Repository navigation
Expand file tree
/
Copy pathdata_preprocess.py
More file actions
194 lines (155 loc) · 7.38 KB
/
Copy pathdata_preprocess.py
File metadata and controls
194 lines (155 loc) · 7.38 KB
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
import random
import shutil
import os
from distutils.dir_util import copy_tree
import pandas as pd
###############################################################
# PARAMETERS
###############################################################
#path_of original dataset
path_original_dataset_train = "/mnt/disco_esterno/brain_tumor_dataset/brain_tumor_images/Training"
path_original_dataset_test = "/mnt/disco_esterno/brain_tumor_dataset/brain_tumor_images/Testing"
#path of final dataset
root = "/mnt/disco_esterno/brain_tumor_dataset/brain_tumor_images/Final_Dataset_V1"
path_dataset_train = os.path.join(root, 'train')
path_dataset_test = os.path.join(root, 'test')
path_dataset_val = os.path.join(root, 'val')
# path of csv files
path_csv_train = os.path.join(root, 'train.csv')
path_csv_test = os.path.join(root, 'test.csv')
path_csv_val = os.path.join(root, 'val.csv')
PERC_TRAIN = 0.8
df_train = pd.DataFrame(columns=['CLASS', 'PATH', 'LABEL'])
df_test = pd.DataFrame(columns=['CLASS', 'PATH', 'LABEL'])
df_val = pd.DataFrame(columns=['CLASS', 'PATH', 'LABEL'])
###############################################################
# FUNZIONI
###############################################################
def create_csv(path_dataset, df, class2label, type_dataset):
for subdir, dirs, files in os.walk(path_dataset):
for cl in dirs:
path_cl = os.path.join(path_dataset, cl)
CHECK_FOLDER = os.path.isdir(path_cl)
if CHECK_FOLDER:
print("class_name: ", cl)
label = class2label[cl]
list_filenames_food = os.listdir(path_cl)
for file_name in list_filenames_food:
relative_path = os.path.join(type_dataset, cl, file_name)
df = df.append({'CLASS': cl,
'PATH': relative_path,
'LABEL': label}, ignore_index=True)
return df
def make_train_val_division(path_original_dataset, path_dataset_train, path_dataset_val):
list_classes = []
print(" TRAIN/ VAL SPLIT")
for subdir, dirs, files in os.walk(path_original_dataset):
for cl in dirs:
path_cl = os.path.join(path_original_dataset, cl)
CHECK_FOLDER = os.path.isdir(path_cl)
if CHECK_FOLDER:
print("class_name: ", cl)
list_classes.append(cl)
list_filenames_cl = os.listdir(path_cl)
number_files = len(list_filenames_cl)
index_list = [i for i in range(number_files)]
print("number_files in directory '{}': {}".format(path_cl, number_files))
number_file_train = round(number_files * PERC_TRAIN)
train_index_list = random.sample(index_list, k=number_file_train)
val_index_list = list(set(index_list) - set(train_index_list))
# copy files in train directory
dst_dir = os.path.join(path_dataset_train, cl)
CHECK_FOLDER = os.path.isdir(dst_dir)
if not CHECK_FOLDER:
os.makedirs(dst_dir)
for idx in train_index_list:
filename = list_filenames_cl[idx]
src_file = os.path.join(path_cl, filename)
dst_file = os.path.join(dst_dir, filename)
shutil.copy2(src_file, dst_file)
print("Number of files in dir '{}': {}".format(dst_dir, len(os.listdir(dst_dir))))
# copy files in val directory
dst_dir = os.path.join(path_dataset_val, cl)
CHECK_FOLDER = os.path.isdir(dst_dir)
if not CHECK_FOLDER:
os.makedirs(dst_dir)
for idx in val_index_list:
filename = list_filenames_cl[idx]
src_file = os.path.join(path_cl, filename)
dst_file = os.path.join(dst_dir, filename)
shutil.copy2(src_file, dst_file)
print("Number of files in dir '{}': {}".format(dst_dir, len(os.listdir(dst_dir))))
print('----------------------------------------------------------------')
return list_classes
def clean_create_dir(path_dir):
CHECK_FOLDER = os.path.isdir(path_dir)
if CHECK_FOLDER:
print("The directory '{}' exists. Delete it".format(path_dir))
try:
shutil.rmtree(path_dir)
except OSError as e:
print("Error: {}".format(e.strerror))
raise e
CHECK_FOLDER = os.path.isdir(path_dir)
if not CHECK_FOLDER:
print("Create the directory '{}'".format(path_dir))
os.makedirs(path_dir)
else:
print("Create the directory '{}'".format(path_dir))
os.makedirs(path_dir)
###############################################################
# EXECUTION
###############################################################
if __name__ == "__main__":
# 1 - clean and create the dataset directory
clean_create_dir(root)
# 2 - copy in the dataset directory the "Testing" dataset renaiming it as Test
copy_tree(path_original_dataset_test, path_dataset_test)
# 3 - do split train/val of the train dataset
list_classes = make_train_val_division(path_original_dataset_train, path_dataset_train, path_dataset_val)
class2label = {k: v for (v, k) in enumerate(list_classes)}
print('-----------------------------------------------')
print('-----------------------------------------------')
print("list_classes: ", list_classes)
print("")
print("list_classes: ", ','.join(list_classes))
print("")
print("class2label: ", class2label)
print('-----------------------------------------------')
print('-----------------------------------------------')
# 4 - create csv files
print("Create csv files")
print("Train set")
df_train = create_csv(path_dataset_train, df_train, class2label, "train")
print("Test set")
df_test = create_csv(path_dataset_test, df_test, class2label, "test")
print("Val set")
df_val = create_csv(path_dataset_val, df_val, class2label, "val")
print('-------------------------------------------------------------')
print('-------------------------------------------------------------')
# 5 - shuffling and saving of csv
df_train = df_train.sample(frac=1).reset_index(drop=True)
df_test = df_test.sample(frac=1).reset_index(drop=True)
df_val = df_val.sample(frac=1).reset_index(drop=True)
df_train.to_csv(path_csv_train, index=False)
df_test.to_csv(path_csv_test, index=False)
df_val.to_csv(path_csv_val, index=False)
print("df_train info")
print(df_train.info())
print('-------------------------------------------------------------')
print("")
print("df_test info")
print(df_test.info())
print('-------------------------------------------------------------')
print("")
print("df_val info")
print(df_val.info())
print('-------------------------------------------------------------')
print("df_train: CLASS values count")
print(df_train[["CLASS"]].value_counts())
print('-------------------------------------------------------------')
print("df_test: CLASS values count")
print(df_test[["CLASS"]].value_counts())
print('-------------------------------------------------------------')
print("df_val: CLASS values count")
print(df_val[["CLASS"]].value_counts())