Dockerfile

FROM gitea/gitea:1.24.6

# 以 root 用户设置权限（Spaces 允许构建时 root）
USER root

# 安装 Python 和 pip
RUN apk update && \
    apk add python3 py3-pip && \
    pip3 install --break-system-packages --upgrade pip

# 安装 Python 包
RUN pip3 install --break-system-packages watchdog huggingface_hub

# 确保数据目录权限，使用默认 git 用户 (UID/GID 1000)
RUN mkdir -p /data/gitea /data/git /data/ssh && \
    chown -R git:git /data && \
    chmod -R 750 /data

# 设置工作目录
WORKDIR /data

# 暴露 Spaces 默认端口（文档用途，实际由 $PORT 控制）
EXPOSE 7860

# Gitea 环境变量：适配 Spaces 端口，禁用 SSH
ENV GITEA__database__DB_TYPE=sqlite3 \
    GITEA__database__PATH=/data/gitea/gitea.db \
    GITEA__server__ROOT_URL=https://deep-sea-02gitea.hf.space/ \
    GITEA__server__HTTP_ADDR=0.0.0.0 \
    GITEA__server__HTTP_PORT=7860 \
    GITEA__server__DISABLE_SSH=true \
    USER=git

# 切换到 git 用户（非 root）
USER git

COPY --chown=git:git <<'EOF' /uploadhf.py
import os
import time
import logging
import threading
import subprocess
from pathlib import Path
from watchdog.observers import Observer
from watchdog.events import FileSystemEventHandler
from huggingface_hub import HfApi

# 配置日志
logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger(__name__)

class DataDirectoryHandler(FileSystemEventHandler):
    """处理 /data 目录文件变化的监控器"""
    
    def __init__(self, repo_id, hf_token, data_directory="/data"):
        self.repo_id = repo_id
        self.hf_token = hf_token
        self.data_directory = data_directory
        self.api = HfApi(token=hf_token)
        self.last_commit_time = 0
        self.commit_delay = 30  # 防抖延迟：30秒内多次变动只提交一次
        logger.info(f"初始化监控器，监控目录: {data_directory}，目标仓库: {repo_id}")
    
    def on_modified(self, event):
        """文件被修改时触发"""
        if not event.is_directory:
            logger.info(f"检测到文件修改: {event.src_path}")
            self.schedule_commit("文件修改")
    
    def on_created(self, event):
        """文件被创建时触发"""
        if not event.is_directory:
            logger.info(f"检测到新文件: {event.src_path}")
            self.schedule_commit("文件创建")
    
    def on_deleted(self, event):
        """文件被删除时触发"""
        if not event.is_directory:
            logger.info(f"检测到文件删除: {event.src_path}")
            self.schedule_commit("文件删除")
    
    def on_moved(self, event):
        """文件被移动或重命名时触发"""
        logger.info(f"检测到文件移动: {event.src_path} -> {event.dest_path}")
        self.schedule_commit("文件移动")
    
    def schedule_commit(self, change_type):
        """安排提交任务，带有防抖机制"""
        current_time = time.time()
        if current_time - self.last_commit_time > self.commit_delay:
            self.last_commit_time = current_time
            # 使用线程执行提交，避免阻塞监控
            thread = threading.Thread(target=self.commit_changes, args=(change_type,))
            thread.daemon = True
            thread.start()
    
    def commit_changes(self, change_type):
        """提交变更到 Hugging Face Hub"""
        try:
            commit_message = f"自动提交: {change_type} - {time.strftime('%Y-%m-%d %H:%M:%S')}"
            
            # 上传整个目录
            self.api.upload_folder(
                folder_path=self.data_directory,
                repo_id=self.repo_id,
                repo_type="dataset",
                commit_message=commit_message,
                ignore_patterns=["*.tmp", "*.log", "*.temp", ".git/*"]  # 忽略临时文件
            )
            
            logger.info(f"✅ 成功提交变更到 Hugging Face Hub: {commit_message}")
            
        except Exception as e:
            logger.error(f"❌ 提交失败: {e}")

def start_directory_monitoring(data_directory="/data", repo_id=None, hf_token=None):
    """启动目录监控服务"""
    # 参数验证
    if not hf_token:
        hf_token = os.getenv('HF_TOKEN')
        if not hf_token:
            raise ValueError("HF_TOKEN 环境变量未设置")
    
    if not repo_id:
        raise ValueError("必须提供 repo_id 参数")
    
    # 检查监控目录是否存在
    if not os.path.exists(data_directory):
        logger.warning(f"监控目录 {data_directory} 不存在，正在创建...")
        os.makedirs(data_directory, exist_ok=True)
    
    # 创建事件处理器和观察器
    event_handler = DataDirectoryHandler(
        repo_id=repo_id,
        hf_token=hf_token,
        data_directory=data_directory
    )
    
    observer = Observer()
    observer.schedule(event_handler, data_directory, recursive=True)
    
    # 启动监控
    observer.start()
    logger.info(f"🎯 目录监控服务已启动: {data_directory}")
    
    return observer

def start_gitea_server(port=7860):
    """启动 Gitea 服务器[1,3](@ref)"""
    def run_gitea():
        try:
            # 使用 subprocess.Popen 启动 Gitea 服务[1,9](@ref)
            gitea_process = subprocess.Popen(
                ['gitea', 'web', '--port', str(port)],
                stdout=subprocess.PIPE,
                stderr=subprocess.PIPE,
                universal_newlines=True
            )
            
            logger.info(f"🚀 Gitea 服务器已启动，端口: {port}")
            logger.info(f"📊 访问地址: http://localhost:{port}")
            
            # 实时输出 Gitea 日志
            while True:
                output = gitea_process.stdout.readline()
                if output == '' and gitea_process.poll() is not None:
                    break
                if output:
                    logger.info(f"Gitea: {output.strip()}")
            
            # 检查进程退出状态
            return_code = gitea_process.poll()
            if return_code != 0:
                error_output = gitea_process.stderr.read()
                logger.error(f"❌ Gitea 服务器异常退出，返回码: {return_code}")
                logger.error(f"错误信息: {error_output}")
            else:
                logger.info("✅ Gitea 服务器正常退出")
                
        except FileNotFoundError:
            logger.error("❌ 未找到 gitea 命令，请确保 Gitea 已正确安装")
        except Exception as e:
            logger.error(f"❌ 启动 Gitea 服务器时发生错误: {e}")
    
    # 在新线程中运行 Gitea[2,6](@ref)
    gitea_thread = threading.Thread(target=run_gitea)
    gitea_thread.daemon = True  # 设置为守护线程，主程序退出时自动结束
    gitea_thread.start()
    
    return gitea_thread

def main():
    """主函数 - 同时启动目录监控和 Gitea 服务器"""
    # 配置参数
    CONFIG = {
        "data_directory": "/data",
        "repo_id": "02engine/02gitea",
        "hf_token": os.getenv('HF_TOKEN'),
        "gitea_port": 7860
    }
    
    # 验证必要参数
    if not CONFIG["hf_token"]:
        logger.error("❌ 请设置 HF_TOKEN 环境变量")
        return
    
    logger.info("🚀 启动集成服务...")
    
    try:
        # 启动 Gitea 服务器[1,3](@ref)
        gitea_thread = start_gitea_server(CONFIG["gitea_port"])
        
        # 启动目录监控服务
        observer = start_directory_monitoring(
            data_directory=CONFIG["data_directory"],
            repo_id=CONFIG["repo_id"],
            hf_token=CONFIG["hf_token"]
        )
        
        logger.info("✅ 所有服务已启动完成！")
        logger.info("调试: HF_TOKEN = %s", CONFIG["hf_token"])
        logger.info("📁 目录监控: /data → Hugging Face Hub")
        logger.info("🌐 Gitea 服务: http://localhost:7860")
        logger.info("🛑 按 Ctrl+C 停止所有服务")
        
        # 主循环，保持程序运行
        try:
            while True:
                time.sleep(1)
        except KeyboardInterrupt:
            logger.info("正在停止所有服务...")
            
    except Exception as e:
        logger.error(f"❌ 启动服务时发生错误: {e}")
    finally:
        # 清理资源
        if 'observer' in locals():
            observer.stop()
            observer.join()
        logger.info("所有服务已停止")

if __name__ == "__main__":
    main()
EOF

# 运行 Gitea，显式指定 Spaces 默认端口
RUN cat /uploadhf.py
CMD ["python3", "/uploadhf.py"]