file_sampling.py

import random
import os
import shutil

class_name = "n02111129"
train_path = f"../../dataset/LGDdataset/imagenet/train/{class_name}/"
val_path = f"../../dataset/LGDdataset/imagenet/val/{class_name}/"

file_list = os.listdir(train_path)
train_destination = f"../../dataset/LGDdataset/220327_relabel_dataset/binary_dataset_ng/train/{class_name}/"
valid_destination = f"../../dataset/LGDdataset/220327_relabel_dataset/binary_dataset_ng/val/{class_name}"

num = 262
val_limit = 45

txt = f"../../dataset/LGDdataset/semi_data/binary/bad/{class_name}.txt"
f = open(txt, "r")

txt_list = []
while True:
    line = f.readline().split()
    if not line: 
        break
    txt_list.append(line[0])


if not (os.path.exists(train_destination)):
    os.makedirs(train_destination)

if not (os.path.exists(valid_destination)):
    os.makedirs(valid_destination)

# train
while(len(os.listdir(train_destination))<num):
    for i in range(num):
        sample = f"{txt_list[i]}"
        sample_name_list = sample.split("/")
        name = sample_name_list[7]


        shutil.copy(f"{train_path}/{name}", f"{train_destination}/{name}")
        
        if(i<val_limit):
            shutil.copy(f"{val_path}/{os.listdir(val_path)[i]}", f"{valid_destination}/{name}")

print("done")

file_sampling2

import random
import os
import shutil

type_name = "multi"
class_name_list = [
"n02098105",
"n02095889",
"n02096051",
"n02093859",
"n02097298",
"n02095570",
"n02093647",
"n02093991",
"n02095314",
"n02097130",
"n02097209",
"n02099267",
"n02105251",
"n02090721",
"n02091635",
"n02105505",
"n02099429",
"n02102973",
"n02092002",
"n02090622"]

num_list = [457, 757, 643, 574, 342, 342, 291, 713, 575, 631, 272, 767, 696, 372, 356, 357, 420, 535, 487, 413]
val_list = [46, 76, 64, 57, 34, 34, 29, 71, 57, 63, 27, 77, 70, 37, 36, 36, 42, 54, 49, 41]



for idx in range(len(class_name_list)):
    class_name = class_name_list[idx]
    num = num_list[idx]
    val_limit = val_list[idx]

    train_path = f"../../dataset/LGDdataset/imagenet/train/{class_name}"
    val_path = f"../../dataset/LGDdataset/imagenet/val/{class_name}"

    train_destination = f"../../dataset/LGDdataset/220418_relabel_dataset/{type_name}_hard/train/{class_name}"
    valid_destination = f"../../dataset/LGDdataset/220418_relabel_dataset/{type_name}_hard/val/{class_name}"
    test_destination = f"../../dataset/LGDdataset/220418_relabel_dataset/{type_name}_hard/test/{class_name}"

    txt = f"../../dataset/LGDdataset/semi_data/{type_name}/{class_name}.txt"
    f = open(txt, "r")

    txt_list = []
    while True:
        line = f.readline().split()
        if not line: 
            break
        txt_list.append(line[0])
 
    if not (os.path.exists(train_destination)):
        os.makedirs(train_destination)

    if not (os.path.exists(valid_destination)):
        os.makedirs(valid_destination)

    if not (os.path.exists(test_destination)):
        os.makedirs(test_destination)

    # train
    print(valid_destination)
    while(len(os.listdir(train_destination))<num):
        for i in range(num):
            sample = f"{txt_list[i]}"
            sample_name_list = sample.split("/")
            name = sample_name_list[7]


            shutil.copy(f"{train_path}/{name}", f"{train_destination}/{name}")
  
            # if(i<val_limit):
            #     shutil.copy(f"{val_path}/{os.listdir(val_path)[i]}", f"{valid_destination}/{name}")

    if(len(os.listdir(train_destination))>=num):
        for j in range(val_list[idx]):
            sample = f"{txt_list[num+j]}"
            sample_name_list = sample.split("/")
            name = sample_name_list[7]


            shutil.copy(f"{train_path}/{name}", f"{test_destination}/{name}")

    if(len(os.listdir(test_destination))>=val_limit):
        for j in range(val_limit):
            sample = f"{txt_list[num+j]}"
            sample_name_list = sample.split("/")
            name = sample_name_list[7]


            shutil.copy(f"{train_path}/{name}", f"{valid_destination}/{name}")

    
    if len(os.listdir(train_destination))==num_list[idx]: print(f"{class_name},train_done")
    else: print(f"{class_name},train not done", len(os.listdir(train_destination)), num_list[idx])

    if len(os.listdir(valid_destination))==val_list[idx]: print(f"{class_name},val_done") 
    else: print(f"{class_name},val not done", len(os.listdir(valid_destination)), val_list[idx])

    if len(os.listdir(test_destination))==val_list[idx]: print(f"{class_name},test_done")
    else : print(f"{class_name},test not done", len(os.listdir(test_destination)), val_list[idx])

print("done")

.py