自动清理Python缓存的脚本

本文目录导读：

1. 基础版本 - 清理Python缓存文件
2. 增强版本 - 带更多功能和配置
3. 配置文件示例 (config.json)
4. 使用方法

我来为你提供一个自动清理Python缓存的脚本。

基础版本 - 清理Python缓存文件

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
自动清理Python缓存的脚本
"""
import os
import shutil
import sys
from pathlib import Path
def clean_python_cache(start_path="."):
    """
    清理Python缓存文件
    包括：__pycache__, .pyc, .pyo, .pyd, .egg-info, .egg
    """
    deleted_count = 0
    freed_space = 0
    # 要删除的目录模式
    dir_patterns = [
        "__pycache__",
        "*.egg-info",
        "*.egg",
        ".mypy_cache",
        ".pytest_cache",
        ".hypothesis"
    ]
    # 要删除的文件模式
    file_patterns = [
        "*.pyc",
        "*.pyo",
        "*.pyd",
        "*.pyc.bak"
    ]
    start_path = Path(start_path).resolve()
    print(f"开始清理: {start_path}")
    print("-" * 50)
    for root, dirs, files in os.walk(start_path):
        # 清理目录
        for dir_name in dirs:
            dir_path = os.path.join(root, dir_name)
            for pattern in dir_patterns:
                if pattern == dir_name or (pattern.startswith("*") and dir_name.endswith(pattern[1:])):
                    try:
                        size = get_folder_size(dir_path)
                        shutil.rmtree(dir_path)
                        deleted_count += 1
                        freed_space += size
                        print(f"删除目录: {dir_path} ({format_size(size)})")
                    except Exception as e:
                        print(f"删除目录失败 {dir_path}: {e}")
        # 清理文件
        for file_name in files:
            file_path = os.path.join(root, file_name)
            for pattern in file_patterns:
                if file_name.endswith(pattern[1:]):
                    try:
                        size = os.path.getsize(file_path)
                        os.remove(file_path)
                        deleted_count += 1
                        freed_space += size
                        print(f"删除文件: {file_path} ({format_size(size)})")
                    except Exception as e:
                        print(f"删除文件失败 {file_path}: {e}")
    print("-" * 50)
    print(f"清理完成!")
    print(f"删除项目数: {deleted_count}")
    print(f"释放空间: {format_size(freed_space)}")
def get_folder_size(folder_path):
    """获取文件夹大小"""
    total_size = 0
    for dirpath, dirnames, filenames in os.walk(folder_path):
        for f in filenames:
            fp = os.path.join(dirpath, f)
            try:
                total_size += os.path.getsize(fp)
            except:
                pass
    return total_size
def format_size(size_bytes):
    """格式化文件大小"""
    if size_bytes == 0:
        return "0 B"
    size_names = ["B", "KB", "MB", "GB"]
    i = 0
    while size_bytes >= 1024 and i < len(size_names) - 1:
        size_bytes /= 1024.0
        i += 1
    return f"{size_bytes:.2f} {size_names[i]}"
if __name__ == "__main__":
    # 如果提供命令行参数，使用指定路径
    if len(sys.argv) > 1:
        target_path = sys.argv[1]
    else:
        target_path = "."
    clean_python_cache(target_path)

增强版本 - 带更多功能和配置

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
高级Python缓存清理工具
"""
import os
import shutil
import sys
import argparse
import json
from pathlib import Path
from datetime import datetime
class CacheCleaner:
    """缓存清理器"""
    def __init__(self, config_file=None):
        self.stats = {
            'directories_deleted': 0,
            'files_deleted': 0,
            'space_freed': 0,
            'errors': 0
        }
        # 默认清理模式
        self.dir_patterns = [
            "__pycache__",
            "*.egg-info",
            "*.egg",
            ".mypy_cache",
            ".pytest_cache",
            ".hypothesis",
            ".tox",
            "build",
            "dist",
            "*.pyc"  # 当作为目录处理
        ]
        self.file_patterns = [
            "*.pyc",
            "*.pyo",
            "*.pyd",
            "*.pyc.bak",
            "*.so",  # Cython编译文件
            "*.pdb"  # 调试文件
        ]
        # 加载配置文件
        if config_file and os.path.exists(config_file):
            self.load_config(config_file)
        # 排除目录
        self.exclude_dirs = ['.git', '.svn', 'venv', 'env', '.env',
                            'node_modules', '.idea', '.vscode']
    def load_config(self, config_file):
        """从JSON文件加载配置"""
        try:
            with open(config_file, 'r') as f:
                config = json.load(f)
                if 'dir_patterns' in config:
                    self.dir_patterns = config['dir_patterns']
                if 'file_patterns' in config:
                    self.file_patterns = config['file_patterns']
                if 'exclude_dirs' in config:
                    self.exclude_dirs = config['exclude_dirs']
        except Exception as e:
            print(f"加载配置文件失败: {e}")
    def clean(self, path=".", verbose=True, dry_run=False):
        """执行清理"""
        start_path = Path(path).resolve()
        if verbose:
            print(f"开始清理: {start_path}")
            print(f"时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
            print("-" * 60)
        for root, dirs, files in os.walk(start_path):
            # 跳过排除目录
            dirs[:] = [d for d in dirs if d not in self.exclude_dirs 
                      and not d.startswith('.')]
            # 清理目录
            for dir_name in dirs[:]:  # 使用切片复制，避免修改遍历列表
                dir_path = os.path.join(root, dir_name)
                for pattern in self.dir_patterns:
                    if self.match_pattern(dir_name, pattern):
                        if dry_run:
                            if verbose:
                                print(f"[模拟] 将删除目录: {dir_path}")
                        else:
                            self.delete_directory(dir_path, verbose)
                        break
            # 清理文件
            for file_name in files:
                file_path = os.path.join(root, file_name)
                for pattern in self.file_patterns:
                    if self.match_pattern(file_name, pattern):
                        if dry_run:
                            if verbose:
                                print(f"[模拟] 将删除文件: {file_path}")
                        else:
                            self.delete_file(file_path, verbose)
                        break
        if verbose:
            self.print_summary()
    def match_pattern(self, name, pattern):
        """检查名称是否匹配模式"""
        if pattern == name:
            return True
        if pattern.startswith("*") and name.endswith(pattern[1:]):
            return True
        return False
    def delete_directory(self, dir_path, verbose=True):
        """删除目录"""
        try:
            size = self.get_folder_size(dir_path)
            shutil.rmtree(dir_path)
            self.stats['directories_deleted'] += 1
            self.stats['space_freed'] += size
            if verbose:
                print(f"删除目录: {dir_path} ({self.format_size(size)})")
        except Exception as e:
            self.stats['errors'] += 1
            if verbose:
                print(f"删除目录失败 {dir_path}: {e}")
    def delete_file(self, file_path, verbose=True):
        """删除文件"""
        try:
            size = os.path.getsize(file_path)
            os.remove(file_path)
            self.stats['files_deleted'] += 1
            self.stats['space_freed'] += size
            if verbose:
                print(f"删除文件: {file_path} ({self.format_size(size)})")
        except Exception as e:
            self.stats['errors'] += 1
            if verbose:
                print(f"删除文件失败 {file_path}: {e}")
    def get_folder_size(self, folder_path):
        """获取文件夹大小"""
        total_size = 0
        for dirpath, dirnames, filenames in os.walk(folder_path):
            for f in filenames:
                fp = os.path.join(dirpath, f)
                try:
                    total_size += os.path.getsize(fp)
                except:
                    pass
        return total_size
    def format_size(self, size_bytes):
        """格式化文件大小"""
        if size_bytes == 0:
            return "0 B"
        size_names = ["B", "KB", "MB", "GB"]
        i = 0
        while size_bytes >= 1024 and i < len(size_names) - 1:
            size_bytes /= 1024.0
            i += 1
        return f"{size_bytes:.2f} {size_names[i]}"
    def print_summary(self):
        """打印清理摘要"""
        print("-" * 60)
        print("清理完成!")
        print(f"删除目录数: {self.stats['directories_deleted']}")
        print(f"删除文件数: {self.stats['files_deleted']}")
        print(f"总释放空间: {self.format_size(self.stats['space_freed'])}")
        if self.stats['errors'] > 0:
            print(f"错误数: {self.stats['errors']}")
        print("-" * 60)
def main():
    parser = argparse.ArgumentParser(
        description='自动清理Python缓存文件',
        formatter_class=argparse.RawDescriptionHelpFormatter,
        epilog="""
示例:
  %(prog)s                          # 清理当前目录
  %(prog)s /path/to/project         # 清理指定项目
  %(prog)s --dry-run                # 模拟运行，不实际删除
  %(prog)s --config my_config.json  # 使用配置文件
  %(prog)s --quiet                  # 静默模式
        """
    )
    parser.add_argument('path', nargs='?', default='.',
                       help='要清理的路径 (默认: 当前目录)')
    parser.add_argument('--dry-run', '-n', action='store_true',
                       help='模拟运行，不实际删除')
    parser.add_argument('--quiet', '-q', action='store_true',
                       help='静默模式，只显示摘要')
    parser.add_argument('--config', '-c', 
                       help='配置文件路径 (JSON格式)')
    parser.add_argument('--version', '-v', action='version',
                       version='Python Cache Cleaner 1.0')
    args = parser.parse_args()
    cleaner = CacheCleaner(args.config)
    cleaner.clean(args.path, verbose=not args.quiet, dry_run=args.dry_run)
if __name__ == "__main__":
    main()

配置文件示例 (config.json)

{
    "dir_patterns": [
        "__pycache__",
        "*.egg-info",
        "*.egg",
        ".mypy_cache",
        ".pytest_cache",
        ".hypothesis",
        ".tox",
        "build",
        "dist"
    ],
    "file_patterns": [
        "*.pyc",
        "*.pyo",
        "*.pyd",
        "*.pyc.bak",
        "*.so",
        "*.pdb"
    ],
    "exclude_dirs": [
        ".git",
        ".svn",
        "venv",
        "env",
        ".env",
        "node_modules",
        ".idea",
        ".vscode"
    ]
}

使用方法

1. 保存脚本为 clean_python_cache.py

2. 基本使用：

   # 清理当前目录
   python clean_python_cache.py
   # 清理指定目录
   python clean_python_cache.py /path/to/your/project

3. 高级功能：

   # 模拟运行（查看会被删除的文件）
   python clean_python_cache.py --dry-run
   # 静默模式
   python clean_python_cache.py --quiet
   # 使用配置文件
   python clean_python_cache.py --config config.json

这个脚本可以帮助你：

• 清理 __pycache__ 目录
• 删除 .pyc、.pyo、.pyd 等编译文件
• 清理测试缓存（pytest、mypy等）
• 清理构建目录
• 支持模拟运行，避免误删
• 可配置文件模式

使用时请谨慎，建议先使用 --dry-run 选项查看将要删除的文件。