\t%f\n" % (key, model.length(G, key),
edge_obs_freq[key], edge_expected_freq[key])
outfile = open('edgeExistence_{0}n_{1}k_{2}d_{3}times_{4}iterations.dat'.format(n, d_mean, D, times, iterations), 'w')
outfile.write(outstring)
outfile.close()
times = 5*10**4
stats_size = 1
alpha = 2
outstring = 'N\tMean Pair-Wise Distance\n'
print 'Working on ',
print 'opPairwise_{0}k_{1}D_{2}alpha_{3}min_{4}max_{5}times_{6}size.dat'.format(k, D, alpha, n_min, n_max, times, stats_size)
print 'N\tMean Pair-Wise Distance'
for n in xrange(n_min, n_max + n_step, n_step):
print "Starting n = %d" % n
pairwise_total = 0
for j in xrange(stats_size):
G = model.make_opinion_graph(n, .5 * n * k, D, alpha)
for i in xrange(n*times):
model.iterate(G)
giant_component = nx.connected_component_subgraphs(G)[0]
iter_pw_distance = 0 #diameter may not be diameter
for vertex_one in giant_component.nodes():
for vertex_two in giant_component.nodes():
if vertex_one == vertex_two:
break
iter_pw_distance += nx.shortest_path_length(
giant_component, vertex_one, vertex_two)
buf = giant_component.number_of_nodes()
pairs = buf * (buf - 1) / 2
pairwise_total += float(iter_pw_distance) / pairs
k = 4
D = 1
times = 1*10**3
stats_size = 50
alpha = 2
outstring = 'N\tDiameter\n'
print 'N\tDiameter'
for n in xrange(n_min, n_max + n_step, n_step):
print "Starting n = %d" % n
diam_total = 0
for j in xrange(stats_size):
G = model.make_opinion_graph(n, .5 * n * k, D, alpha)
for i in xrange(n*times):
model.iterate(G)
subgraphs = nx.connected_component_subgraphs(G)
max_diameter = 0 #diameter may not be diameter
for SG in subgraphs: #of largest connected subgraph
if SG.number_of_nodes() < 2:
break
diameter = nx.diameter(SG)
if diameter > max_diameter:
max_diameter = diameter
diam_total += max_diameter
outstring += '{0}\t{1}\n'.format(n, diam_total / float(stats_size))
print '{0}\t{1}'.format(n, diam_total / float(stats_size))