数据分类处理---python

说明：读取数据的名字，然后根据数据的名字进行分类，还涉及到图像的名字在csv文件中的查找，用到的有glob，os，会有就是处理数据的pandas工具。

import glob
import os
import pandas as pd
import shutil
import os.path as op


#不动原来的文件，避免数据混乱
#只需要在原来的文件中创建三个文件
#失败的
#封面
#不是封面
#分别放进去就可以了


def main():
    file_path = 'mnt'
    image_paths = glob.glob(os.path.join(file_path, '*/*/*/*/*/*/*/*/*.jpg'))
    image_file = glob.glob(os.path.join(file_path, '*/*/*/*/*/*/*/*'))
    #generate_file(image_file)

    '''对应的文件夹下生成书，内页和没有书的文件夹，如果生成了，那么此句不运行'''

    match = pd.read_csv('templates.csv', names=['pathid', 'imageid', 'bookid', 'group', 'isdel'])
    match_df = pd.DataFrame(match)
    split_file(image_paths, match_df)

    '''
    读取对应的csv文件，然后转化到dataframe下，接着进行图像分类
    
    图像分类：
    success ： imageid :(group:1 封面，group：other numbers 内页)
    fail： without_imageid :(-1:检测不是书的， 不是-1：什么都不是，系统问题的数据）
    
    '''

def generate_file(image_paths):
    names = ['cover','inside_pages','nobooks']
    for image_file in image_paths:
        for n in names:
            generate_path = os.path.join(image_file,n)
            os.mkdir(generate_path)


def split_file(paths,match):
    names = ['cover', 'inside_pages', 'nobooks']
    for image_name in paths:
        imageid_i = image_name.split('_')[3]
        files = image_name.split('/')

        if imageid_i == '-1' and files[-2] != names[2]: #检测不是书的
            try:
                new_path = os.path.join(image_name.replace(op.basename(image_name),names[2])) #不用strip容易出bug
                shutil.move(image_name,new_path)
            except FileNotFoundError as e:
                print(image_name)
        else:
            location = match[(match.imageid == imageid_i)].index.values
            try:
                if match.at[int(location), 'group'] == 1 and files[-2] != names[0]: #检测出来封面的
                    new_path = os.path.join(image_name.replace(op.basename(image_name), names[0]))
                    shutil.move(image_name, new_path)
                    #print(new_path)
                    #print(image_name)
                elif match.at[int(location), 'group'] != 1 and files[-2] != names[1]: #大部分是内页的
                    new_path = os.path.join(image_name.replace(op.basename(image_name), names[1]))
                    #print(new_path)
                    #print(image_name)
                    shutil.move(image_name, new_path)
            except TypeError as e:
                pass
                #print(e)
                #print(image_name)

if __name__ == '__main__':
    main()