Update sync_data.sh

sync_data.sh

@@ -1,38 +1,35 @@
 #!/bin/sh
+set -e
 
-# 检查环境变量
+# 必须设置的环境变量
 if [ -z "$HF_TOKEN" ] || [ -z "$DATASET_ID" ]; then
-    echo "未检测到 HF_TOKEN 或 DATASET_ID，备份功能不可用"
+    echo "请设置环境变量 HF_TOKEN 和 DATASET_ID"
     exit 1
 fi
 
-# 激活虚拟环境
-. $HOME/venv/bin/activate
+# 激活 Python 虚拟环境（请根据实际路径修改）
+. "$HOME/venv/bin/activate"
 
-# 生成同步脚本
+STORAGE_PATH="$HOME/app/data"
+FLAG_FILE="$HOME/.hf_backup_first_done"
+
+# 生成 hf_sync.py 脚本
 cat > hf_sync.py << 'EOL'
-# HuggingFace 同步脚本
 from huggingface_hub import HfApi
-import sys
-import os
-import tarfile
-import tempfile
+import sys, os, tarfile, 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):
     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.sort()
     if len(backup_files) >= max_files:
-        files_to_delete = backup_files[:(len(backup_files) - max_files + 1)]
-        for file_to_delete in files_to_delete:
+        for file_to_delete in backup_files[:(len(backup_files) - max_files + 1)]:
             try:
                 api.delete_file(path_in_repo=file_to_delete, repo_id=repo_id, repo_type="dataset")
                 print(f'已删除旧备份: {file_to_delete}')
             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):
     api = HfApi(token=token)
     try:
@@ -47,8 +44,6 @@ def upload_backup(file_path, file_name, token, repo_id):
     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):
     try:
         api = HfApi(token=token)
@@ -72,7 +67,6 @@ def download_latest_backup(token, repo_id, extract_path):
     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)
@@ -81,7 +75,6 @@ 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]
@@ -96,33 +89,43 @@ if __name__ == "__main__":
     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 hf_sync.py download "${HF_TOKEN}" "${DATASET_ID}" "$HOME/app"  
 
-# 同步函数
+upload_backup() {
+    if [ -d "${STORAGE_PATH}" ]; then
+        echo "上传备份数据..."
+        timestamp=$(date +%Y%m%d_%H%M%S)
+        backup_file="backup_${timestamp}.tar.gz"
+        tar -czf "/tmp/${backup_file}" -C "$(dirname "${STORAGE_PATH}")" "$(basename "${STORAGE_PATH}")"
+        python hf_sync.py upload "${HF_TOKEN}" "${DATASET_ID}" "/tmp/${backup_file}" "${backup_file}"
+        rm -f "/tmp/${backup_file}"
+    else
+        echo "数据目录不存在，无法上传备份"
+    fi
+}
+
+download_restore() {
+    echo "从 HuggingFace 下载最新备份并恢复..."
+    python hf_sync.py download "${HF_TOKEN}" "${DATASET_ID}" "$HOME/app"
+}
+
+if [ ! -f "$FLAG_FILE" ]; then
+    echo "首次运行，开始上传备份..."
+    upload_backup
+    touch "$FLAG_FILE"
+    echo "首次备份完成，程序退出。"
+    exit 0
+fi
+
 sync_data() {
     while true; do
         echo "同步进程启动于 $(date)"
-        
-        # 确保数据目录存在（请根据实际路径修改）
-        STORAGE_PATH="$HOME/app/data"
-        if [ -d "${STORAGE_PATH}" ]; then
-            # 创建备份
-            timestamp=$(date +%Y%m%d_%H%M%S)
-            backup_file="backup_${timestamp}.tar.gz"
-            
-            # 压缩目录（使用-C避免包含父路径）
-            tar -czf "/tmp/${backup_file}" -C "$(dirname "${STORAGE_PATH}")" "$(basename "${STORAGE_PATH}")"
-            
-            # 上传到 HuggingFace
-            echo "正在上传备份到 HuggingFace..."
-            python hf_sync.py upload "${HF_TOKEN}" "${DATASET_ID}" "/tmp/${backup_file}" "${backup_file}"
-            
-            # 合并历史提交
+        upload_backup
+        download_restore
+
         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..."
@@ -132,25 +135,17 @@ sync_data() {
             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..."
+                echo "超过7天，合并历史提交..."
                 python 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。"
+                echo "未满7天，跳过合并历史提交。"
             fi
         fi
 
-            # 清理临时文件
-            rm -f "/tmp/${backup_file}"
-        else
-            echo "存储目录 ${STORAGE_PATH} 不存在，等待中..."
-        fi
-        
-        # 同步间隔
         SYNC_INTERVAL=${SYNC_INTERVAL:-7200}
         echo "下次同步将在 ${SYNC_INTERVAL} 秒后进行..."
         sleep $SYNC_INTERVAL
     done
 }
 
-# 启动同步进程
-sync_data &
+sync_data &