Optimize File System Management for Docker Launcher and Resolve class_registry Dependency Issue

[lcolok]

Optimize File System Management for Docker Launcher and Resolve class_registry Dependency Issue

This PR focuses on two improvements for the Docker launcher, without affecting the core logic of MindSearch:

1. Addition of a FileSystemManager class to optimize file system operations in the Docker launcher
2. Inclusion of class_registry installation command in the Dockerfile to resolve dependency issues

Detailed Changes

1. File System Management Optimization for Docker Launcher

   • Added FileSystemManager class with ensure_dir and ensure_file methods
   • These methods ensure necessary directories and files exist in the Docker environment, creating them if they don't
   • Improved code structure of the Docker launcher, enhancing reliability and maintainability of file system operations

2. Docker Image Dependency Resolution

   • Added pip install class_registry command to the Dockerfile
   • Resolved the ImportError: cannot import name 'AutoRegister' from 'class_registry' error occurring in Docker containers
   • Ensured correct importation of required dependencies for the lagent library in the Docker environment

优化Docker启动器的文件系统管理并解决class_registry依赖问题

本次PR主要针对Docker启动器进行了两个方面的改进，不涉及MindSearch的核心逻辑：

1. 新增 FileSystemManager 类以优化Docker启动器的文件系统操作
2. 在 Dockerfile 中添加 class_registry 的安装命令以解决依赖问题

详细改动

1. Docker启动器文件系统管理优化

   • 新增 FileSystemManager 类，提供了 ensure_dir 和 ensure_file 方法
   • 这些方法用于确保Docker环境中必要的目录和文件存在，如果不存在则创建
   • 优化了Docker启动器的代码结构，提高了文件系统操作的可靠性和可维护性

2. Docker镜像依赖问题解决

   • 在 Dockerfile 中新增 pip install class_registry 命令
   • 解决了Docker容器中出现的 ImportError: cannot import name 'AutoRegister' from 'class_registry' 错误
   • 确保在Docker环境中 lagent 库能够正确导入所需的依赖