MultiTest/copy_data.py at master · paprays/MultiTest · GitHub

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
import os
import shutil

import numpy as np
from natsort import natsort
from tqdm import tqdm

import config as cf
import config.common_config
from init import symlink


def copy_data4epnet(dir_seq):
    system_name = "EPNet"
    return copy_data(dir_seq, system_name)


def copy_data4fconv(dir_seq):
    system_name = "FConv"
    return copy_data(dir_seq, system_name)


def copy_data4clocs(dir_seq):
    system_name = "CLOCs"
    return copy_data(dir_seq, system_name)


def copy_data4second(dir_seq):
    system_name = "Second"
    return copy_data(dir_seq, system_name)


def copy_data4rcnn(dir_seq):
    system_name = "Rcnn"
    return copy_data(dir_seq, system_name)


def copy_data(dir_seq, system_name):
    kitti_trainval_txt_file = "trainval.txt"
    kitti_val_txt_file = "val.txt"
    kitti_root_dir = cf.common_config.kitti_dataset_root
    kitti_aug_dir = cf.common_config.workplace_dir
    kitti_aug_dir = os.path.join(kitti_aug_dir, system_name, "kitti")
    os.makedirs(kitti_aug_dir, exist_ok=True)
    if os.path.exists(kitti_aug_dir):
        shutil.rmtree(kitti_aug_dir)
    kitti_imagesets_dir = os.path.join(kitti_aug_dir, "ImageSets")
    kitti_testing = os.path.join(kitti_aug_dir, "testing")
    os.makedirs(kitti_imagesets_dir, exist_ok=True)
    os.makedirs(kitti_testing, exist_ok=True)
    kitti_aug_train_val = os.path.join(kitti_imagesets_dir, kitti_trainval_txt_file)
    kitti_aug_val = os.path.join(kitti_imagesets_dir, kitti_val_txt_file)
    for _n in ["train", "test", "trainval"]:
        _p = os.path.join(kitti_imagesets_dir, _n + ".txt")
        with open(_p, "w") as f:
            ...

    if os.path.exists(kitti_aug_train_val):
        os.remove(kitti_aug_train_val)
    if os.path.exists(kitti_aug_val):
        os.remove(kitti_aug_val)
    kitti_aug_dir = os.path.join(kitti_aug_dir, "training")
    kitti_root_dir = os.path.join(kitti_root_dir, "training")
    os.makedirs(kitti_aug_dir, exist_ok=True)

    kitti_aug_dir_calib = os.path.join(kitti_aug_dir, "calib")
    kitti_root_dir_calib = os.path.join(kitti_root_dir, "calib")

    sub_dirs = ["image_2", "label_2", "calib", "velodyne", "velodyne_reduced"]
    for sub_dir in sub_dirs:
        os.makedirs(os.path.join(kitti_aug_dir, sub_dir), exist_ok=True)
    queue_dir = cf.common_config.save_dir_guided
    queue_dir = os.path.join(queue_dir, system_name)
    dir_seq_arr = []
    if not os.path.exists(os.path.join(queue_dir, dir_seq)):
        raise ValueError(os.path.join(queue_dir, dir_seq))
    image_dir = os.path.join(queue_dir, dir_seq, "image_2")
    to_image_dir = os.path.join(kitti_aug_dir, "image_2")

    pc_dir = os.path.join(queue_dir, dir_seq, "velodyne")
    to_pc_dir = os.path.join(kitti_aug_dir, "velodyne")

    insert_label_dir = os.path.join(queue_dir, dir_seq, "label_2")
    to_label_dir = os.path.join(kitti_aug_dir, "label_2")

    fns = natsort.natsorted(os.listdir(image_dir))
    if len(fns) == 0:
        raise ValueError(f"{image_dir}")
    else:

        dir_seq_arr.append(dir_seq)
        fn = fns[-1]

    shutil.copyfile(os.path.join(image_dir, fn), os.path.join(to_image_dir, "{}.png".format(dir_seq)))

    shutil.copyfile(os.path.join(kitti_root_dir_calib, "{}.txt".format(dir_seq)),
                    os.path.join(kitti_aug_dir_calib, "{}.txt".format(dir_seq)))

    shutil.copyfile(os.path.join(pc_dir, "{}.bin".format(dir_seq)),
                    os.path.join(to_pc_dir, "{}.bin".format(dir_seq)))

    shutil.copyfile(os.path.join(insert_label_dir, "{}.txt".format(dir_seq)),
                    os.path.join(to_label_dir, "{}.txt".format(dir_seq)))

    with open(kitti_aug_train_val, "a") as f:
        for seq in dir_seq_arr:
            f.writelines(str(seq) + "\n")

    with open(kitti_aug_val, "a") as f:
        for seq in dir_seq_arr:
            f.writelines(str(seq) + "\n")


def copy_ori_data4retrain(target_dir, train_size=100):
    kitti_dataset_root = config.common_config.kitti_dataset_root
    kitti_dataset_data_root = os.path.join(kitti_dataset_root, "training")
    kitti_imagesets_dir = os.path.join(kitti_dataset_root, "ImageSets")
    kitti_train_split = os.path.join(kitti_imagesets_dir, "train.txt")
    with open(kitti_train_split, "r") as f:
        idx_arr = f.readlines()
    idx_arr = [x.strip() for x in idx_arr]
    if idx_arr[-1] == "":
        del idx_arr[-1]
    np.random.seed(0)
    if train_size == "ALL":
        idx_train = np.array(idx_arr)
    else:
        idx_selected = np.random.permutation(len(idx_arr))[:train_size]
        idx_train = np.array(idx_arr)[idx_selected]
    sub_dirs = ["image_2", "label_2", "calib", "velodyne"]
    suffixs = ["png", "txt", "txt", "bin"]
    for idx in idx_train:
        for sub_dir, suffix in zip(sub_dirs, suffixs):
            pi_sub = os.path.join(kitti_dataset_data_root, sub_dir, f"{idx}.{suffix}")
            po_sub = os.path.join(target_dir, sub_dir, f"{idx}.{suffix}")
            shutil.copyfile(pi_sub, po_sub)
    return idx_train


def copy_data4retrain(kitti_aug_dir, kitti_workplace_dir, dir_seq_arr_train, dir_seq_arr_val, ori_train_size):
    kitti_aug_dir_data = os.path.join(kitti_aug_dir, "training")
    kitti_workplace_dir_data = os.path.join(kitti_workplace_dir, "training")
    kitti_workplace_dir_test = os.path.join(kitti_workplace_dir, "testing")

    sub_dirs = ["image_2", "label_2", "calib", "velodyne", "label_2_new"]
    for sub_dir in sub_dirs:
        po_sub = os.path.join(kitti_workplace_dir_data, sub_dir)
        po_test = os.path.join(kitti_workplace_dir_test, sub_dir)
        os.makedirs(po_test, exist_ok=True)
        if os.path.exists(po_sub):
            shutil.rmtree(po_sub)
        pi_sub = os.path.join(kitti_aug_dir_data, sub_dir)
        shutil.copytree(pi_sub, po_sub)

    ori_train_idx = copy_ori_data4retrain(kitti_workplace_dir_data, train_size=ori_train_size)
    dir_seq_arr_train = np.sort(np.concatenate([dir_seq_arr_train, ori_train_idx], axis=0))
    assert len(set(dir_seq_arr_train).intersection(set(dir_seq_arr_val))) == 0

    kitti_imagesets_dir = os.path.join(kitti_workplace_dir, "ImageSets")
    os.makedirs(kitti_imagesets_dir, exist_ok=True)

    kitti_aug_train = os.path.join(kitti_imagesets_dir, "train.txt")
    kitti_aug_val = os.path.join(kitti_imagesets_dir, "val.txt")
    kitti_aug_train_val = os.path.join(kitti_imagesets_dir, "trainval.txt")
    kitti_aug_test = os.path.join(kitti_imagesets_dir, "test.txt")

    with open(kitti_aug_train, "w") as f:
        pre = ""
        for seq in dir_seq_arr_train:
            f.writelines(pre + str(seq))
            pre = "\n"

    with open(kitti_aug_val, "w") as f:
        pre = ""
        for seq in dir_seq_arr_val:
            f.writelines(pre + str(seq))
            pre = "\n"

    with open(kitti_aug_test, "w") as f:
        ...

    with open(kitti_aug_train_val, "w") as f:
        pre = ""
        for seq in np.concatenate([dir_seq_arr_train, dir_seq_arr_val], axis=0):
            f.writelines(pre + str(seq))
            pre = "\n"


if __name__ == '__main__':
    ...