分类目录结构转csv结构

分类网络目录结构转pandas csv结构:

import os
import time

import pandas as pd

if __name__ == '__main__':

    dir_path=r'F:\project\huajie\abnormal_img\dataset\abnormal_data\val/'



    if os.path.basename(os.path.dirname(dir_path))=="train":
        val_type = 1
    elif os.path.basename(os.path.dirname(dir_path))=="val":
        val_type = 0
    else:
        print("val/train dir is error",os.path.dirname(dir_path))
        exit(1123)

    imgs = ['%s/%s' % (i[0], j) for i in os.walk(dir_path) for j in i[-1] if j.endswith(('.jpg', '.jpeg'))]

    print("data_len",len(imgs))

    #多分类用/分开
    columns=["img_path","cls_0/cls_1","val/train"]

    cla_num=10
    ids=[str(i) for i in range(cla_num)]
    # ids = ["0", "1","2","3","4","","6","7"]

    datas=[]
    start=time.time()
    for img_path in imgs:
        label_txt = os.path.basename(os.path.dirname(img_path))
        id = ids.index(label_txt)
        data_row = pd.Series([img_path.replace(dir_path,"/"), id, val_type], index=columns)
        datas.append(data_row)

    gdp4 =pd.concat(datas ,axis=1).T

    gdp4.to_csv(dir_path+"/label.csv",index=False)
    print("time" ,time.time( ) -start,"s")