-
Notifications
You must be signed in to change notification settings - Fork 1
/
loaddatas.py
133 lines (116 loc) · 5.32 KB
/
loaddatas.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
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
import torch_geometric.datasets
from torch_geometric.data import Data
import torch_geometric.transforms as T
import torch
import sys
import networkx as nx
import os
import numpy as np
import scipy.sparse as sp
from torch_geometric.utils import remove_self_loops
import torch_geometric.datasets
from sg2dgm import riccidist2dgm as sg2dgm
#from sg2dgm import riccidist2dgm_c as sg2dgm
def loaddatas(d_name):
if d_name in ["PPI"]:
dataset = torch_geometric.datasets.PPI('./data/' + d_name)
elif d_name == 'Cora':
dataset = torch_geometric.datasets.Planetoid('./data/'+d_name,d_name,transform=T.NormalizeFeatures())
elif d_name in ['Citeseer', 'PubMed']:
dataset = torch_geometric.datasets.Planetoid('./data/' + d_name, d_name)
elif d_name in ["Computers", "Photo"]:
dataset = torch_geometric.datasets.Amazon('./data/'+d_name,d_name)
return dataset
def get_edges_split(data, val_prop = 0.2, test_prop = 0.2, seed = 1234):
g = nx.Graph()
g.add_nodes_from([i for i in range(len(data.y))])
ricci_edge_index_ = np.array((data.edge_index))
ricci_edge_index = [(ricci_edge_index_[0, i], ricci_edge_index_[1, i]) for i in
range(np.shape(ricci_edge_index_)[1])]
g.add_edges_from(ricci_edge_index)
adj = nx.adjacency_matrix(g)
return get_adj_split(adj,val_prop = val_prop, test_prop = test_prop, seed = seed)
#def get_adj_split(adj, val_prop = 0.05, test_prop = 0.1, seed=1234):
def get_adj_split(adj, val_prop=0.05, test_prop=0.1, seed=1234):
np.random.seed(seed) # get tp edges
x, y = sp.triu(adj).nonzero()
pos_edges = np.array(list(zip(x, y)))
np.random.shuffle(pos_edges)
# get tn edges
x, y = sp.triu(sp.csr_matrix(1. - adj.toarray())).nonzero()
neg_edges = np.array(list(zip(x, y)))
np.random.shuffle(neg_edges)
m_pos = len(pos_edges)
n_val = int(m_pos * val_prop)
n_test = int(m_pos * test_prop)
val_edges, test_edges, train_edges = pos_edges[:n_val], pos_edges[n_val:n_test + n_val], pos_edges[n_test + n_val:]
val_edges_false, test_edges_false = neg_edges[:n_val], neg_edges[n_val:n_test + n_val]
train_edges_false = np.concatenate([neg_edges, val_edges, test_edges], axis=0)
return train_edges, train_edges_false, val_edges, val_edges_false, test_edges, test_edges_false
def compute_persistence_image(data, train_edges, train_edges_false, val_edges, val_edges_false, test_edges, test_edges_false, data_name, hop = 1):
if data_name == "photo":
data_name = "Photo"
if data_name == "computers":
data_name = "Computers"
filename = './data/TLCGNN/' + data_name + '.npy'
if os.path.exists(filename):
return np.load(filename)
total_edges = np.concatenate(
(train_edges, train_edges_false, val_edges, val_edges_false, test_edges, test_edges_false))
data.train_pos, data.train_neg = len(train_edges), len(train_edges_false)
data.val_pos, data.val_neg = len(val_edges), len(val_edges_false)
data.test_pos, data.test_neg = len(test_edges), len(test_edges_false)
data.total_edges = total_edges
# delete val_pos and test_pos
edge_list = np.array(data.edge_index).T.tolist()
for edges in val_edges:
edges = edges.tolist()
if edges in edge_list:
edge_list.remove(edges)
edge_list.remove([edges[1], edges[0]])
for edges in test_edges:
edges = edges.tolist()
if edges in edge_list:
edge_list.remove(edges)
edge_list.remove([edges[1], edges[0]])
data.edge_index = torch.Tensor(edge_list).long().transpose(0, 1)
data.edge_index, _ = remove_self_loops(data.edge_index)
# generate graph for computing persistence diagram
g = nx.Graph()
ricci_edge_index_ = np.array(remove_self_loops((data.edge_index.cpu()))[0])
ricci_edge_index = [(ricci_edge_index_[0, i], ricci_edge_index_[1, i]) for i in
range(np.shape(ricci_edge_index_)[1])]
g.add_edges_from(ricci_edge_index)
print(len(g.edges()))
# ricci_cur = compute_ricci_flow(data, d_name)
ricci_cur = compute_ricci_curvature(data)
# compute sg2dgm and save in a dict
pi = sg2dgm.graph2pi(g, ricci_curv=ricci_cur)
pi.get_pimg_for_all_edges(total_edges, cores=16, hop=hop, norm=True, extended_flag=True,
resolution=5, descriptor='sum')
np.save(filename,pi.pi_sg)
return pi.pi_sg
def compute_ricci_curvature(data):
from GraphRicciCurvature.OllivierRicci import OllivierRicci
print("start writing ricci curvature")
Gd = nx.Graph()
ricci_edge_index_ = np.array(data.edge_index)
ricci_edge_index = [(ricci_edge_index_[0, i],
ricci_edge_index_[1, i]) for i in
range(np.shape(data.edge_index)[1])]
Gd.add_edges_from(ricci_edge_index)
Gd_OT = OllivierRicci(Gd, alpha=0.5, method="Sinkhorn", verbose="INFO")
print("adding edges finished")
Gd_OT.compute_ricci_curvature()
ricci_list = []
for n1, n2 in Gd_OT.G.edges():
ricci_list.append([n1, n2, Gd_OT.G[n1][n2]['ricciCurvature']])
ricci_list.append([n2, n1, Gd_OT.G[n1][n2]['ricciCurvature']])
ricci_list = sorted(ricci_list)
print("computing ricci curvature finished")
return ricci_list
def num(strings):
try:
return int(strings)
except ValueError:
return float(strings)