def createGraphML():
global g
g = Graph()
c.execute('select uid from userdata')
dataList = c.fetchall()
gnodes=[]
edges=[]
for i in dataList:
i=str(i)
i = i.replace("(","").replace(",)","").replace("L","")
i= int(i)
c.execute('select distinct low from graphdata where high=?', (i,))
relate=c.fetchall()
if not i in gnodes:
g.add_node(i)
gnodes.append(i)
for e in relate:
e=str(e)
e = e.replace("(","").replace(",)","").replace("L","")
e=int(e)
if not e in gnodes:
g.add_node(e)
gnodes.append(e)
g.add_edge_by_label(str(i), str(e))
parser = GraphMLParser()
parser.write(g, "myGraph.graphml")
def createGraphML():
global g
g = Graph()
c.execute('select uid from userdata')
dataList = c.fetchall()
gnodes = []
edges = []
for i in dataList:
i = str(i)
i = i.replace("(", "").replace(",)", "").replace("L", "")
i = int(i)
c.execute('select distinct low from graphdata where high=?', (i, ))
relate = c.fetchall()
if not i in gnodes:
g.add_node(i)
gnodes.append(i)
for e in relate:
e = str(e)
e = e.replace("(", "").replace(",)", "").replace("L", "")
e = int(e)
if not e in gnodes:
g.add_node(e)
gnodes.append(e)
# edges.append(i)
# edges.append(e)
#1 if edges2.count(e) > 1:
g.add_edge_by_label(str(i), str(e))
parser = GraphMLParser()
parser.write(g, "myGraph.graphml")
def __write_graphML(self):
"""Writes the .graphml file.
Uses class Graph from pygraphml library to create the graph.
Also it uses class GraphMLParser from pygraphml library to write the file.
"""
gr = Graph()
# Adding nodes in the graph with properties.
for sub in self.nodes_prop.keys():
n = gr.add_node(sub)
for pair_prop in self.nodes_prop[sub]:
n[pair_prop[0]] = pair_prop[1]
# Adding nodes in the graph without properties.
for node in self.nodes.values():
if node not in self.nodes_prop.keys():
gr.add_node(node)
# Checking the relations between nodes and creating respective edges.
for relation in self.relations:
source = self.nodes[relation[0]]
target = self.nodes[relation[2]]
edge = gr.add_edge_by_label(source, target)
edge.set_directed(True)
edge['model'] = relation[1]
# Writting the file.
parser = GraphMLParser()
file_name = self.file_name.split(".")
file_name = file_name[0]
parser.write(gr, file_name + ".graphml")
print("File " + file_name + ".graphml is succesfully written")
def generate_graph(similarity, attr): from pygraphml import Graph items = similarity.items() labels = np.array([artist for artist, obj in items]) network = np.zeros((labels.size, labels.size)) g = Graph() for artist in labels: g.add_node(artist) for artist_id, x in enumerate(items): network[artist_id, artist_id] = 1 artist, obj = x if attr == "emotion": for idx, score in enumerate(obj["emotion_sim"]): if network[artist_id, idx] == 0 and network[idx, artist_id] == 0: edge = g.add_edge_by_label(labels[artist_id], labels[idx]) if score == 0: edge["weight"] = 0.001 else: edge["weight"] = score network[artist_id, idx] = 1 network[idx, artist_id] = 1 elif attr == "topic": for idx, score in enumerate(obj["topic_sim"]): if network[artist_id, idx] == 0 and network[idx, artist_id] == 0: edge = g.add_edge_by_label(labels[artist_id], labels[idx]) if score == 0: edge["weight"] = 0.001 else: edge["weight"] = score network[artist_id, idx] = 1 network[idx, artist_id] = 1 return g
def generate_graph_with_clusters(summary, attr): from pygraphml import Graph items = summary.items() labels = np.array([artist for artist, obj in items]) g = Graph() for artist in labels: g.add_node(artist) if attr == "emotion": for category in EMOTION_CATEGORIES: g.add_node(category) elif attr == "topic": for category in TOPIC_CATEGORIES: g.add_node(category) for artist_id, x in enumerate(items): arist, obj = x if attr == "emotion": for idx, score in enumerate(obj["emotions"]): edge = g.add_edge_by_label(EMOTION_CATEGORIES[idx], labels[artist_id]) if score == 0: edge["weight"] = 0.001 # Set to very small value else: edge["weight"] = score elif attr == "topic": for idx, score in enumerate(obj["topics"]): edge = g.add_edge_by_label(TOPIC_CATEGORIES[idx], labels[artist_id]) if score == 0: edge["weight"] = 0.001 # Set to very small value else: edge["weight"] = score return g