使用Python轻松构建一个Windows11系统智能垃圾清理系统

{
  "settings": {
    "dry_run": false,
    "days_old": 7,
    "file_extensions": [".tmp", ".log", ".cache", ".old", ".etl"],
    "max_file_size_mb": 500
  },
  "targets": [
    {
      "name": "User Temp",
      "path": "%TEMP%",
      "recursive": true
    },
    {
      "name": "System Temp",
      "path": "C:\\Windows\\Temp",
      "recursive": true
    },
    {
      "name": "Prefetch",
      "path": "C:\\Windows\\Prefetch",
      "recursive": false
    }
  ],
  "exclusions": [
    "desktop.ini",
    "thumbs.db"
  ]
}

import os
import sys
import json
import time
import logging
import shutil
import argparse
from datetime import datetime, timedelta
from pathlib import Path

# 初始化日志
logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(levelname)s - %(message)s',
    handlers=[
        logging.FileHandler('cleaner.log'),
        logging.StreamHandler(sys.stdout)
    ]
)

class SmartCleaner:
    def __init__(self, config_path):
        self.config = self.load_config(config_path)
        self.stats = {'deleted_files': 0, 'freed_space': 0}
        self.now = time.time()
        
        # 计算时间阈值 (秒)
        self.days_threshold = self.config['settings'].get('days_old', 7)
        self.time_threshold = self.now - (self.days_threshold * 86400)

    def load_config(self, path):
        try:
            with open(path, 'r', encoding='utf-8') as f:
                return json.load(f)
        except FileNotFoundError:
            logging.error(f"配置文件 {path} 未找到")
            sys.exit(1)

    def get_file_size(self, filepath):
        return os.path.getsize(filepath)

    def is_excluded(self, filename):
        """检查文件是否在排除列表中"""
        exclusions = self.config.get('exclusions', [])
        return filename.lower() in [e.lower() for e in exclusions]

    def should_delete(self, filepath):
        """核心判断逻辑：是否满足删除条件"""
        try:
            # 1. 检查文件名白名单
            if self.is_excluded(os.path.basename(filepath)):
                return False
            
            # 2. 检查后缀 (如果配置了特定后缀)
            exts = self.config['settings'].get('file_extensions', [])
            if exts:
                if os.path.splitext(filepath)[1].lower() not in exts:
                    return False

            # 3. 检查访问时间 (是否为旧文件)
            # stat.st_atime 是访问时间，st_mtime 是修改时间。这里用 mtime 更稳妥。
            file_mtime = os.path.getmtime(filepath)
            if file_mtime > self.time_threshold:
                return False
            
            # 4. 检查文件大小 (防止误删正在写入的大文件)
            max_size = self.config['settings'].get('max_file_size_mb', 500) * 1024 * 1024
            if os.path.getsize(filepath) > max_size:
                logging.warning(f"文件过大跳过: {filepath}")
                return False

            return True
        except Exception as e:
            logging.warning(f"检查文件 {filepath} 时出错: {e}")
            return False

    def clean_path(self, target_info):
        """清理指定目录"""
        raw_path = target_info['path']
        # Windows环境变量展开
        path = os.path.expandvars(raw_path)
        recursive = target_info.get('recursive', True)

        if not os.path.exists(path):
            logging.warning(f"目录不存在，跳过: {path}")
            return

        logging.info(f"正在扫描: {path} (递归: {recursive})")
        
        if recursive:
            walker = os.walk(path, topdown=False)
        else:
            # 非递归模式下，只处理当前目录文件
            walker = [(path, [], os.listdir(path))]

        for root, dirs, files in walker:
            for file in files:
                file_path = os.path.join(root, file)
                
                if self.should_delete(file_path):
                    self.perform_delete(file_path)

    def perform_delete(self, filepath):
        """执行删除动作"""
        try:
            size = self.get_file_size(filepath)
            
            if self.config['settings']['dry_run']:
                logging.info(f"[模拟] 将删除: {filepath} ({size/1024/1024:.2f} MB)")
            else:
                # 先尝试删除文件，如果文件被占用会抛出异常
                os.remove(filepath)
                logging.info(f"[已删除] {filepath}")
                
            self.stats['deleted_files'] += 1
            self.stats['freed_space'] += size
            
        except PermissionError:
            logging.debug(f"权限不足/文件被占用，跳过: {filepath}")
        except Exception as e:
            logging.error(f"删除 {filepath} 失败: {e}")

    def clean_empty_dirs(self, target_info):
        """可选：清理空文件夹"""
        path = os.path.expandvars(target_info['path'])
        if not target_info.get('recursive', True): return
        
        logging.info(f"正在清理空文件夹: {path}")
        for root, dirs, files in os.walk(path, topdown=False):
            for dir_name in dirs:
                dir_path = os.path.join(root, dir_name)
                try:
                    if not os.listdir(dir_path): # 目录为空
                        if not self.config['settings']['dry_run']:
                            os.rmdir(dir_path)
                            logging.info(f"[已删除空目录] {dir_path}")
                        else:
                            logging.info(f"[模拟] 将删除空目录: {dir_path}")
                except OSError:
                    pass

    def run(self):
        start_time = datetime.now()
        logging.info(f"===== 任务开始 (模拟模式: {self.config['settings']['dry_run']}) =====")

        for target in self.config['targets']:
            self.clean_path(target)
            # self.clean_empty_dirs(target) # 如需删除空文件夹可取消注释

        end_time = datetime.now()
        duration = (end_time - start_time).total_seconds()
        
        logging.info("===== 任务完成 =====")
        logging.info(f"扫描文件数: {self.stats['deleted_files']}")
        logging.info(f"释放空间: {self.stats['freed_space'] / 1024 / 1024:.2f} MB")
        logging.info(f"耗时: {duration:.2f} 秒")

if __name__ == "__main__":
    parser = argparse.ArgumentParser(description="Windows 智能清理脚本")
    parser.add_argument('--dry-run', action='store_true', help='模拟运行，不实际删除文件')
    args = parser.parse_args()

    # 如果命令行传参，覆盖配置文件
    config_file = 'config.json'
    cleaner = SmartCleaner(config_file)
    
    if args.dry_run:
        cleaner.config['settings']['dry_run'] = True
        
    cleaner.run()