添加了重启时的强制同步拉取功能

start.sh

@@ -1,13 +1,15 @@
 
-
 set -ex
 
-python3 /app/sync.py &
-
-sleep 3
+# 首先执行初始化并等待完成
+python3 /app/sync.py --mode init
 
-ls /data
+# 如果初始化成功，启动同步服务
+python3 /app/sync.py --mode sync &
 
+# 等待几秒确保同步服务正常启动
+sleep 1
 
+# 启动 one-api 服务
+ls /data
 /one-api
-

sync.py

@@ -3,6 +3,9 @@ import time
 import tarfile
 import hashlib
 import shutil
+import argparse
+import sys
+from enum import Enum, auto
 from pathlib import Path
 from typing import Optional
 from dataclasses import dataclass
@@ -11,6 +14,11 @@ import logging
 from dotenv import load_dotenv
 from huggingface_hub import CommitScheduler, HfApi
 
+class SyncMode(Enum):
+    INIT_ONLY = auto()  # 只执行初始化
+    SYNC_ONLY = auto()  # 只执行同步
+    BOTH = auto()       # 执行初始化和同步
+
 @dataclass
 class Config:
     repo_id: str
@@ -123,9 +131,11 @@ class ArchiveManager:
                     path=self.config.data_path,
                     filter=self._tar_filter
                 )
+            return True
         except Exception as e:
-            self.logger.info(f"No existing archive found or download failed: {e}")
+            self.logger.error(f"No existing archive found or download failed: {e}")
             self.config.data_path.mkdir(parents=True, exist_ok=True)
+            return False
 
     @staticmethod
     def _tar_filter(tarinfo, path):
@@ -133,8 +143,7 @@ class ArchiveManager:
         if tarinfo.name.startswith('data/'):
             tarinfo.name = tarinfo.name[5:]
             return tarinfo
-        else:
-            return None
+        return None
 
 class SyncService:
     def __init__(self, config: Config, logger: Logger):
@@ -142,14 +151,29 @@ class SyncService:
         self.logger = logger.logger
         self.monitor = DirectoryMonitor(config.data_path)
         self.archive_manager = ArchiveManager(config, logger)
-        
-    def run(self):
+    
+    def init(self) -> bool:
+        """
+        执行初始化操作
+        返回: 是否成功初始化
+        """
+        try:
+            self.logger.info("Starting initialization...")
+            self.config.sync_path.mkdir(parents=True, exist_ok=True)
+            success = self.archive_manager.extract_archive()
+            if success:
+                self.logger.info("Initialization completed successfully")
+            else:
+                self.logger.warning("Initialization completed with warnings")
+            return success
+        except Exception as e:
+            self.logger.error(f"Initialization failed: {e}")
+            return False
+
+    def sync(self):
+        """执行持续同步操作"""
         self.logger.info(f"Starting sync process for repo: {self.config.repo_id}")
         self.logger.info(f"Sync interval: {self.config.sync_interval} minutes")
-        
-        # 初始化目录和下载数据
-        self.config.sync_path.mkdir(parents=True, exist_ok=True)
-        self.archive_manager.extract_archive()
 
         scheduler = CommitScheduler(
             repo_id=self.config.repo_id,
@@ -175,11 +199,38 @@ class SyncService:
             self.logger.info("Stopping sync process...")
             scheduler.stop()
 
+def parse_args():
+    parser = argparse.ArgumentParser(description='Data synchronization service')
+    parser.add_argument(
+        '--mode',
+        type=str,
+        choices=['init', 'sync', 'both'],
+        default='both',
+        help='Operation mode: init (initialization only), sync (synchronization only), both (default)'
+    )
+    return parser.parse_args()
+
 def main():
+    args = parse_args()
     config = Config.from_env()
     logger = Logger()
     service = SyncService(config, logger)
-    service.run()
+
+    mode = {
+        'init': SyncMode.INIT_ONLY,
+        'sync': SyncMode.SYNC_ONLY,
+        'both': SyncMode.BOTH
+    }[args.mode]
+
+    if mode in (SyncMode.INIT_ONLY, SyncMode.BOTH):
+        success = service.init()
+        if not success:
+            sys.exit(1)
+        if mode == SyncMode.INIT_ONLY:
+            return
+
+    if mode in (SyncMode.SYNC_ONLY, SyncMode.BOTH):
+        service.sync()
 
 if __name__ == "__main__":
     main()