sync_run.py

import os
import time
import subprocess
import threading
from huggingface_hub import HfApi, snapshot_download

TOKEN = os.getenv("HF_TOKEN")
DATASET_ID = os.getenv("DATASET_ID")

# 定义需要持久化的文件夹和文件
# 该项目主要数据在 data，配置可能在根目录的 .json 或 .env
PERSIST_FILES = ["data", "output", "config.json", "settings.json", ".env"]

api = HfApi(token=TOKEN)

def download_data():
    if not DATASET_ID: return
    try:
        print(f"[System] 正在从 Dataset 拉取持久化数据...")
        # 仅下载数据文件，不下载代码文件 (.py, .html, .js)
        snapshot_download(
            repo_id=DATASET_ID,
            repo_type="dataset",
            local_dir=".",
            token=TOKEN,
            allow_patterns=["data/*", "output/*", "*.json", "*.env", "*.db"],
            ignore_patterns=["*.py", "static/*", "templates/*", "logs/*"]
        )
        print("[System] 拉取完成。")
    except Exception as e:
        print(f"[System] 拉取失败: {e}")

def upload_data():
    while True:
        time.sleep(120) # 每 2 分钟备份一次
        if not DATASET_ID: continue
        try:
            # 上传所有数据和配置文件
            api.upload_folder(
                folder_path=".",
                repo_id=DATASET_ID,
                repo_type="dataset",
                allow_patterns=["data/*", "output/*", "*.json", "*.env", "*.db"],
                ignore_patterns=["logs/*", "__pycache__/*", "*.py", "static/*", "templates/*"],
                run_as_future=True
            )
            print(f"[Backup] {time.strftime('%H:%M:%S')} 同步已提交至 Dataset")
        except Exception as e:
            if "No files have been modified" not in str(e):
                print(f"[Backup] 备份出错: {e}")

if __name__ == "__main__":
    os.makedirs("data", exist_ok=True)
    os.makedirs("output", exist_ok=True)

    download_data()
    
    threading.Thread(target=upload_data, daemon=True).start()

    print("[System] 正在启动 Web UI...")
    # 强制端口 7860
    subprocess.run(["python", "webui.py", "--host", "0.0.0.0", "--port", "7860"])