Python3 实现完全备份、增量备份目录以及文件

import os
from time import strftime
import pickle as p
import tarfile
import hashlib
#运行脚本，每周四进行完全备份，其他时间进行增量备份。

src_dir = '/root/security'                            #定义源目录
dst_dir = '/tmp/mybackup'                             #定义目标目录
md5_file = '/tmp/mybackup/md5.data'                   #采用MD5进行校验
nowdate = strftime('%Y-%m-%d')


def check_md5(fname):                      #定义MD5处理函数，每次处理4096字节更新MD5值
    m = hashlib.md5()
    with open(fname, 'rb') as fobj:
        data = fobj.read(4096)
        if data:
            m.update(data)
    return m.hexdigest()


def full_backup(src_dir, dst_dir, md5_file):
    md5_dict = {}
    fname = 'full_backup' + os.path.basename(src_dir) + '-' + nowdate + '.tar.gz'
    path_file = os.path.join(dst_dir, fname)
    tar = tarfile.open(path_file, 'w:gz')
    tar.add(src_dir)
    tar.close()
    for path, folders, files in os.walk(src_dir):   #采用os的walk遍历目录
        for each_file in files:
            key = os.path.join(path, each_file)
            md5_dict[key] = check_md5(key)
    with open(md5_file, 'wb') as fobj:
        p.dump(md5_dict, fobj)


def incr_backup(src_dir, dst_dir, md5_file):
    fname = 'incr_backup' + os.path.basename(src_dir) + '-' + nowdate + '.tar.gz'
    path_file = os.path.join(dst_dir, fname)
    with open(md5_file, 'rb') as fobj:
        ori_md5_dict = p.load(fobj)
    new_md5_dict = {}
    tar = tarfile.open(path_file, 'w:gz')
    for path, folders, files in os.walk(src_dir):
        for each_file in files:
            key = os.path.join(path, each_file)
            new_md5_dict[key] = check_md5(key)
            if ori_md5_dict.get(key) != new_md5_dict[key]:
                tar.add(key)
    tar.close()


if __name__ == '__main__':
    if not os.path.exists(dst_dir):
        os.mkdir(dst_dir)
    if strftime('%a') == 'Thu':
        full_backup(src_dir, dst_dir, md5_file)
    else:
        incr_backup(src_dir, dst_dir, md5_file)