-
Notifications
You must be signed in to change notification settings - Fork 0
/
embed.py
87 lines (76 loc) · 2.95 KB
/
embed.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
import getopt
from sys import argv
import numpy as np
import matplotlib.pyplot as plt
import networkx as nx
import random
import pickle as pkl
from sklearn.model_selection import ShuffleSplit
from sklearn.multiclass import OneVsRestClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import roc_auc_score, precision_recall_curve, auc, f1_score, accuracy_score, precision_score
from sklearn.metrics.pairwise import linear_kernel, cosine_similarity
from karateclub import BoostNE, NodeSketch, GraRep, NetMF, Node2Vec, DeepWalk, RandNE, GLEE, Walklets, Role2Vec
import argparse
parser = argparse.ArgumentParser(description='Create Graph Embeddings')
parser.add_argument("--graRep", "-g", help="Use GraRep; only use with small datasets", action="store_true")
parser.add_argument("dataset", help="Dataset Name")
parser.add_argument("dim", help="number of dimensions", type=int)
args = parser.parse_args()
dataset = args.dataset
dim = args.dim
if args.graRep:
embedding_names = ['node2vec', 'deepwalk', 'netmf', 'GraRep', 'BoostNE', 'RandNE', 'Walklets', 'Role2Vec']
else:
embedding_names = ['node2vec', 'deepwalk', 'netmf', 'BoostNE', 'RandNE', 'Walklets', 'Role2Vec']
train_g = nx.read_adjlist(f"{dataset}-train.adjlist", nodetype=int)
#train_g = nx.convert_node_labels_to_integers(train_g)
numeric_indices = [index for index in range(train_g.number_of_nodes())]
node_indices = sorted([node for node in train_g.nodes()])
print(numeric_indices == node_indices)
for emb in embedding_names:
print(emb)
if emb == 'node2vec':
model = Node2Vec(dimensions=dim)
model.fit(train_g)
vecs = model.get_embedding()
#vecs = Node2Vec(adj)
elif emb == 'deepwalk':
model = DeepWalk(dimensions=dim)
model.fit(train_g)
vecs = model.get_embedding()
#vecs = DeepWalk(adj)
elif emb == 'netmf':
model = NetMF(dimensions=dim)
model.fit(train_g)
vecs = model.get_embedding()
#vecs = NetMF(adj)
elif emb == 'BoostNE':
model = BoostNE(dimensions=dim, iterations=8)
model.fit(train_g)
vecs = model.get_embedding()
elif emb == 'GraRep':
model = GraRep(dimensions=dim)
model.fit(train_g)
vecs = model.get_embedding()
elif emb == 'NodeSketch':
model = NodeSketch(dimensions=dim, iterations=6, decay=0.01)
model.fit(train_g)
vecs = model.get_embedding()
elif emb == 'GLEE':
model = GLEE(dimensions=dim)
model.fit(train_g)
vecs = model.get_embedding()
elif emb == 'RandNE':
model = RandNE(dimensions=dim)
model.fit(train_g)
vecs = model.get_embedding()
elif emb == 'Walklets':
model = Walklets(dimensions=dim)
model.fit(train_g)
vecs = model.get_embedding()
elif emb == 'Role2Vec':
model = Role2Vec(dimensions=dim)
model.fit(train_g)
vecs = model.get_embedding()
np.save(f'{emb}-{dataset}.npy', vecs)