def compute_and_measure_graphs(p):
graph = nx.DiGraph()
t_triples, t_cardinality = hdt.search_triples("", p, "")
print("amount of triples: ", t_cardinality)
for (l, p, r) in t_triples:
if p == "http://www.w3.org/2000/01/rdf-schema#subClassOf":
graph.add_edge(r, l)
# avc = nx.average_clustering(graph) # average clustering coefficient
# trans = nx.transitivity(graph)
# pearson = nx.degree_pearson_correlation_coefficient(graph)
reaching = nx.global_reaching_centrality(graph)
return reaching
# compute SCC
# mydict = {}
# for n in graph.nodes:
# collect_succssor = []
# for s in graph.successors(n):
# collect_succssor.append(s)
# mydict[n] = collect_succssor
# sccs = tarjan(mydict)
# filter_scc = [x for x in sccs if len(x)>1]
# # compute the counter for ths filtered SCC
# ct = Counter()
# for f in filter_scc:
# ct[len(f)] += 1
# print ('SCC info: ', ct)
# # obtain the corresponding scc graphs
# scc_graphs = []
# for s in filter_scc:
# g = graph.subgraph(s).copy()
# scc_graphs.append(g)
return filter_scc, scc_graphs
def func(G):
"""
The function being optimized by composite graphs.
"""
n = G.number_of_nodes()
return (logpaths(G) / n)**0.5 - 2 * nx.global_reaching_centrality(
G) - nx.average_clustering(G)
def draw_graph(g):
weight = [g[u][v]['weight'] / 5 for u, v in g.edges()]
# pos = nx.kamada_kawai_layout(g)
pos = nx.spring_layout(g)
# plt.subplot(111)
# nx.draw(g, pos, with_labels=True, width=weight)
# plt.show()
connectivity = nx.algebraic_connectivity(g)
centrality = nx.global_reaching_centrality(g)
# print(f'density = {nx.density(g):.4f}')
return connectivity, centrality
def erdos_renyi_grc(G):
grc_test = []
N = nx.number_of_nodes(G)
E = nx.number_of_edges(G)
prob_edges = ((N * (N - 1)) / 2) / E
for i in range(500):
er_graph = nx.erdos_renyi_graph(
N, prob_edges,
seed=10) #number of nodes, probability for edge creation
grc = nx.global_reaching_centrality(er_graph)
grc_test.append(grc)
print grc_test
def obtenerValores(dirigido, noDirigido):
# variables locales
datos = []
m = 0
c = 0
dm = 0
com = 0
# 1; orden - ambas
#print("orden")
datos.append(str(dirigido.order()))
# 2; tamaño - dirigida
#print("tamaño")
datos.append(str(dirigido.size()))
# 3; densidad, dirigida
#print("densidad")
datos.append(str(nx.density(dirigido)))
# 4; grado promedio - dirigido
#print("grado promedio")
datos.append(str((dirigido.size()) / (dirigido.order())))
# 5; diametro - no dirigido
#print("diametro")
datos.append(str(nx.diameter(noDirigido)))
# 6; radio - no dirigido
#print("radio")
datos.append(str(nx.radius(noDirigido)))
# 7; tamaño de clique mas grande - no dirigida
#print("clique mas grande")
datos.append(str(nx.graph_clique_number(noDirigido)))
# 8; numero de cliques maximales - no dirigida
#print("cliques maximales")
datos.append(str(nx.graph_number_of_cliques(noDirigido)))
# 9; global reaching centrality - dirigido
#print("reachability")
datos.append(str(nx.global_reaching_centrality(dirigido)))
# 10; clustering coefficient - dirigida
#print("clustering")
datos.append(str(nx.average_clustering(dirigido)))
# 11; transitividad - dirigida
#print("transitivity")
datos.append(str(nx.transitivity(dirigido)))
# 12; 13; 14; datos MODC: modularidad, dependencia minima, total de comunidades - no dirigido
#print("MODC")
(m, dm, com) = MODC(noDirigido, True)
datos.append(str(m))
datos.append(str(dm))
datos.append(str(com))
# fin de funcion
return (datos)
def features_vector(file): #self
base_list = name_base(path_pdb, file)
for i in range(len(base_list)):
print base_list[i]
G = input_nx(path_data, base_list[i])
L = laplacian_walk(G)
er_lap = erdos_renyi_lap(G)
H = load_hydrophobic(path_data, base_list[i])
C = load_charged(path_data, base_list[i])
A = alpha_content(path_pdb, base_list[i])
B = beta_content(path_pdb, base_list[i])
with open((feat_path + ('features_' + base_list[i] + '.txt')),
'w') as w:
#w.write(str(float(nx.number_of_edges(G))/float(nx.number_of_nodes(G)))+'\n') #removed avg degree
try:
w.write(
str(
nx.average_shortest_path_length(G) /
(((math.log(float(nx.number_of_nodes(G)))) - 0.5772) /
((math.log(float(nx.number_of_edges(G))) /
(float(nx.number_of_nodes(G)))) + 0.5))) +
'\n') #avg shortest path length
w.write(
str(
nx.diameter(G) /
(math.log((float(nx.number_of_nodes(G)))) / (math.log(
float(nx.number_of_edges(G)) /
float(nx.number_of_nodes(G)))))) + '\n') #diameter
w.write(
str(nx.radius(G) / nx.average_shortest_path_length(G)) +
'\n') #radius (min. shortest path)
w.write(
str(
nx.average_clustering(G) /
((float(nx.number_of_edges(G)) /
float(nx.number_of_nodes(G))) /
float(nx.number_of_nodes(G)))) +
'\n') #clustering coefficient of random graph
#***Number of quasi-rigid domains
w.write(str(nx.degree_assortativity_coefficient(G)) +
'\n') #assortativity coefficient
w.write(str(nx.global_reaching_centrality(G)) +
'\n') #nx.degree_centrality(G)
#***Residue intrinsic dimensionality (may be used to compute 6.?)
w.write(
str(L / er_lap) + '\n'
) #Normalized laplacian walk from function, N = D^{-1/2} L D^{-1/2} as well with ER
w.write(str(float(A) / (float(A) + float(B))) +
'\n') #alpha content
w.write(str(float(B) / (float(A) + float(B))) +
'\n') #beta content
w.write(
str((float(nx.number_of_edges(H)) /
float(nx.number_of_nodes(H))) /
(float(nx.number_of_edges(G)) /
float(nx.number_of_nodes(G)))) + '\n'
) #***Average degree of hydrophobic residues (F,M,W,I,V,L,P,A) norm
w.write(
str((float(nx.average_clustering(G)) /
((float(nx.number_of_edges(G)) /
float(nx.number_of_nodes(G))) /
float(nx.number_of_nodes(G)))) /
(float(nx.average_clustering(H)) /
((float(nx.number_of_edges(H)) /
float(nx.number_of_nodes(H))) /
float(nx.number_of_nodes(H))))) + '\n'
) #clustering coefficient of random graph - hydrophobic residues norm
w.write(
str(nx.global_reaching_centrality(H)) + '\n'
) #global, Average local reaching centrality of hydrophobic residues
w.write(
str((float(nx.number_of_edges(C)) /
float(nx.number_of_nodes(C))) /
(float(nx.number_of_edges(G)) /
float(nx.number_of_nodes(G)))) + '\n'
) #***Average degree of charged residues (R,D,E,H,K) norm
try:
w.write(
str((float(nx.average_clustering(G)) /
((float(nx.number_of_edges(G)) /
float(nx.number_of_nodes(G))) /
float(nx.number_of_nodes(G)))) /
(float(nx.average_clustering(C)) /
((float(nx.number_of_edges(C)) /
float(nx.number_of_nodes(C))) /
float(nx.number_of_nodes(C))))) + '\n'
) #clustering coefficient of random graph - charged residues norm
except ZeroDivisionError:
w.write('0.0' + '\n')
w.write(
str(nx.global_reaching_centrality(C)) + '\n'
) #global, Average local reaching centrality of charged residues
p = (float(nx.number_of_edges(G)) /
float(nx.number_of_nodes(G))) / (
float(nx.number_of_nodes(G)) - 1)
X = [x[1] for x in G.degree()]
w.write(
str(
np.var(X, dtype=np.float64) /
(float(nx.number_of_nodes(G)) * p * (1 - p))) +
'\n') #variance
w.write(
str(
skew(X, bias=False) / ((1 - (2 * p)) / (math.sqrt(
((float(nx.number_of_nodes(G)) - 1) * p) *
(1 - p))))) + '\n') #skewness
except nx.exception.NetworkXError:
pass #graph is not connected
print base_list[
i], ' *.txt file has been saved! (original PDB, hydrophobic, charged)'
}
fun_assortativity_measures = {
"average_neighbor_degree": nx.average_neighbor_degree
}
fun_global_measures = {
"global_density":
lambda G: dict(zip(G.nodes, [nx.density(G.to_undirected())] * len(G))),
"global_average_shortest_path_length":
lambda G: dict(
zip(G.nodes, [nx.average_shortest_path_length(G.to_undirected())] *
len(G))),
"global_reaching_centrality":
lambda G: dict(
zip(G.nodes, [nx.global_reaching_centrality(G.to_undirected())] * len(
G))),
"global_degree_assortativity_coefficient":
lambda G: dict(
zip(G.nodes, [nx.degree_assortativity_coefficient(G.to_undirected())] *
len(G))),
"global_degree_pearson_correlation_coefficient":
lambda G: dict(
zip(G.nodes, [
nx.degree_pearson_correlation_coefficient(G.to_undirected())
] * len(G))),
"global_transitivity":
lambda G: dict(zip(G.nodes, [nx.transitivity(G.to_undirected())] * len(G))
),
"global_average_clustering":
lambda G: dict(
def features_part1(info):
"""
second set of features.
"""
G = info['G']
n = info['num_nodes']
num_units = info['num_units']
edges = info['edges']
nedges = len(edges)
res = dict()
res['num_nodes'] = n
res['num_edges'] = nedges
res['reduce_frac'] = num_units / n - 1
res['edges_per_node'] = nedges / n
res['density'] = nx.density(G)
res['transitivity'] = nx.transitivity(G)
res['average_clustering'] = nx.average_clustering(G)
res['average_node_connectivity'] = nx.average_node_connectivity(G)
res['average_shortest_path_length'] = nx.average_shortest_path_length(G)
res['s_metric_norm'] = np.sqrt(nx.s_metric(G, normalized=False) / nedges)
res['global_reaching_centrality'] = nx.global_reaching_centrality(G)
res['edge_connectivity'] = nx.edge_connectivity(G, 0, n - 1)
res['modularity_trace'] = np.real(np.sum(nx.modularity_spectrum(G)))
stages = info['stages']
stagedict = get_stage_dict(n, stages)
edges_diff = np.array([stagedict[j] - stagedict[i] for (i, j) in edges])
n0 = np.sum(edges_diff == 0)
n1 = np.sum(edges_diff == 1)
n2 = np.sum(edges_diff == 2)
res['intrastage'] = n0 / nedges
res['interstage'] = n1 / nedges
res['hops_per_node'] = n2 / n
degrees = np.array(nx.degree(G))[:, 1]
res['mean_degree'] = np.mean(degrees)
res['std_degree'] = np.std(degrees)
res['span_degree'] = np.amax(degrees) / np.amin(degrees)
triadic = nx.triadic_census(G)
res['021D'] = triadic['021D'] / nedges
res['021U'] = triadic['021U'] / nedges
res['021C'] = triadic['021C'] / nedges
res['030T'] = triadic['030T'] / nedges
paths_nums = pathhist(G)
ns = np.arange(1, n + 1)
paths_total = np.sum(paths_nums)
mean_path = np.sum(ns * paths_nums) / paths_total
mean_pathsqr = np.sum(ns**2 * paths_nums) / paths_total
std_path = np.sqrt(mean_pathsqr - mean_path**2)
nonz = np.nonzero(paths_nums)[0]
shortest_path = nonz[0] + 1
longest_path = nonz[-1] + 1
res['log_paths'] = np.log(paths_total)
res['mean_path'] = mean_path
res['std_paths'] = std_path
res['min_path'] = shortest_path
res['max_path'] = longest_path
res['span_path'] = longest_path / shortest_path
return res
def test_non_positive_weights(self):
with pytest.raises(nx.NetworkXError):
G = nx.DiGraph()
nx.global_reaching_centrality(G, weight="weight")
def test_cycle_directed_weighted(self):
G = nx.DiGraph()
G.add_weighted_edges_from([(1, 2, 1), (2, 1, 1)])
assert nx.global_reaching_centrality(G) == 0
def test_cycle_undirected_unweighted(self):
G = nx.Graph()
G.add_edge(1, 2)
assert nx.global_reaching_centrality(G, weight=None) == 0
def test_negatively_weighted(self):
with pytest.raises(nx.NetworkXError):
G = nx.Graph()
G.add_weighted_edges_from([(0, 1, -2), (1, 2, +1)])
nx.global_reaching_centrality(G, weight="weight")
