def test_is_empty():
graphs = [nx.Graph(), nx.DiGraph(), nx.MultiGraph(), nx.MultiDiGraph()]
for G in graphs:
assert_true(nx.is_empty(G))
G.add_nodes_from(range(5))
assert_true(nx.is_empty(G))
G.add_edges_from([(1, 2), (3, 4)])
assert_false(nx.is_empty(G))
def test_is_empty():
graphs = [nx.Graph(), nx.DiGraph(), nx.MultiGraph(), nx.MultiDiGraph()]
for G in graphs:
assert nx.is_empty(G)
G.add_nodes_from(range(5))
assert nx.is_empty(G)
G.add_edges_from([(1, 2), (3, 4)])
assert not nx.is_empty(G)
def analysis(edges, node_dic, class_dic={}):
G = nx.Graph()
for edge in edges:
x = edge[0]
y = edge[1]
G.add_edge(x, y)
graph_degrees = G.degree()
degree_sequence = sorted([d for n, d in G.degree()],
reverse=False) # mac default ?
degreeCount = collections.Counter(degree_sequence)
degreeDist = list(degreeCount.items())
# # remove disconnected nodes to calculate Radius and Graph centers.
if not nx.is_empty(G):
largest_cc = max(nx.connected_components(G))
nodelist = G.nodes()
to_remove = set(nodelist) - set(largest_cc)
for nd in to_remove:
G.remove_node(nd)
radius, eccentricity, center = "", "", ""
if not nx.is_empty(G):
if nx.is_connected(G):
radius = nx.radius(G)
eccentricity = nx.eccentricity(G)
center = nx.center(
G
) # The center is the set of nodes with eccentricity equal to radius.
mod_score = -2
# check if it makes sense to calculate the modularity score
aTestNodeName = "" if len(list(node_dic.keys())) == 0 else node_dic[list(
node_dic.keys())[0]]
if bool(class_dic.keys()
) and aTestNodeName != "" and aTestNodeName in class_dic:
inv_dic = {}
nodelist = G.nodes()
for n in nodelist:
converted_num = node_dic[n]
class_tmp = class_dic[converted_num]
if class_tmp not in inv_dic:
inv_dic[class_tmp] = set([n])
else:
inv_dic[class_tmp].add(n)
groups = [x for k, x in inv_dic.items()]
mod_score = nx_comm.modularity(G, groups)
return radius, eccentricity, center, degreeDist, mod_score, G
def graph(listMovies):
G = nx.Graph()
couple=[]
count=0
for i in range(len(listMovies)):
for j in range(len(listMovies[i])):
for y in range(1,len(listMovies[i])):
couple=[listMovies[i][j],listMovies[i][y]]
count=0
for k in listMovies:
if (couple[0] in k) and (couple[1] in k):
count+=1
if count>=2 and not couple[0] == couple[1]:
G.add_edge(couple[0],couple[1])
break
if not nx.is_empty(G):
pos = nx.spring_layout(G) #<<<<<<<<<< Initialize this only once
nx.draw(G,pos=pos, with_labels=True, node_size = 100, font_size=10)
nx.draw_networkx_nodes(G,pos=pos, with_labels=True, node_size = 1500, font_size=10)
nx.draw_networkx_edges(G, pos, alpha=0.3)#<<<<<<<<< pass the pos variable
#plt.draw()
plt.figure(figsize=(8, 8)) # image is 8 x 8 inches
# To plot the next graph in a new figure
plt.show()
def isGraphEmtpy(self):
try:
if nx.is_empty(self.G):
raise nx.NetworkXError('No graph found!')
except Exception as e:
messagebox.showerror(title=None, message=e)
raise e
def drawGraph_Func1(result, typeDistance):
G = nx.Graph()
E_color = ''
for i in result:
G.add_node(i)
for i in result:
for j in result:
edge = CAL.get_edge(i, j, [typeDistance])
if not edge == None:
G.add_edge(i, j)
if not nx.is_empty(G):
colors = ["red"] + (["cyan"] * (len(result) - 1))
pos = nx.spring_layout(G) #<<<<<<<<<< Initialize this only once
nx.draw(G, pos, with_labels=True, font_size=25)
nx.draw_networkx_nodes(G, pos, node_color=colors, node_size=1000)
if typeDistance == "Distance":
nx.draw_networkx_edges(G, pos, edge_color='blue', width=3)
elif typeDistance == "Time":
nx.draw_networkx_edges(G, pos, edge_color='purple', width=3)
else:
nx.draw_networkx_edges(G, pos, edge_color='red', width=3)
plt.show()
def load_graph(filename: str,
graph_name: str,
format_type='edge_list',
is_directed=False):
"""
Load directed or undirected graph from file with specific format
:param filename: File to read graph
:param graph_name: Graph name
:param format_type: File format of saved graph e.g.: `edge_list`
:param is_directed: Load as directed graph?
:return: g: nx.Graph
"""
if not os.path.isfile(filename):
logger.error(f"File: [{filename}] does not exist")
return
if format_type == 'edge_list':
graph_type = nx.DiGraph if is_directed else nx.Graph
g = nx.read_edgelist(filename, create_using=graph_type)
add_graph_name(g, graph_name)
if nx.is_empty(g):
logger.info("Loaded graph is empty")
return g
else:
logger.error(f"Unknown graph input format: [{format_type}]")
return
def network_graph_nx(root):
"""Purpose of this function is to create a Network graph for visualization of each individual object given in the KNOSSOS XML input file"""
# Need to iterate over each of the objects in the XML file to add individual skeletons to the Networkx graph:
final_nx_graph = nx.Graph()
position_labels = xy_node_position_dict(root)
for thing in parse.root.iter('thing'):
# Creating the dataframe with node information as input:
node_df = ed.ed_per_node(thing)
# Designating the nodes to be used in the Network graph, using if statement to add nodes as part of separate object:
network_graph_nx = nx.Graph()
if nx.is_empty(network_graph_nx):
network_graph_first = nx.from_pandas_edgelist(node_df, 'Source ID', 'Target ID', edge_attr='Euclidean Distance')
network_graph_nx.add_edges_from(network_graph_first.edges())
network_graph_nx.add_nodes_from(network_graph_first.nodes())
else:
network_graph_add = nx.from_pandas_edgelist(node_df, 'Source ID', 'Target ID', edge_attr='Euclidean Distance')
network_graph_nx.update(network_graph_add)
# Appending the final networkx graph and saving it as a .graphml file type:
final_nx_graph.update(network_graph_nx)
# Adding node attributes such as position and skeleton group:
# positions = nx.spring_layout(final_nx_graph, dim=3, k=None, pos=position_labels, fixed=None, iterations=50, weight='weight', scale=1.0)
nx.set_node_attributes(final_nx_graph, position_labels, name='Position')
return final_nx_graph
def ampiezza(grafo, start, end):
'''
Apply BFS to find the path from a node to another
:param grafo: networkx.Graph
:param start: name of start node
:param end: name of destination node
:return: Path from start to end if exists, empty list otherwise
'''
if nx.is_empty(grafo) or not grafo.has_node(start) or not grafo.has_node(
end) or start == end:
return []
queue = [start]
while queue:
curr = queue.pop(0)
for node in grafo.neighbors(curr):
# skip iteration if node already visited
if grafo.nodes[node]['visitato']:
continue
if node == end:
return grafo.nodes(data=True)[curr]['percorso'] + [node]
else:
grafo.nodes[node][
'percorso'] = grafo.nodes[curr]['percorso'] + [node]
grafo.nodes[node]['visitato'] = True
queue.append(node)
# no path found
return []
def run_inference(self, debug=True, return_results=True):
"""
Run Inference on network with given evidences.
"""
g_temp = copy.deepcopy(self.g)
self._log = self.log.setup_logger(debug=debug)
self._log.debug("Started")
if all(x == None for x in self.evidences.values()):
self._log.debug(
"No evidences were set. Proceeding without evidence")
self.parameters = dict.fromkeys(self.nodes)
self.calculated_means = copy.deepcopy(self.evidences)
self.calculated_vars = dict.fromkeys(self.nodes)
self.done_flags = dict.fromkeys(self.nodes)
it = 0
while not nx.is_empty(g_temp):
it += 1
pure_children = self.__get_pure_root_nodes(g_temp)
for child in pure_children:
if self.evidences[child] is None:
self.calculated_means[child], self.calculated_vars[
child] = self.__get_node_values(child)
self.__print_message(self._log, child)
else:
self._log.debug(
f"Skipped Calculating:'{child}' as evidence is available."
)
g_temp.remove_nodes_from(list(g_temp.pred[child]))
return self.__build_results()
def pajek_to_files(name, url, pajek_lines, dir_name):
if pajek_lines:
try:
G = nx.parse_pajek(pajek_lines)
if not nx.is_empty(G):
old_attributes = list(G.nodes)
G = nx.convert_node_labels_to_integers(G)
id_mapping = []
node_list = list(G.nodes)
for i in range(len(node_list)):
id_mapping.append([old_attributes[i], str(node_list[i])])
mapping_file = open('..' + dir_name +
'/node_id_mappings/mapping_' +
url.split('/')[-1] + '.csv',
'w',
newline='')
mapping_file_writer = csv.writer(mapping_file)
mapping_file_writer.writerow(['id', 'name'])
for tup in id_mapping:
mapping_file_writer.writerow(list(tup))
nx.write_edgelist(G,
'..' + dir_name + '/edge_lists/' +
url.split('/')[-1] + '.csv',
delimiter=',')
insert_into_db(
name, url,
dir_name + '/edge_lists/' + url.split('/')[-1] + '.csv',
dir_name + '/node_id_mappings/mapping_' +
url.split('/')[-1] + '.csv', G.is_directed(),
G.is_multigraph(), int(G.number_of_nodes()),
int(nx.number_of_selfloops(G)))
except Exception as e:
traceback.print_exc()
print(e)
print("Couldn't parse " + url)
def network_five_function(year):
"""具体函数"""
file_path = '../data/生成数据/03关系矩阵/五年期/' + str(year) + '-' + str(
year + 4) + '年竞争关系矩阵.csv'
if os.path.exists(file_path):
graph = create_diagrams(file_path)
if not nx.is_empty(graph):
network_indicators = calculate_networks_indicators(
graph, year, '五年期')
excel_path = '../data/生成数据/04关系矩阵_网络指标/五年期/' + file_path[-20:-4]
folder = os.path.exists(excel_path)
if not folder:
os.makedirs(excel_path)
network_indicators.to_excel(excel_writer=excel_path + '/相关指标.xlsx',
index=False)
address = pd.read_excel(
io='../data/生成数据/01企业名称修改表/13整车数据ZZCMCZ编码表.xlsx',
usecols=[0, 1],
converters={
'ZZCMC编号': str,
'ZZCMC企业名称': lambda x: x.strip()
})
address = address.rename(columns={
'ZZCMC编号': 'nodes',
'ZZCMC企业名称': 'institution_name'
})
network_indicators_address = pd.merge(left=network_indicators,
right=address,
how='left',
on=['nodes'])
network_indicators_address['year'] = year + 4
network_indicators_address.to_excel(excel_writer=excel_path +
'/相关指标(企业名称).xlsx',
index=False)
def exchange_shifts(workers: List[Worker]):
# init graph
g = nx.DiGraph()
g.add_nodes_from(workers) # worker node
g.add_nodes_from([0, len(workers) - 1]) # shift node
for w in workers:
g.add_edge(w.current_shift, w) # add edge from shift to worker
g.add_edge(w, w.wanted_shift) # add edge from worker to wanted shift
while not nx.is_empty(g):
# find cycle and update shifts and remove nodes
for cycle in nx.simple_cycles(g):
for node in cycle:
if not isinstance(node,
int): # go on the workers and match shifts
if not node.wanted_shift == node.current_shift: # if the shift was switched
print(node.name, "moves from shift",
node.current_shift, "to shift",
node.wanted_shift)
else:
print(node.name, "stays with shift",
node.current_shift)
g.remove_nodes_from(cycle)
# update new wanted shifts
for w in list(g.nodes):
if not isinstance(w, int):
while not g.has_node(
w.wanted_shift): # search the first shift that exists
w.delete_first()
g.add_edge(
w, w.wanted_shift) # add edge from worker to wanted shift
def solve(G):
"""
Args:
G: networkx.Graph
Returns:
T: networkx.Graph
"""
end_time = datetime.now() + timedelta(seconds=TIMEOUT)
graphs_to_consider = []
graphs_to_consider.append(minimum_spanning_tree(G))
index = 0
while datetime.now() < end_time and index < len(G.nodes):
resulting_graph = create_graph_from_dominating_set(
G, dominating_set(G, start_with=index))
#print( average_pairwise_distance(resulting_graph))
#resulting_graph = optimize_pairwise_distances(G, resulting_graph, average_pairwise_distance(resulting_graph))
if not nx.is_empty(resulting_graph) and nx.is_connected(
resulting_graph):
graphs_to_consider.append(resulting_graph)
index += 1
return min(graphs_to_consider, key=average_pairwise_distance)
def spectral_gap_sparse(G):
if (not nx.is_empty(G)):
if (nx.is_connected(G)):
d = []
# start = time.time()
for elem in list(G.nodes):
deg = len(list(G.neighbors(elem)))
d.append(1 / deg)
Aa100 = nx.to_scipy_sparse_matrix(G)
invD = np.diag(d)
n = len(d)
out = sp.sparse.csr_matrix(invD)
P = out * Aa100
if (len(G.nodes()) > 2):
spettro = sp.sparse.linalg.eigsh(P,
k=2,
which="LA",
return_eigenvectors=False)
spettro = sorted(spettro, reverse=True)
sg = spettro[0] - spettro[1]
else:
sg = 0
else:
sg = 0
else:
sg = 0
sg_trans = float(sg)
return (sg_trans)
def prune_graph(self):
import graspologic.utils as gu
from pynets.statistics.individual.algorithms import defragment, \
prune_small_components, most_important
hardcoded_params = utils.load_runconfig()
if int(self.prune) not in range(0, 4):
raise ValueError(f"Pruning option {self.prune} invalid!")
if self.prune != 0:
# Remove isolates
G_tmp = self.G.copy()
self.G = defragment(G_tmp)[0]
del G_tmp
if int(self.prune) == 1:
try:
self.G = prune_small_components(
self.G, min_nodes=hardcoded_params["min_nodes"][0])
except BaseException:
print(
UserWarning(f"Warning: pruning {self.est_path} "
f"failed..."))
elif int(self.prune) == 2:
try:
hub_detection_method = \
hardcoded_params["hub_detection_method"][0]
print(f"Filtering for hubs on the basis of "
f"{hub_detection_method}...\n")
self.G = most_important(self.G, method=hub_detection_method)[0]
except FileNotFoundError as e:
import sys
print(e, "Failed to parse advanced.yaml")
elif int(self.prune) == 3:
print("Pruning all but the largest connected "
"component subgraph...")
self.G = gu.largest_connected_component(self.G)
else:
print("No graph defragmentation applied...")
self.G = nx.from_numpy_array(self.in_mat)
if nx.is_empty(self.G) is True or \
(np.abs(self.in_mat) < 0.0000001).all() or \
self.G.number_of_edges() == 0:
print(
UserWarning(f"Warning: {self.est_path} "
f"empty after pruning!"))
return self.in_mat, None
# Saved pruned
if (self.prune != 0) and (self.prune is not None):
final_mat_path = f"{self.est_path.split('.npy')[0]}{'_pruned'}"
utils.save_mat(self.in_mat, final_mat_path, self.out_fmt)
print(f"{'Source File: '}{final_mat_path}")
return self.in_mat, final_mat_path
def find_friends(self):
#find the neighbors in the graph not in the grid
if nx.is_empty(self.model.G) == False:
self.friends = self.model.G[self]
self.friend_attendance = np.array(
[f.attendance for f in self.friends])
else:
return (None)
def calculate_enterprise_indicators_three_year_ipcmg():
"""计算企业网络指标(三年期IPCMG)"""
institutions = pd.read_excel(io='../data/生成数据/06专利数据/汽车产业(企业名称).xlsx')
description_columns = [str(i) + '-' + str(i + 2) for i in range(1985, 2013)]
description_information = pd.DataFrame(columns=description_columns, index=list(institutions['企业名称']))
statistical_columns = ['企业名称', '年份', 'ipcmgdgr3yw', 'ipcmgbtw3yw', 'ipcmglcrch3yw', 'ipcmgcnstr3yw',
'ipcmgeffsz3yw', 'ipcmgtrgl3yw', 'ipcmgclust3yw', 'ipcmgtirstr3yw', 'ipcmgntsz3yw',
'ipcmgntdnst3yw', 'ipcmgcliq3yw', 'ipcmgeffc3yw', 'ipcmgisolt3yw']
statistical_information = pd.DataFrame(columns=statistical_columns)
index_symbol = 0
for i in range(len(institutions)):
each_institutions = institutions['企业名称'].iloc[i]
for j in range(1985, 2013):
years_interval = str(j) + '-' + str(j + 2)
file_path = '../data/生成数据/07IPCMGSG网络/三年期IPCMG/' + each_institutions + '/' + years_interval + '年IPCMG矩阵.csv'
if os.path.exists(file_path):
graph = create_diagrams(file_path)
if not nx.is_empty(graph):
description_information.loc[each_institutions, years_interval] = 1
network_indicators = calculate_networks_indicators(graph)
excel_path = '../data/生成数据/07IPCMGSG网络/三年期IPCMG(网络指标)/' + each_institutions + '/' + years_interval + '年IPCMG矩阵'
folder = os.path.exists(excel_path)
if not folder:
os.makedirs(excel_path)
network_indicators.to_excel(excel_writer=excel_path + '/相关指标.xlsx',
index=False)
del network_indicators['nodes']
network_indicators_average = network_indicators.mean()
row_information = [each_institutions, j + 2,
network_indicators_average['degree_centrality'],
network_indicators_average['betweenness_centrality'],
network_indicators_average['local_reaching_centrality'],
network_indicators_average['constraint'],
network_indicators_average['effective_size'],
network_indicators_average['triangles'],
network_indicators_average['clustering'],
network_indicators_average['tie_strength'],
network_indicators_average['number_of_node'],
network_indicators_average['density'],
network_indicators_average['cliques'],
network_indicators_average['efficiency'],
network_indicators_average['isolates']]
statistical_information.loc[index_symbol] = row_information
index_symbol += 1
else:
description_information.loc[each_institutions, years_interval] = 0
else:
description_information.loc[each_institutions, years_interval] = None
description_information = description_information.reset_index(drop=False)
description_information.to_excel(excel_writer='../data/生成数据/07IPCMGSG网络/统计信息/01三年期IPCMG矩阵描述信息.xlsx',
index=False)
statistical_information.to_excel(excel_writer='../data/生成数据/07IPCMGSG网络/统计信息/02三年期IPCMG网络指标信息.xlsx',
index=False)
def detect(self, traces_df, kpis, visualize=False):
traces_df = self.process(traces_df)
# Parse the traces and kpis
parsed_traces = self.parse_traces(traces_df)
# FIXME possible case where system doesn't answer for a long time and wasn't called
#check for anomaly
# 1 - find outlier in elapsed
# 1.1 microRCA
traces = self.get_anomalous_traces(traces_df)
# Hosts + Service
# Each service connects to all the services it communicates with and all hosts it connects to (no need to differentiate!)
DG = nx.DiGraph()
for trace in traces:
DG = self.trace_graph(parsed_traces[trace], DG)
if visualize:
print(DG.nodes(data=True), len(DG.nodes()))
plt.figure(figsize=(9, 9))
#pos = nx.spring_layout(DG)
# pos = nx.draw_shell(DG)
# nx.draw(DG, pos, with_labels=True, cmap=plt.get_cmap('jet'), node_size=0, arrows=True)
nx.draw_shell(DG, with_labels=True)
# nx.draw_networkx_nodes(DG, pos, nodelist=hosts, node_color="r", node_size=1500)
# nx.draw_networkx_nodes(DG, pos, nodelist=services, node_color="b", node_size=500)
# nx.draw_networkx_edges(DG, pos, width=1.0, alpha=0.5)
labels = nx.get_edge_attributes(DG, 'weight')
# nx.draw_networkx_edge_labels(DG, pos, edge_labels=labels)
plt.savefig('output.png')
# print(f'[DEBUG] Graph is {"connected" if nx.is_weakly_connected(DG) else "not connected"}')
# Extract anomalous subgraph
anomaly_DG, anomalous_edges = self.get_anomalous_graph(
DG, traces, parsed_traces)
if nx.is_empty(anomaly_DG):
raise ValueError('No anomaly detected')
# Faulty service localization
# Update weights of anomalous graph
# Use cases from the paper
# Get personalization vector (Transition Probability Matrix)
# Reverse the service-service edges
# Apply pagerank
parsed_kpis = self.parse_kpis(kpis)
result = self.get_fault_service(anomaly_DG, anomalous_edges, traces_df,
parsed_kpis)
return result
def connected(g, index):
if nx.is_empty(g) or nx.is_connected(g):
return g
else:
# largest_cc = max(nx.connected_components(g), key=len)
for component in nx.connected_components(g):
if index in component:
largest_cc = component
subG = g.subgraph(largest_cc)
return subG
def check_graph(graph, shape, max_stitch_length):
if networkx.is_empty(graph) or not networkx.is_eulerian(graph):
if shape.area < max_stitch_length**2:
message = "This shape is so small that it cannot be filled with rows of stitches. " \
"It would probably look best as a satin column or running stitch."
raise InvalidPath(_(message))
else:
message = "Cannot parse shape. " \
"This most often happens because your shape is made up of multiple sections that aren't connected."
raise InvalidPath(_(message))
def test_conncet_components_error(self):
g = nx.Graph()
gc = graphs.connect_components(g, 'pos')
self.assertTrue(nx.is_empty(gc))
g.add_node(0)
g.add_node(1, pos=(1, 1))
g.add_edge(0, 1)
self.assertRaises(ValueError,
lambda: graphs.connect_components(g, 'pos'))
def save_file_edgelist(G):
if not nx.is_empty(G):
nome = input("Inserire un nome per il salvataggio del grafo: ")
nome = '' + nome + '.txt'
f = open(nome, 'wb')
nx.write_edgelist(G, f)
print("Sotto-grafo salvato correttamente.")
f.close()
else:
print("Non c'è nessun sotto-grafo da salvare.")
def addNodeToGraph(self):
self.figure.clf()
if nx.is_empty(self.the_graph):
self.the_graph.add_node(0)
# self.color_map[0] = 'blue'
self.the_graph.add_node(list(self.the_graph.nodes)[-1] + 1)
# self.color_map[list(self.the_graph.nodes)[-1]] = 'blue'
nx.draw(self.the_graph, with_labels=True)
self.canvas.draw_idle()
def get_matching_polynomial_recursive(graph):
if nx.is_empty(graph):
return np.asarray([1] + [0] * graph.number_of_nodes())
graph_copy = graph.copy()
x = list(graph_copy.out_edges)[0]
graph_copy.remove_edges_from([x])
edge_removed_matching = get_matching_polynomial_recursive(graph_copy)
graph_copy.remove_nodes_from([x[0], x[1]])
vertex_removed_matching = get_matching_polynomial_recursive(graph_copy)
vertex_removed_matching = np.pad(
vertex_removed_matching,
(edge_removed_matching.shape[0] - vertex_removed_matching.shape[0], 0),
'constant')
return edge_removed_matching - vertex_removed_matching
def determine_safer_backbones(g, pruning_threshold=100):
""""Determine the backbones of the graph
with ambiguous nodes being removed """
g = g.copy()
backbones = []
while not nx.is_empty(g):
gmst = g
paths = determine_safer_backbones_of_trees(gmst, pruning_threshold)
backbones.extend(paths)
vertices = [u for path in paths for u in path]
neighbors = [v for u in vertices for v in g.neighbors(u)]
g.remove_nodes_from(vertices)
g.remove_nodes_from(neighbors)
backbones.sort(key=len, reverse=True)
return backbones
def partition_here(graph):
if nx.is_empty(graph):
return 0, 0
Gcc = sorted(nx.connected_components(graph), key=len, reverse=True)
G = graph.subgraph(Gcc[0])
settings = nxmetis.MetisOptions(ncuts=4, niter=200, ufactor=280)
par = nxmetis.partition(G, 2, options=settings)
community1 = par[1][0]
community2 = par[1][1]
rwc = np.mean(randomwalk_polarization(G, 100, 0.02, 1000, community1, community2))
prc = len(G)/len(graph)
return rwc, prc
def construct_state_graph(assignments_dict, id_metatile_file):
id_metatile_map = read_pickle(id_metatile_file)
state_graph = nx.DiGraph()
for (tile_x, tile_y), tile_id in assignments_dict.items():
metatile = Metatile.from_str(id_metatile_map.get(tile_id))
metatile_graph = nx.DiGraph(metatile.graph_as_dict)
if nx.is_empty(metatile_graph):
pass
else:
unnormalized_graph = Metatile.get_normalized_graph(
metatile_graph,
coord=(tile_x * TILE_DIM, tile_y * TILE_DIM),
normalize=False)
state_graph = nx.compose(state_graph, unnormalized_graph)
return state_graph
def clu_coe_attack(graph, centrality_metric):
iters = 0
graph = graph.copy()
clu_coe = []
ranks = centrality_metric(graph)
nodes = sorted(graph.nodes(), key=lambda n: ranks[n])
pos = nx.spring_layout(graph)
while nx.is_connected(graph) and not nx.is_empty(graph):
clu_coe.append(nx.average_clustering(graph))
graph.remove_node(nodes.pop())
file_name = './pics/' + str(iters) + '.png'
# graph_snap(graph, pos, file_name)
iters += 1
else:
file_name = './pics/' + str(iters) + '.png'
# graph_snap(graph, pos, file_name)
return clu_coe
def rand_cleu_coe_attack(graph, n):
iters = 0
graph = graph.copy()
clu_coe = []
nodes = graph.nodes()
pos = nx.spring_layout(graph)
while nx.is_connected(graph) and not nx.is_empty(graph):
sample = random.sample(nodes, n)
clu_coe.append(nx.average_clustering(graph))
graph.remove_nodes_from(sample)
file_name = './pics/' + str(iters) + '.png'
# graph_snap(graph, pos, file_name)
iters += 1
else:
file_name = './pics/' + str(iters) + '.png'
# graph_snap(graph, pos, file_name)
return clu_coe
def rand_den_attack(graph, n):
iters = 0
graph = graph.copy()
density = []
nodes = graph.nodes()
pos = nx.spring_layout(graph)
while nx.is_connected(graph) and not nx.is_empty(graph):
sample = random.sample(nodes, min(len(nodes), n))
density.append(nx.density(graph))
graph.remove_nodes_from(sample)
file_name = './pics/' + str(iters) + '.png'
# graph_snap(graph, pos, file_name)
iters += 1
else:
file_name = './pics/' + str(iters) + '.png'
# graph_snap(graph, pos, file_name)
return density
def removeNodeFromGraph(self):
if nx.is_empty(self.the_graph):
return
i, okPressed = QInputDialog.getInt(self, "Node to delete",
"Delete which node?",
list(self.the_graph.nodes)[0],
list(self.the_graph.nodes)[0],
list(self.the_graph.nodes)[-1], 1)
if okPressed:
self.figure.clf()
try:
self.the_graph.remove_node(i)
except:
pass
nx.draw(self.the_graph, with_labels=True)
self.canvas.draw_idle()
def connect_components(graph, pos_attr, length_attr=None):
""" connects all components by iteratively inserting edges between
components that have minimal distance.
`graph` is a networkx graph object with positions assigned to nodes
`pos_attr` gives the key for the node attribute that stores the position
`length_attr` stores the length of the new edges in this edge attribute
"""
if nx.is_empty(graph):
return graph
# build a distance matrix for all nodes
vertices = nx.get_node_attributes(graph, pos_attr)
nodes = to_array(vertices.keys())
positions = to_array(vertices.values())
dists = distance.squareform(distance.pdist(positions))
if len(vertices) != nx.number_of_nodes(graph):
raise ValueError("Not all nodes have a position specified by the node "
"attribute `%s`" % pos_attr)
# get all subgraphs and build a list of indices into the distance matrix
subgraphs = list(nx.connected_component_subgraphs(graph))
num_subgraphs = len(subgraphs)
# find the index of each node of each subgraph in the nodes array
sg_nids_list = [[np.flatnonzero(nodes == n)[0] for n in sg.nodes()]
for sg in subgraphs]
assert sum(len(s) for s in sg_nids_list) == nx.number_of_nodes(graph)
# initialize result with first subgraph
result = subgraphs.pop(0)
result_nids = sg_nids_list.pop(0)
# iterate until all subgraphs have been added
while subgraphs:
# find subgraph that is closest to `result`
sg_min, nid_sg, nid_res, dist_min = None, None, None, np.inf
for k, sg_nids in enumerate(sg_nids_list):
dist_mat = dists[sg_nids, :][:, result_nids]
x, y = np.unravel_index(dist_mat.argmin(), dist_mat.shape)
dist = dist_mat[x, y]
if dist < dist_min:
sg_min = k # index into the subgraph
dist_min = dist # its distance to `result`
# store the node indices for the connecting edge
nid_sg = sg_nids[x]
nid_res = result_nids[y]
# add graph to `result`
result.add_nodes_from(subgraphs[sg_min].nodes(data=True))
result.add_edges_from(subgraphs[sg_min].edges(data=True))
# add a new edge between the subgraph and `result`
if length_attr is not None:
attr_dict = {length_attr: dists[nid_res, nid_sg]}
else:
attr_dict = None
result.add_edge(nodes[nid_res], nodes[nid_sg], **attr_dict)
# remove the subgraph from the to-do list
result_nids.extend(sg_nids_list.pop(sg_min))
del subgraphs[sg_min]
assert nx.is_connected(result)
assert nx.number_of_nodes(result) == len(vertices)
assert nx.number_of_edges(result) == \
nx.number_of_edges(graph) + num_subgraphs - 1
return result