Python解析压缩包内部文件的后缀名并分类存放

if zipfile.is_zipfile(archive_path):
    with zipfile.ZipFile(archive_path, 'r') as zip_ref:
        for file_info in zip_ref.infolist():
            ext = os.path.splitext(file_info.filename)[1].lower()
            if ext:
                extensions.add(ext[1:])  # 去掉点号

elif rarfile.is_rarfile(archive_path):
    with rarfile.RarFile(archive_path) as rar_ref:
        for file_info in rar_ref.infolist():
            ext = os.path.splitext(file_info.filename)[1].lower()
            if ext:
                extensions.add(ext[1:])

elif py7zr.is_7zfile(archive_path):
    with py7zr.SevenZipFile(archive_path, mode='r') as z7_ref:
        for filename in z7_ref.getnames():
            ext = os.path.splitext(filename)[1].lower()
            if ext:
                extensions.add(ext[1:])

# -*- coding: UTF-8 -*-
'''
@Project ：测试 
@File    ：main.py
@IDE     ：PyCharm 
@Author  ：一晌小贪欢(278865463@qq.com)
@Date    ：2024/12/23 15:32 
'''


import os
import zipfile
import py7zr
import rarfile
import shutil

# 设置源文件夹和目标文件夹
source_folder = '压缩包数据源'
target_folder = '分类文件夹'

# 检查目标文件夹是否存在，不存在则创建
if not os.path.exists(target_folder):
    os.makedirs(target_folder)


def create_folder(folder_name):
    folder_path = os.path.join(target_folder, folder_name)
    if not os.path.exists(folder_path):
        os.makedirs(folder_path)
    return folder_path


def get_archive_extensions(archive_path):
    extensions = set()

    try:
        if zipfile.is_zipfile(archive_path):
            with zipfile.ZipFile(archive_path, 'r') as zip_ref:
                for file_info in zip_ref.infolist():
                    ext = os.path.splitext(file_info.filename)[1].lower()
                    if ext:
                        extensions.add(ext[1:])  # 去掉点号

        elif rarfile.is_rarfile(archive_path):
            with rarfile.RarFile(archive_path) as rar_ref:
                for file_info in rar_ref.infolist():
                    ext = os.path.splitext(file_info.filename)[1].lower()
                    if ext:
                        extensions.add(ext[1:])

        elif py7zr.is_7zfile(archive_path):
            with py7zr.SevenZipFile(archive_path, mode='r') as z7_ref:
                for filename in z7_ref.getnames():
                    ext = os.path.splitext(filename)[1].lower()
                    if ext:
                        extensions.add(ext[1:])
    except Exception as e:
        print(f"读取文件 {archive_path} 时出错: {str(e)}")
        return set()

    return extensions


# 主程序
for filename in os.listdir(source_folder):
    file_path = os.path.join(source_folder, filename)

    if os.path.isfile(file_path):
        # 获取压缩包内所有文件的扩展名
        extensions = get_archive_extensions(file_path)

        if extensions:
            # 如果只有一种文件类型，移动到对应文件夹
            if len(extensions) == 1:
                ext = extensions.pop()
                dest_folder = create_folder(ext)
                dest_path = os.path.join(dest_folder, filename)
                shutil.move(file_path, dest_path)
                print(f"移动 {filename} 到 {ext} 文件夹")
            else:
                # 如果有多种文件类型，移动到 mixed 文件夹
                dest_folder = create_folder('mixed')
                dest_path = os.path.join(dest_folder, filename)
                shutil.move(file_path, dest_path)
                print(f"移动 {filename} 到 mixed 文件夹 (包含文件类型: {', '.join(extensions)})")
        else:
            print(f"跳过文件: {filename} (无法读取或没有有效文件)")