-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathnewdataset.py
224 lines (206 loc) · 9.2 KB
/
newdataset.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
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
import torch
import torch.utils.data as data
import numpy as np
import os
from tqdm import tqdm
from xarray import open_dataset
def read_file(file):
vars_name = [name for name in file]
return [file.data_vars[name].values for name in vars_name]
#TODO:
#改成对范围内每个站点的dataset :tick
#增加时间监督信息
# start_time分成8类[0,3,6,9,12,15,18,21]
class gridNewDataset(data.Dataset):
def __init__(self, data_path, isTrain=True, isFirstTime=False):
time_class = [0, 3, 6, 9, 12, 15, 18, 21]
self.mean = torch.load(
'/mnt/pami23/stma/weather/processed_data/mean.pth').numpy()
self.std = torch.load(
'/mnt/pami23/stma/weather/processed_data/std.pth').numpy()
self.needed = [0, 8, 14, 17, 22, 28, 31, 35, 40]
if isFirstTime:
# 在这边只扫一遍文件名
self.inputfile = []
self.rainfile = []
self.tempfile = []
for i in tqdm(range(1962), desc="Scanning dataset files"):
file_name = data_path + 'example' + '{:0>5d}'.format(i +
1) + '/'
#由于是预测12-36小时，所以起始时间+12
start_time = 12
if os.path.exists(os.path.join(file_name, '12_12-36h')):
start_time = 0
for j in range(9):
input_file_name = file_name + 'grid_inputs_' + '{:0>2d}'.format(
j + 1) + '.nc'
rain_file_name = file_name + 'obs_grid_rain' + '{:0>2d}'.format(
j + 1) + '.nc'
temp_file_name = file_name + 'obs_grid_temp' + '{:0>2d}'.format(
j + 1) + '.nc'
#如果某个文件不存在，就跳过不要这个数据了吧
if not os.path.isfile(
input_file_name) or not os.path.isfile(
rain_file_name) or not os.path.isfile(
temp_file_name):
start_time += 3
continue
time_classification = time_class.index(start_time)
self.inputfile.append([input_file_name, 0])
self.rainfile.append([rain_file_name, time_classification])
self.tempfile.append([temp_file_name, time_classification])
start_time += 3
if start_time >= 24:
start_time -= 24
np.save('/mnt/pami23/stma/weather/processed_data/newInputFile.npy',
self.inputfile)
np.save('/mnt/pami23/stma/weather/processed_data/newRainFile.npy',
self.rainfile)
np.save('/mnt/pami23/stma/weather/processed_data/newTempFile.npy',
self.tempfile)
else:
self.inputfile = np.load(
'/mnt/pami23/stma/weather/processed_data/newInputFile.npy')
self.rainfile = np.load(
'/mnt/pami23/stma/weather/processed_data/newRainFile.npy')
self.tempfile = np.load(
'/mnt/pami23/stma/weather/processed_data/newTempFile.npy')
self.length = 0
self.input = []
self.rain = []
self.temp = []
self.time = []
print("file len:", len(self.inputfile), len(self.rainfile),
len(self.tempfile))
if isFirstTime:
for idx, [filename, _] in enumerate(tqdm(self.inputfile)):
input_list = self.get_input_list(filename)
rain_list = self.get_label_file(self.rainfile[idx][0])
temp_list = self.get_label_file(self.tempfile[idx][0])
time_classification = self.rainfile[idx][1]
self.input.append(input_list)
self.rain.append(rain_list)
self.temp.append(temp_list)
self.time.append(time_classification)
'''for i in range(8, 61, 1):
for j in range(8, 65, 1):
self.input.append(input_list[:, i - 8:i + 8,
j - 8:j + 8])
self.rain.append(rain_list[i, j])
self.temp.append(temp_list[i, j])
self.time.append(time_classification)
self.length += 1'''
#由于数据量过于大了，我们只存17646个总体文件
np.save('/mnt/pami23/stma/weather/processed_data/newInput.npy',
self.input)
np.save('/mnt/pami23/stma/weather/processed_data/newRain.npy',
self.rain)
np.save('/mnt/pami23/stma/weather/processed_data/newTemp.npy',
self.temp)
np.save('/mnt/pami23/stma/weather/processed_data/newTime.npy',
self.time)
else:
#self.input = np.load('processed_data/newInput.npy')
print('loading input!')
self.input = np.load(
'/mnt/pami23/stma/weather/processed_data/newInput.npy')
print('loading rain!')
self.rain = np.load(
'/mnt/pami23/stma/weather/processed_data/newRain.npy')
print('loading temp!')
self.temp = np.load(
'/mnt/pami23/stma/weather/processed_data/newTemp.npy')
print('loading time!')
self.time = np.load(
'/mnt/pami23/stma/weather/processed_data/newTime.npy')
file_len = len(self.inputfile)
train_file_len = int(0.9 * file_len)
total_len = file_len * 53 * 57
train_len = train_file_len * 53 * 57
if isTrain:
self.input = self.input[:train_file_len]
self.rain = self.rain[:train_file_len]
self.temp = self.temp[:train_file_len]
self.time = self.time[:train_file_len]
self.length = train_len
else:
self.input = self.input[train_file_len:]
self.rain = self.rain[train_file_len:]
self.temp = self.temp[train_file_len:]
self.time = self.time[train_file_len:]
self.length = total_len - train_len
print('length:', self.length)
print(len(self.input))
def get_input_list(self, input_file_name):
input = open_dataset(input_file_name)
input_values = read_file(input)
temp_list = []
i = 0
for values in input_values:
if values.ndim == 3:
values = np.transpose(values, (2, 0, 1))
for num in range(values.shape[0]):
#temp_list.append(values[num].flatten().tolist())
if i in self.needed:
temp_list.append((values[num] - self.mean[i]) /
self.std[i].tolist())
i += 1
else:
#temp_list.append(values.flatten().tolist())
temp_list.append(
(values - self.mean[i]) / self.std[i].tolist())
i += 1
input = np.asarray(temp_list)
return input
def get_label_file(self, label_file_name):
label = open_dataset(label_file_name)
label_values = read_file(label)
return label_values[0]
def __len__(self):
return self.length
def __getitem__(self, idx):
#53*57=3021
real_id = int(idx / 3021)
inner_id = int(idx % 3021)
total_input = torch.from_numpy(self.input[real_id])
total_rain = torch.from_numpy(self.rain[real_id])
total_temp = torch.from_numpy(self.temp[real_id])
time = self.time[real_id]
i = int(inner_id / 57) + 8
j = int(inner_id % 57) + 8
input = total_input[:, i - 8:i + 9, j - 8:j + 9]
rain = total_rain[i, j]
temp = total_temp[i, j]
return input, rain, temp, time
if __name__ == "__main__":
dataset = gridNewDataset("/mnt/pami23/stma/weather/train/",
isTrain=True,
isFirstTime=False)
'''mean = torch.zeros(58)
std = torch.zeros(58)
for idx in tqdm(range(dataset.length)):
input = open_dataset(dataset.input[idx])
input_values = read_file(input)
i = 0
for values in input_values:
if values.ndim == 3:
values = np.transpose(values, (2, 0, 1))
for num in range(values.shape[0]):
#temp_list.append(values[num].flatten().tolist())
#temp_list.append(values[num].tolist())
mean[i] += values[num].mean()
std[i] += values[num].std()
i += 1
else:
#temp_list.append(values.flatten().tolist())
#temp_list.append(values.tolist())
mean[i] += values.mean()
std[i] += values.std()
i += 1
mean.div_(dataset.length)
std.div_(dataset.length)
print(mean, std)
torch.save(mean, 'processed_data/mean.pth')
torch.save(std, 'processed_data/std.pth')'''
input, rain, temp, time = dataset.__getitem__(47976500)
print(input.shape, rain, temp, time)