def get_train_data():
train_x_y = []
global flag0_count
global flag1_count
flag0_count = 0
flag1_count = 0
sess = tf.InteractiveSession()
for i in range(1, 10):
file_path_community = './0814data/%d/community-standard.txt' % i
file_path_network = './0814data/%d/network.txt' % i
# print(file_path_network)
social_list = cnn_socialNet_read_data.get_standard_network(
file_path_community)
my_graph = cnn_socialNet_read_data.get_graph(file_path_network)
cnn_socialNet_read_data.add_flag_graph(my_graph, social_list)
edges = []
for (u, v, flag) in my_graph.edges.data('flag'):
# print(u, v, flag)
if int(flag) == 0:
flag0_count = flag0_count + 1
else:
flag1_count = flag1_count + 1
edges.append(((u, v), flag))
for j in range(len(edges)):
matrix1, row1, clown1 = cnn_socialNet_deal_data.get_jump1_3dimension_different_size_matrix(
my_graph, edges[j][0])
image1 = tf.convert_to_tensor(matrix1)
image1 = tf.image.convert_image_dtype(image1, tf.float32)
resize_image1 = tf.image.resize_images(image1, [128, 128],
method=3)
img_numpy1 = resize_image1.eval(session=sess)
matrix2, row2, clown2 = cnn_socialNet_deal_data.get_jump2_3dimension_different_size_matrix(
my_graph, edges[j][0])
image2 = tf.convert_to_tensor(matrix2)
image2 = tf.image.convert_image_dtype(image2, tf.float32)
resize_image2 = tf.image.resize_images(image2, [128, 128],
method=3)
img_numpy2 = resize_image2.eval(session=sess)
# print('resize_iamge', img_numpy)
# matrix1 = tf.constant(resize_image).eval()
# print(edges[j][1])
if int(edges[j][1]) == 1:
label = [1, 0]
else:
label = [0, 1]
train_x_y.append((img_numpy1, img_numpy2, label))
sess.close()
return train_x_y, flag0_count, flag1_count
def get_train_data():
train_x_y = []
global flag0_count
global flag1_count
flag0_count = 0
flag1_count = 0
sess = tf.InteractiveSession()
file_path_community = './small_data/karate-standard.txt'
file_path_network = './small_data/karate-edges.txt'
# print(file_path_network)
social_list = cnn_socialNet_read_data.get_standard_network(
file_path_community)
print(social_list)
my_graph = cnn_socialNet_read_data.get_graph(file_path_network, split=',')
print(my_graph.edges)
cnn_socialNet_read_data.add_flag_graph(my_graph, social_list)
edges = []
for (u, v, flag) in my_graph.edges.data('flag'):
print(u, v, flag)
if int(flag) == 0:
flag0_count = flag0_count + 1
else:
flag1_count = flag1_count + 1
edges.append(((u, v), flag))
for j in range(len(edges)):
matrix, row, clown = cnn_socialNet_deal_data.get_jump1_3dimension_different_size_matrix(
my_graph, edges[j][0])
image = tf.convert_to_tensor(matrix)
image = tf.image.convert_image_dtype(image, tf.float32)
resize_image = tf.image.resize_images(image, [128, 128], method=3)
img_numpy = resize_image.eval(session=sess)
# print('resize_iamge', img_numpy)
# matrix1 = tf.constant(resize_image).eval()
# print(edges[j][1])
if int(edges[j][1]) == 1:
label = [1, 0]
else:
label = [0, 1]
train_x_y.append((img_numpy, label))
sess.close()
return train_x_y, flag0_count, flag1_count
import networkx as nx
import cnn_socialNet_read_data
file_path = './0814data/1/community-standard.txt'
test_file_path = './0814data/test_community.txt'
social_list = cnn_socialNet_read_data.get_standard_network(test_file_path)
print(social_list)
# 测试获取网络
test_network = './0814data/test_nodes.txt'
network_file_path = '0814data/1/network.txt'
test_G = cnn_socialNet_read_data.get_graph(test_network)
print(test_G.edges())
# 测试为网络中每条边添加flag属性
edges = []
cnn_socialNet_read_data.add_flag_graph(test_G, social_list)
for (u, v, flag) in test_G.edges.data('flag'):
print(u, v, flag)
edges.append(((u, v), int(flag)))
# 划分社区算法
visited = {}
def breadth_first_search(root=None):
queue = []
social = []
nodes = test_G.nodes
def bfs(first_node):
order = [first_node]