-
Notifications
You must be signed in to change notification settings - Fork 5
/
Copy pathgenerate_PEMS_data.py
141 lines (113 loc) · 5.35 KB
/
generate_PEMS_data.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
131
132
133
134
135
136
137
138
139
140
141
import torch
from torch_geometric_temporal.dataset import METRLADatasetLoader, PemsBayDatasetLoader
import pandas as pd
import networkx as nx
import numpy as np
def get_adjacency_matrix(distance_df_filename, num_of_vertices, id_filename=None):
'''
Parameters
----------
distance_df_filename: str, path of the csv file contains edges information
num_of_vertices: int, the number of vertices
Returns
----------
A: np.ndarray, adjacency matrix
'''
if 'npy' in distance_df_filename:
adj_mx = np.load(distance_df_filename)
return adj_mx, None
else:
import csv
A = np.zeros((int(num_of_vertices), int(num_of_vertices)),
dtype=np.float32)
distaneA = np.zeros((int(num_of_vertices), int(num_of_vertices)),
dtype=np.float32)
if id_filename:
with open(id_filename, 'r') as f:
id_dict = {int(i): idx for idx, i in enumerate(f.read().strip().split('\n'))} # 把节点id（idx）映射成从0开始的索引
with open(distance_df_filename, 'r') as f:
f.readline()
reader = csv.reader(f)
for row in reader:
if len(row) != 3:
continue
i, j, distance = int(row[0]), int(row[1]), float(row[2])
A[id_dict[i], id_dict[j]] = 1
distaneA[id_dict[i], id_dict[j]] = distance
return A, distaneA
else:
with open(distance_df_filename, 'r') as f:
f.readline()
reader = csv.reader(f)
for row in reader:
if len(row) != 3:
continue
i, j, distance = int(row[0]), int(row[1]), float(row[2])
A[i, j] = 1
distaneA[i, j] = distance
return A, distaneA
def generate_PEMS_dataset_adj(dataset_name,device, data_used_subset= False, subset_percent = 0.2):
file_path = "/home/mingxi/deep_learning_implementation/flow_data/data/"+ dataset_name +"/" + dataset_name
if dataset_name == "PEMS04_speed":
file_path = "/home/mingxi/deep_learning_implementation/flow_data/data/" + 'PEMS04' + "/" + 'PEMS04'
if dataset_name == "PEMS08_speed":
file_path = "/home/mingxi/deep_learning_implementation/flow_data/data/" + 'PEMS08' + "/" + 'PEMS08'
if dataset_name == 'PEMSD7':
# PEMSD7: Time(5/1/2012 - 6/30/2012, weekdays), Nodes(228)
speed_matrix = pd.read_csv("~/deep_learning_implementation/flow_data/PEMSD7/V_228.csv")
speed_matrix = speed_matrix.values
adj_matrix = pd.read_csv("~/deep_learning_implementation/flow_data/PEMSD7/weighted_adj.csv", index_col=None,header=None)
adj_matrix = adj_matrix.values
if dataset_name != 'PEMSD7':
##dataset_name
# PEMS04: Time(1/1/2018 - 2/28/2018), Nodes(307)
# PEMS07: Time(5/1/2017 - 8/31/2017), Nodes(883)
# PEMS08: Time(7/1/2016 - 8/31/2016), Nodes(170)
speed_matrix = np.load( file_path + ".npz")['data']
if dataset_name != "PEMS04_speed" or dataset_name != "PEMS08_speed":
speed_matrix = speed_matrix[:, :, 0]
print(speed_matrix)
else:
speed_matrix = speed_matrix[:, :, 1]
speed_matrix = speed_matrix.clip(0,100)
max_dp = speed_matrix.max()
num_nodes = speed_matrix.shape[1]
_ , adj_matrix = get_adjacency_matrix(file_path + ".csv", num_nodes)
if data_used_subset == True:
start_index = np.random.randint(0, (1 - subset_percent) * speed_matrix.shape[0])
end_index = int(start_index + subset_percent * speed_matrix.shape[0])
speed_matrix = speed_matrix.iloc[start_index, end_index]
#speed_matrix = torch.tensor(speed_matrix)
# mean, std = speed_matrix.mean(), speed_matrix.std()
# speed_matrix = (speed_matrix - mean)/std
max = np.max(speed_matrix)
speed_matrix = torch.tensor(speed_matrix) / max
X = torch.unsqueeze(speed_matrix, 2)
num_nodes = X.shape[1]
adj = nx.convert_matrix.from_numpy_matrix(adj_matrix,parallel_edges=True,create_using=nx.MultiGraph)
sp_L = nx.linalg.laplacianmatrix.normalized_laplacian_matrix(adj)
rows, cols = sp_L .nonzero()
data = sp_L[rows, cols]
indicies = []
indicies.append(rows)
indicies.append(cols)
indicies= torch.tensor(indicies)
data = torch.tensor(data, dtype=torch.float32).squeeze(0)
sp_L = torch.sparse_coo_tensor(indicies, data).to(device)
edgelist = [(u, v) for (u, v) in adj.edges()]
edge_index = torch.tensor(edgelist)
edge_index = edge_index.transpose(0,1)
edge_attr = []
for edge in edgelist:
edge_attr.append(adj_matrix[edge[0]][edge[1]])
edge_attr = torch.tensor(edge_attr, dtype=torch.float32).unsqueeze(1)
return X, sp_L, edgelist, edge_index, edge_attr, num_nodes, max
# X, sp_L, edgelist, edge_index, edge_attr, num_nodes, mean, std = generate_flow_dataset_adj('PEMSD7', device)
#
# X_, sp_L_, edgelist_, edge_index_, edge_attr_, num_nodes_, mean_, std_ = generate_flow_dataset_adj('PEMS04', device)
# X_1, _, _, _, _, _, mean__, std__ = generate_flow_dataset_adj('PEMS07', device)
# X_2, _, _, _, _, _, _mean, _std= generate_flow_dataset_adj('PEMS08', device)
# print(X.shape)
# print(X_.shape)
# print(X_1.shape)
# print(X_2.shape)