def entrance(in_tree, out_tree, twigs, l_in, l_out, node_set, q_in, q_out,
q_in_menu, q_out_menu, fo):
global head_map, END
heads = []
END = len(twigs)
for twig in twigs:
heads.append(twig.head)
print("开始获取head映射...")
print("开始获取head映射...", file=statics.f_console)
t1 = time.clock() # timer
result = __get_head_map(in_tree, out_tree, heads, q_in, q_out, q_in_menu,
q_out_menu)
t2 = time.clock() # timer
print("获取head映射耗时 " + str(t2 - t1))
print("获取head映射耗时 " + str(t2 - t1), file=statics.f_console)
# 对照, 朴素的head获取方法
__naive_get_head_map(l_in, l_out, node_set, heads, q_in, q_out)
if result == common.INVALID_CANDIDATE:
return common.INVALID_CANDIDATE
querybox_list = CQueue()
querybox_list.put(common.QueryBox(0))
print("查询内核启动...")
print("查询内核启动...", file=statics.f_console)
t3 = time.clock() # timer
while not querybox_list.is_empty():
x_box = querybox_list.get()
result = __core(l_in=l_in, l_out=l_out, box=x_box, twigs=twigs, fo=fo)
for each in result:
querybox_list.put(each)
t4 = time.clock() # timer
print("查询内核运行耗时 " + str(t4 - t3))
print("查询内核运行耗时 " + str(t4 - t3), file=statics.f_console)
def entrance(in_tree, out_tree, twigs, l_in, l_out, node_set, q_in, q_out, q_in_menu, q_out_menu, fo):
global head_map, END
heads = []
END = len(twigs)
for twig in twigs:
heads.append(twig.head)
print("开始获取head映射...")
print("开始获取head映射...", file=statics.f_console)
t1 = time.clock() # timer
result = __get_head_map(in_tree, out_tree, heads, q_in, q_out, q_in_menu, q_out_menu)
t2 = time.clock() # timer
print("获取head映射耗时 " + str(t2 - t1))
print("获取head映射耗时 " + str(t2 - t1), file=statics.f_console)
# 对照, 朴素的head获取方法
__naive_get_head_map(l_in, l_out, node_set, heads, q_in, q_out)
if result == common.INVALID_CANDIDATE:
return common.INVALID_CANDIDATE
querybox_list = CQueue()
querybox_list.put(common.QueryBox(0))
print("查询内核启动...")
print("查询内核启动...", file=statics.f_console)
t3 = time.clock() # timer
while not querybox_list.is_empty():
x_box = querybox_list.get()
result = __core(l_in=l_in, l_out=l_out, box=x_box, twigs=twigs, fo=fo)
for each in result:
querybox_list.put(each)
t4 = time.clock() # timer
print("查询内核运行耗时 " + str(t4 - t3))
print("查询内核运行耗时 " + str(t4 - t3), file=statics.f_console)
def __build_tree(table, menu, node_set, neighbor):
root = TreeNode(data=[])
for node in node_set:
pin = root
if node not in menu:
root.data.append(node)
continue
if not menu[node]:
root.data.append(node)
continue
for label in menu[node]:
for i in range(0, table[node][label]):
if label not in pin.children:
child = TreeNode(data=[])
pin.add_child(label, child)
pin = child
else:
pin = pin.children[label]
pin.data.append(node)
print("开始调整出入度树以最大化差异...")
print("开始调整出入度树以最大化差异...", file=statics.f_console)
# 梳理树中的目录信息, 并根据邻居信息将树中同一树节点上的点按照最大novelty间隔排开
q = CQueue()
q.put(root)
t_start_intervein = time.clock()
while not q.is_empty():
x_node = q.get()
__intervein_node(x_node.data, neighbor)
x_node.label_menu.sort()
for edge in x_node.children:
q.put(x_node.children[edge])
t_end_intervein = time.clock()
print("差异化出入度表耗时 " + str(t_end_intervein - t_start_intervein))
print("差异化出入度表耗时 " + str(t_end_intervein - t_start_intervein), file=statics.f_console)
print("开始计算prophecy...")
print("开始计算prophecy...", file=statics.f_console)
t_start_prophecy = time.clock()
__dfs_for_prophecy(root)
t_end_prophecy = time.clock()
print("计算prophecy耗时: " + str(t_end_prophecy - t_start_prophecy))
print("计算prophecy耗时: " + str(t_end_prophecy - t_start_prophecy), file=statics.f_console)
print("开始深先遍历出入度树...")
print("开始深先遍历出入度树...", file=statics.f_console)
statics.io_tree_max_dep = -1
dfs_for_depth(root, 1)
print("得到最大深度:" + str(statics.io_tree_max_dep))
print("得到最大深度:" + str(statics.io_tree_max_dep), file=statics.f_console)
return root
def __build_tree(table, menu, node_set, neighbor):
root = TreeNode(data=[])
for node in node_set:
pin = root
if node not in menu:
root.data.append(node)
continue
if not menu[node]:
root.data.append(node)
continue
for label in menu[node]:
for i in range(0, table[node][label]):
if label not in pin.children:
child = TreeNode(data=[])
pin.add_child(label, child)
pin = child
else:
pin = pin.children[label]
pin.data.append(node)
print("开始调整出入度树以最大化差异...")
print("开始调整出入度树以最大化差异...", file=statics.f_console)
# 梳理树中的目录信息, 并根据邻居信息将树中同一树节点上的点按照最大novelty间隔排开
q = CQueue()
q.put(root)
t_start_intervein = time.clock()
while not q.is_empty():
x_node = q.get()
__intervein_node(x_node.data, neighbor)
x_node.label_menu.sort()
for edge in x_node.children:
q.put(x_node.children[edge])
t_end_intervein = time.clock()
print("差异化出入度表耗时 " + str(t_end_intervein - t_start_intervein))
print("差异化出入度表耗时 " + str(t_end_intervein - t_start_intervein), file=statics.f_console)
print("开始计算prophecy...")
print("开始计算prophecy...", file=statics.f_console)
t_start_prophecy = time.clock()
__dfs_for_prophecy(root)
t_end_prophecy = time.clock()
print("计算prophecy耗时: " + str(t_end_prophecy - t_start_prophecy))
print("计算prophecy耗时: " + str(t_end_prophecy - t_start_prophecy), file=statics.f_console)
print("开始深先遍历出入度树...")
print("开始深先遍历出入度树...", file=statics.f_console)
statics.io_tree_max_dep = -1
dfs_for_depth(root, 1)
print("得到最大深度:" + str(statics.io_tree_max_dep))
print("得到最大深度:" + str(statics.io_tree_max_dep), file=statics.f_console)
return root
def filter(self):
"""
Filter all candidate by single projection
if null set caused, return common.INVALIDATE_CANDIDAT
:return:
"""
pipe_line = CQueue()
for node in self.candidate:
if len(self.candidate[node]) == 1:
griddle = None
for each in self.candidate[node]:
griddle = each
if griddle is None:
return INVALID_CANDIDATE
for other in self.candidate:
if node != other:
try:
self.candidate[other].remove(griddle)
new_len = len(self.candidate[other])
if new_len < 1:
return INVALID_CANDIDATE
if new_len == 1:
pipe_line.put(other)
except:
pass
while not pipe_line.is_empty():
griddle = None
x_node = pipe_line.get()
for each in self.candidate[x_node]:
griddle = each
if griddle is None:
return INVALID_CANDIDATE
for other in self.candidate:
if other != x_node:
try:
self.candidate[other].remove(griddle)
new_len = len(self.candidate[other])
if new_len < 1:
return INVALID_CANDIDATE
if new_len == 1:
pipe_line.put(other)
except:
pass
return VALID_CANDIDATE
def filter(self):
"""
Filter all candidate by single projection
if null set caused, return common.INVALIDATE_CANDIDAT
:return:
"""
pipe_line = CQueue()
for node in self.candidate:
if len(self.candidate[node]) == 1:
griddle = None
for each in self.candidate[node]:
griddle = each
if griddle is None:
return INVALID_CANDIDATE
for other in self.candidate:
if node != other:
try:
self.candidate[other].remove(griddle)
new_len = len(self.candidate[other])
if new_len < 1:
return INVALID_CANDIDATE
if new_len == 1:
pipe_line.put(other)
except:
pass
while not pipe_line.is_empty():
griddle = None
x_node = pipe_line.get()
for each in self.candidate[x_node]:
griddle = each
if griddle is None:
return INVALID_CANDIDATE
for other in self.candidate:
if other != x_node:
try:
self.candidate[other].remove(griddle)
new_len = len(self.candidate[other])
if new_len < 1:
return INVALID_CANDIDATE
if new_len == 1:
pipe_line.put(other)
except:
pass
return VALID_CANDIDATE
def load_map():
"""
Pass L_in, L_out, noted for in/out degree dictionary and in/out label-marked dictionary
NOTE! The initialization must be done outside the function
"""
print("开始加载文件...")
print("开始加载文件...", file=statics.f_console)
t_start_load_file = time.clock() # timer
l_in = {}
l_out = {}
l_in_menu = {}
l_out_menu = {}
node_set = set()
f = open(ROOT_PATH + "data.txt")
while True:
line = f.readline()
if len(line) < 1:
break
tpl = line.split('\t')
ori = int(tpl[0])
edg = int(tpl[1])
des = int(tpl[2])
node_set.add(ori)
node_set.add(des)
if ori not in l_out:
l_out[ori] = {}
if ori not in l_out_menu:
l_out_menu[ori] = []
if des not in l_in:
l_in[des] = {}
if des not in l_in_menu:
l_in_menu[des] = []
try:
l_in[des][edg].append(ori)
except:
l_in[des][edg] = []
l_in[des][edg].append(ori)
try:
l_out[ori][edg].append(des)
except:
l_out[ori][edg] = []
l_out[ori][edg].append(des)
if edg not in l_in_menu[des]:
l_in_menu[des].append(edg)
if edg not in l_out_menu[ori]:
l_out_menu[ori].append(edg)
f.close()
t_end_load_file = time.clock() # timer
print("结束加载文件...")
print("结束加载文件...", file=statics.f_console)
print('加载文件耗时 ' + str(t_end_load_file - t_start_load_file))
print('加载文件耗时 ' + str(t_end_load_file - t_start_load_file), file=statics.f_console)
t_start_sort_raw_data = time.clock() # timer
__inner_sort(l_in, l_in_menu)
__inner_sort(l_out, l_out_menu)
t_end_sort_raw_data = time.clock() # timer
print("原始数据内部排序耗时 " + str(t_end_sort_raw_data - t_start_sort_raw_data))
print("原始数据内部排序耗时 " + str(t_end_sort_raw_data - t_start_sort_raw_data), file=statics.f_console)
print("开始统计邻居信息")
print("开始统计邻居信息", file=statics.f_console)
t_start_neighbor = time.clock() # timer
# 统计邻居节点信息
neighbor = {}
# 先统计入度
visited = set()
q_node = CQueue()
q_dep = CQueue()
for node in node_set:
visited.clear()
visited.add(node)
neighbor[node] = NeighborInfo()
q_node.clear()
q_dep.clear()
q_node.put(node)
q_dep.put(0)
while not q_node.is_empty():
x_node = q_node.get()
x_dep = q_dep.get()
next_dep = x_dep + 1
if x_node not in l_in:
continue
for l in l_in[x_node]:
for ori in l_in[x_node][l]:
if ori not in visited:
neighbor[node].safe_add(ori, statics.fade_factor_pow[next_dep], target='in')
visited.add(ori)
if next_dep < statics.neighbor_threshold:
q_node.put(ori)
q_dep.put(next_dep)
neighbor[node].cal_module(target='in')
# 先不统计出度
t_end_neighbor = time.clock() # timer
print("统计邻居信息耗时 " + str(t_end_neighbor - t_start_neighbor))
print("统计邻居信息耗时 " + str(t_end_neighbor - t_start_neighbor), file=statics.f_console)
return l_in, l_out, l_in_menu, l_out_menu, node_set, neighbor
def load_map():
"""
Pass L_in, L_out, noted for in/out degree dictionary and in/out label-marked dictionary
NOTE! The initialization must be done outside the function
"""
print("开始加载文件...")
print("开始加载文件...", file=statics.f_console)
t_start_load_file = time.clock() # timer
l_in = {}
l_out = {}
l_in_menu = {}
l_out_menu = {}
node_set = set()
f = open(ROOT_PATH + "data.txt")
while True:
line = f.readline()
if len(line) < 1:
break
tpl = line.split('\t')
ori = int(tpl[0])
edg = int(tpl[1])
des = int(tpl[2])
node_set.add(ori)
node_set.add(des)
if ori not in l_out:
l_out[ori] = {}
if ori not in l_out_menu:
l_out_menu[ori] = []
if des not in l_in:
l_in[des] = {}
if des not in l_in_menu:
l_in_menu[des] = []
try:
l_in[des][edg].append(ori)
except:
l_in[des][edg] = []
l_in[des][edg].append(ori)
try:
l_out[ori][edg].append(des)
except:
l_out[ori][edg] = []
l_out[ori][edg].append(des)
if edg not in l_in_menu[des]:
l_in_menu[des].append(edg)
if edg not in l_out_menu[ori]:
l_out_menu[ori].append(edg)
f.close()
t_end_load_file = time.clock() # timer
print("结束加载文件...")
print("结束加载文件...", file=statics.f_console)
print('加载文件耗时 ' + str(t_end_load_file - t_start_load_file))
print('加载文件耗时 ' + str(t_end_load_file - t_start_load_file),
file=statics.f_console)
t_start_sort_raw_data = time.clock() # timer
__inner_sort(l_in, l_in_menu)
__inner_sort(l_out, l_out_menu)
t_end_sort_raw_data = time.clock() # timer
print("原始数据内部排序耗时 " + str(t_end_sort_raw_data - t_start_sort_raw_data))
print("原始数据内部排序耗时 " + str(t_end_sort_raw_data - t_start_sort_raw_data),
file=statics.f_console)
print("开始统计邻居信息")
print("开始统计邻居信息", file=statics.f_console)
t_start_neighbor = time.clock() # timer
# 统计邻居节点信息
neighbor = {}
# 先统计入度
visited = set()
q_node = CQueue()
q_dep = CQueue()
for node in node_set:
visited.clear()
visited.add(node)
neighbor[node] = NeighborInfo()
q_node.clear()
q_dep.clear()
q_node.put(node)
q_dep.put(0)
while not q_node.is_empty():
x_node = q_node.get()
x_dep = q_dep.get()
next_dep = x_dep + 1
if x_node not in l_in:
continue
for l in l_in[x_node]:
for ori in l_in[x_node][l]:
if ori not in visited:
neighbor[node].safe_add(
ori,
statics.fade_factor_pow[next_dep],
target='in')
visited.add(ori)
if next_dep < statics.neighbor_threshold:
q_node.put(ori)
q_dep.put(next_dep)
neighbor[node].cal_module(target='in')
# 先不统计出度
t_end_neighbor = time.clock() # timer
print("统计邻居信息耗时 " + str(t_end_neighbor - t_start_neighbor))
print("统计邻居信息耗时 " + str(t_end_neighbor - t_start_neighbor),
file=statics.f_console)
return l_in, l_out, l_in_menu, l_out_menu, node_set, neighbor