Reconstruct Graph from Input data

builder.py

+'''Reads the graph and node contexts'''
+
+import networkx as nx
+
+def build_graph():
+    '''Read input network''' 
+
+    G = nx.read_edgelist('input/cora.edgelist', nodetype=int, create_using=nx.Graph())
+    for e in G.edges:
+        G.edges[e]['weight'] = 1 
+
+    return G
+
+def node_labels():
+    '''Read node classes'''
+    
+    G = build_graph()
+    d = {}
+    with open('input/coralabels.edgelist', 'r') as lab:    
+        l = lab.readlines()     
+    
+        if len(l[0].split(' ')) == 2:
+                x = [int(w.split(' ')[0]) for w in l]
+                y = [int(w.split(' ')[-1].rstrip('\n')) for w in l]  
+                for j in range(len(x)):
+                    d[x[j]] = y[j]      #For label data with node          
+        elif len(l[0].split(' ')) == 1: 
+            for i, k in enumerate(l, 1):    #start counting from 1
+                d[i] = int(k.rstrip('\n'))    #For label data without node
+        else: print('Edges more than 2 nodes')
+        labels_dict = {}
+        for node in G.nodes:
+            labels_dict[node] = d[node]
+        return labels_dict
+            
+        
+def node_neighbor_labels():
+    G = build_graph()
+    labels_dict = node_labels()
+    neighbor_labels = {}
+    for node in G.nodes:
+        nn = list(G.neighbors(node))    
+        nnlabel = [labels_dict[n] for n in nn]  #Superimpose labels on node id
+        neighbor_labels[node] = nnlabel
+    return neighbor_labels, labels_dict
\ No newline at end of file