Update sync_data.sh

sync_data.sh

@@ -18,10 +18,9 @@ import os
 import tarfile
 import tempfile
 
-# 管理备份文件数量，超出最大数量则自动删除最旧的备份
-def manage_backups(api, repo_id, max_files=50):
+def manage_backups(api, repo_id, prefix, max_files=50):
     files = api.list_repo_files(repo_id=repo_id, repo_type="dataset")
-    backup_files = [f for f in files if f.startswith('backup_') and f.endswith('.tar.gz')]
+    backup_files = [f for f in files if f.startswith(prefix) and f.endswith('.tar.gz')]
     backup_files.sort()
     if len(backup_files) >= max_files:
         files_to_delete = backup_files[:(len(backup_files) - max_files + 1)]
@@ -32,8 +31,7 @@ def manage_backups(api, repo_id, max_files=50):
             except Exception as e:
                 print(f'删除 {file_to_delete} 时出错: {str(e)}')
 
-# 上传备份文件到 HuggingFace
-def upload_backup(file_path, file_name, token, repo_id):
+def upload_backup(file_path, file_name, token, repo_id, prefix):
     api = HfApi(token=token)
     try:
         api.upload_file(
@@ -43,19 +41,17 @@ def upload_backup(file_path, file_name, token, repo_id):
             repo_type="dataset"
         )
         print(f"成功上传 {file_name}")
-        manage_backups(api, repo_id)
+        manage_backups(api, repo_id, prefix)
     except Exception as e:
         print(f"上传文件出错: {str(e)}")
 
-
-# 下载最新备份
-def download_latest_backup(token, repo_id, extract_path):
+def download_latest_backup(token, repo_id, prefix, extract_path):
     try:
         api = HfApi(token=token)
         files = api.list_repo_files(repo_id=repo_id, repo_type="dataset")
-        backup_files = [f for f in files if f.startswith('backup_') and f.endswith('.tar.gz')]
+        backup_files = [f for f in files if f.startswith(prefix) and f.endswith('.tar.gz')]
         if not backup_files:
-            print("未找到任何备份文件")
+            print(f"未找到任何备份文件（前缀: {prefix}）")
             return
         latest_backup = sorted(backup_files)[-1]
         with tempfile.TemporaryDirectory() as temp_dir:
@@ -68,17 +64,13 @@ def download_latest_backup(token, repo_id, extract_path):
             if filepath and os.path.exists(filepath):
                 try:
                     with tarfile.open(filepath, 'r:gz') as tar:
-                        tar.extractall("/mcp-proxy-server")  # 或其它有权限的目录
-                        tar.extractall("/tools")  # 如有需要
-                    print(f"已成功恢复备份: {latest_backup}")
-                except PermissionError as e:
-                    print(f"解压备份时权限不足: {str(e)}，请确保容器有写入根目录的权限。")
+                        tar.extractall(extract_path)
+                    print(f"已成功恢复备份: {latest_backup} 到 {extract_path}")
                 except Exception as e:
                     print(f"解压备份时出错: {str(e)}")
     except Exception as e:
         print(f"下载备份出错: {str(e)}")
 
-# 合并历史提交
 def super_squash_history(token, repo_id):
     try:
         api = HfApi(token=token)
@@ -87,7 +79,6 @@ def super_squash_history(token, repo_id):
     except Exception as e:
         print(f"合并历史出错: {str(e)}")
 
-# 主函数
 if __name__ == "__main__":
     action = sys.argv[1]
     token = sys.argv[2]
@@ -95,67 +86,67 @@ if __name__ == "__main__":
     if action == "upload":
         file_path = sys.argv[4]
         file_name = sys.argv[5]
-        upload_backup(file_path, file_name, token, repo_id)
+        prefix = sys.argv[6]
+        upload_backup(file_path, file_name, token, repo_id, prefix)
     elif action == "download":
-        extract_path = sys.argv[4] if len(sys.argv) > 4 else '.'
-        download_latest_backup(token, repo_id, extract_path)
+        prefix = sys.argv[4]
+        extract_path = sys.argv[5]
+        download_latest_backup(token, repo_id, prefix, extract_path)
     elif action == "super_squash":
         super_squash_history(token, repo_id)
 EOL
 
-# 首次启动时从 HuggingFace 下载最新备份（解压到应用目录）
-echo "正在从 HuggingFace 下载最新备份..."
-python /tmp/hf_sync.py download "${HF_TOKEN}" "${DATASET_ID}" "/"  
+# 首次启动时分别还原 config 和 tools
+echo "正在从 HuggingFace 下载 config 备份..."
+python /tmp/hf_sync.py download "${HF_TOKEN}" "${DATASET_ID}" "config_backup_" "/mcp-proxy-server/config"
+echo "正在从 HuggingFace 下载 tools 备份..."
+python /tmp/hf_sync.py download "${HF_TOKEN}" "${DATASET_ID}" "tools_backup_" "/tools"
 
 # 同步函数
 sync_data() {
     while true; do
         echo "同步进程启动于 $(date)"
-        
-        # 确保数据目录存在（请根据实际路径修改）
+
         STORAGE_PATH1="/mcp-proxy-server/config"
         STORAGE_PATH2="/tools"
-        if [ -d "${STORAGE_PATH1}" ] || [ -d "${STORAGE_PATH2}" ]; then
-            # 创建备份
-            timestamp=$(date +%Y%m%d_%H%M%S)
-            backup_file="backup_${timestamp}.tar.gz"
-            
-            # 压缩目录（使用-C避免包含父路径），同时包含/mcp-proxy-server/config和/tools
-            tar -czf "/tmp/${backup_file}" \
-                -C "/mcp-proxy-server" "config" \
-                -C "/" "tools"
-            
-            # 上传到 HuggingFace
-            echo "正在上传备份到 HuggingFace..."
-            python /tmp/hf_sync.py upload "${HF_TOKEN}" "${DATASET_ID}" "/tmp/${backup_file}" "${backup_file}"
-            
-            # 合并历史提交
-            SQUASH_FLAG_FILE="/tmp/last_squash_time"
-            NOW=$(date +%s)
-            SEVEN_DAYS=$((7*24*60*60))
-            if [ ! -f "$SQUASH_FLAG_FILE" ]; then
+        timestamp=$(date +%Y%m%d_%H%M%S)
+        backup_file1="config_backup_${timestamp}.tar.gz"
+        backup_file2="tools_backup_${timestamp}.tar.gz"
+
+        if [ -d "${STORAGE_PATH1}" ]; then
+            tar -czf "/tmp/${backup_file1}" -C "/mcp-proxy-server" "config"
+            echo "正在上传 config 备份到 HuggingFace..."
+            python /tmp/hf_sync.py upload "${HF_TOKEN}" "${DATASET_ID}" "/tmp/${backup_file1}" "${backup_file1}" "config_backup_"
+            rm -f "/tmp/${backup_file1}"
+        fi
+
+        if [ -d "${STORAGE_PATH2}" ]; then
+            tar -czf "/tmp/${backup_file2}" -C "/" "tools"
+            echo "正在上传 tools 备份到 HuggingFace..."
+            python /tmp/hf_sync.py upload "${HF_TOKEN}" "${DATASET_ID}" "/tmp/${backup_file2}" "${backup_file2}" "tools_backup_"
+            rm -f "/tmp/${backup_file2}"
+        fi
+
+        # 合并历史提交
+        SQUASH_FLAG_FILE="/tmp/last_squash_time"
+        NOW=$(date +%s)
+        SEVEN_DAYS=$((7*24*60*60))
+        if [ ! -f "$SQUASH_FLAG_FILE" ]; then
+            echo $NOW > "$SQUASH_FLAG_FILE"
+            echo "首次合并历史提交..."
+            python /tmp/hf_sync.py super_squash "${HF_TOKEN}" "${DATASET_ID}"
+        else
+            LAST=$(cat "$SQUASH_FLAG_FILE")
+            DIFF=$((NOW - LAST))
+            if [ $DIFF -ge $SEVEN_DAYS ]; then
                 echo $NOW > "$SQUASH_FLAG_FILE"
-                echo "首次合并历史提交..."
+                echo "距离上次合并已超过7天，正在合并历史提交..."
                 python /tmp/hf_sync.py super_squash "${HF_TOKEN}" "${DATASET_ID}"
             else
-                LAST=$(cat "$SQUASH_FLAG_FILE")
-                DIFF=$((NOW - LAST))
-                if [ $DIFF -ge $SEVEN_DAYS ]; then
-                    echo $NOW > "$SQUASH_FLAG_FILE"
-                    echo "距离上次合并已超过7天，正在合并历史提交..."
-                    python /tmp/hf_sync.py super_squash "${HF_TOKEN}" "${DATASET_ID}"
-                else
-                    echo "距离上次合并未满7天，本次跳过合并历史提交。"
-                fi
+                echo "距离上次合并未满7天，本次跳过合并历史提交。"
             fi
-
-            # 清理临时文件
-            rm -f "/tmp/${backup_file}"
-        else
-            echo "存储目录 ${STORAGE_PATH1} 或 ${STORAGE_PATH2} 不存在，等待中..."
         fi
-        
-        # 同步间隔
+
         SYNC_INTERVAL=${SYNC_INTERVAL:-7200}
         echo "下次同步将在 ${SYNC_INTERVAL} 秒后进行..."
         sleep $SYNC_INTERVAL