/
data_partition.py
144 lines (104 loc) · 4.79 KB
/
data_partition.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
import os
import numpy as np
def get_dir_label(data_dir):
cate_list = os.listdir(os.path.join(data_dir, 'Dataset_v3_vision'))
if '.DS_Store' in cate_list:
cate_list.remove('.DS_Store')
class_id_dict = {}
for i in range(len(cate_list)):
current_class = cate_list[i]
class_dir = os.path.join(data_dir, 'Dataset_v3_vision', current_class)
sample_list = os.listdir(class_dir)
if '.DS_Store' in sample_list:
sample_list.remove('.DS_Store')
data_samples = []
for sample in sample_list:
data_samples.append(os.path.join(current_class, sample[:-4]))
class_id_dict[i] = data_samples
return class_id_dict
def integrate_sample_id(sample_list):
sample_id_dict = {}
for i in range(len(sample_list)):
sample_ = sample_list[i].split('/')
sample_id = sample_[1][:5]
if sample_id not in sample_id_dict:
sample_id_dict[sample_id] = []
sample_id_dict[sample_id].append(sample_list[i])
return sample_id_dict
def get_samples(sample_id_dict, samples):
sample_list = []
for i in range(len(samples)):
sample_list.extend(sample_id_dict[samples[i]])
return sample_list
def data_splitter(class_id_dict, train_ratio, val_ratio):
id_list = list(class_id_dict.keys())
train_sample = []
train_label = []
test_sample = []
test_label = []
val_sample = []
val_label = []
for i in range(len(id_list)):
samples = class_id_dict[i]
np.random.seed(i)
np.random.shuffle(samples)
train_num = int(np.floor(len(samples)*train_ratio))
val_num = int(np.floor(len(samples)*val_ratio))
train_sample.extend(samples[:train_num])
train_label.extend([i for k in range(train_num)])
val_sample.extend(samples[train_num:(train_num+val_num)])
val_label.extend([i for k in range(val_num)])
test_sample.extend(samples[(train_num+val_num):])
test_label.extend([i for k in range(len(samples)-train_num-val_num)])
return (train_sample, train_label, val_sample, val_label, test_sample, test_label)
def data_splitter_by_id(class_id_dict, train_ratio, val_ratio):
id_list = list(class_id_dict.keys())
train_sample = []
train_label = []
test_sample = []
test_label = []
val_sample = []
val_label = []
for i in range(len(id_list)):
samples = class_id_dict[i]
sample_id_dict = integrate_sample_id(samples)
sample_ids = list(sample_id_dict.keys())
np.random.seed(i)
np.random.shuffle(sample_ids)
train_num = int(np.floor(len(sample_ids)*train_ratio))
val_num = int(np.floor(len(sample_ids)*val_ratio))
current_train_sample_id = sample_ids[:train_num]
current_val_sample_id = sample_ids[train_num:(train_num+val_num)]
current_test_sample_id = sample_ids[(train_num+val_num):]
current_train_sample = get_samples(sample_id_dict, current_train_sample_id)
train_sample.extend(current_train_sample)
train_label.extend([i for k in range(len(current_train_sample))])
current_val_sample = get_samples(sample_id_dict, current_val_sample_id)
val_sample.extend(current_val_sample)
val_label.extend([i for k in range(len(current_val_sample))])
current_test_sample = get_samples(sample_id_dict, current_test_sample_id)
test_sample.extend(current_test_sample)
test_label.extend([i for k in range(len(current_test_sample))])
return (train_sample, train_label, val_sample, val_label, test_sample, test_label)
def cate_data_sample(class_id_dict, train_ratio, class_id):
id_list = list(class_id_dict.keys())
samples = class_id_dict[class_id]
np.random.seed(class_id)
np.random.shuffle(samples)
train_num = int(np.floor(len(samples)*train_ratio))
train_sample = samples[:train_num]
return train_sample
def data_construction(data_dir, train_ratio=0.7, val_ratio=0.1):
class_id_dict = get_dir_label(data_dir)
(train_sample, train_label, val_sample, val_label, test_sample, test_label) = data_splitter_by_id(class_id_dict, train_ratio, val_ratio)
return (train_sample, train_label, val_sample, val_label, test_sample, test_label)
def audio_construction(data_dir, train_ratio=0.7, val_ratio=0.1):
class_id_dict = get_dir_label(data_dir)
(train_sample, train_label, val_sample, val_label, test_sample, test_label) = data_splitter_by_id(class_id_dict, train_ratio, val_ratio)
train_sample.extend(val_sample)
train_sample.extend(test_sample)
return train_sample
def single_category_construction(data_dir, train_ratio=0.7, class_id=0):
class_id_dict = get_dir_label(data_dir)
data_sample = cate_data_sample(class_id_dict, train_ratio, class_id)
return data_sample