feat: 修复数据库持久化，完善服务降级机制

- 恢复使用 AsyncPostgresSaver 持久化短期记忆
- 添加 LLM 作为 Rerank 服务的最后降级方案
- 完善降级链：Local llama.cpp → Zhipu Rerank → LLM Fallback

backend/app/backend.py

@@ -4,8 +4,7 @@ FastAPI 后端 - 支持动态模型切换，使用 PostgreSQL 持久化记忆
 """
 
 import os
-# from app.config import DB_URI, BACKEND_PORT
-from app.config import BACKEND_PORT
+from app.config import DB_URI, BACKEND_PORT
 import uuid
 import json
 from contextlib import asynccontextmanager
@@ -15,7 +14,7 @@ from fastapi import FastAPI, HTTPException, WebSocket, WebSocketDisconnect, Depe
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi.responses import StreamingResponse
 from pydantic import BaseModel
-from langgraph.checkpoint.memory import MemorySaver
+from langgraph.checkpoint.postgres.aio import AsyncPostgresSaver
 from .agent.service import AIAgentService
 from .agent.history import ThreadHistoryService
 from app.core.human_review import (
@@ -27,42 +26,56 @@ from app.core.human_review import (
 from app.subgraphs.contact.api_client import ContactAPIClient
 from app.subgraphs.dictionary.api_client import DictionaryAPIClient
 from app.subgraphs.news_analysis.api_client import NewsAPIClient
-# from .db.init_db import init_subgraph_tables
-# from .db.models import ContactRepository, DictionaryRepository, NewsRepository
+from .db.init_db import init_subgraph_tables
+from .db.models import ContactRepository, DictionaryRepository, NewsRepository
 from app.logger import info, error
 
 @asynccontextmanager
 async def lifespan(app: FastAPI):
-    """应用生命周期管理：创建并注入全局服务（临时用内存 checkpoint）"""
-    # 1. 创建内存 checkpointer（临时测试）
-    checkpointer = MemorySaver()
+    """应用生命周期管理：创建并注入全局服务"""
+    # 1. 创建数据库连接池并初始化表（仅 checkpointer）
+    async with AsyncPostgresSaver.from_conn_string(DB_URI) as checkpointer:
+        await checkpointer.setup()
         
-    # 2. 构建 AI Agent 服务
-    agent_service = AIAgentService(checkpointer)
-    await agent_service.initialize()
+        # 1.5 初始化子图表
+        await init_subgraph_tables(checkpointer.conn)
         
-    # 3. 创建历史查询服务（保持原有的 checkpointer 参数）
-    history_service = ThreadHistoryService(checkpointer)
+        # 2. 构建 AI Agent 服务
+        agent_service = AIAgentService(checkpointer)
+        await agent_service.initialize()
         
-    # 4. 创建审核管理器
-    review_manager = ReviewManager(InMemoryReviewStore())
+        # 3. 创建历史查询服务（保持原有的 checkpointer 参数）
+        history_service = ThreadHistoryService(checkpointer)
         
-    # 5. 将服务实例存入 app.state
-    app.state.agent_service = agent_service
-    app.state.history_service = history_service
-    app.state.review_manager = review_manager
-    app.state.contact_api = ContactAPIClient()
-    app.state.dictionary_api = DictionaryAPIClient()
-    app.state.news_api = NewsAPIClient()
-    app.state.contact_repo = None
-    app.state.dictionary_repo = None
-    app.state.news_repo = None
+        # 3.5 创建子图 Repositories
+        contact_repo = ContactRepository(checkpointer.conn)
+        dictionary_repo = DictionaryRepository(checkpointer.conn)
+        news_repo = NewsRepository(checkpointer.conn)
         
-    # 应用运行中...
-    yield
+        # 3.6 创建子图 API 客户端（真实数据库模式）
+        contact_api = ContactAPIClient(checkpointer.conn)
+        dictionary_api = DictionaryAPIClient(word_repository=dictionary_repo)
+        news_api = NewsAPIClient(news_repository=news_repo)
         
-    # 6. 关闭时清理
-    info("🛑 应用关闭")
+        # 4. 创建审核管理器
+        review_manager = ReviewManager(InMemoryReviewStore())
+        
+        # 5. 将服务实例存入 app.state
+        app.state.agent_service = agent_service
+        app.state.history_service = history_service
+        app.state.review_manager = review_manager
+        app.state.contact_api = contact_api
+        app.state.dictionary_api = dictionary_api
+        app.state.news_api = news_api
+        app.state.contact_repo = contact_repo
+        app.state.dictionary_repo = dictionary_repo
+        app.state.news_repo = news_repo
+        
+        # 应用运行中...
+        yield
+        
+        # 6. 关闭时自动清理数据库连接（async with 负责）
+        info("🛑 应用关闭，数据库连接池已释放")
 
 app = FastAPI(lifespan=lifespan)
 

backend/app/model_services/rerank_services.py

@@ -136,6 +136,92 @@ class ZhipuRerankService(BaseRerankService):
             raise
 
 
+class LLMFallbackRerankService(BaseRerankService):
+    """
+    使用 LLM 作为最后的降级方案进行重排
+    通过让 LLM 评估文档相关性并给出分数
+    """
+
+    def __init__(self, llm=None):
+        from .chat_services import get_chat_service
+        self.llm = llm or get_chat_service()
+
+    def compute_scores(self, query: str, documents: List[str]) -> List[float]:
+        """
+        使用 LLM 评估文档相关性并打分
+        """
+        if not documents:
+            return []
+
+        scores = []
+        for doc in documents:
+            score = self._score_single_document(query, doc)
+            scores.append(score)
+        
+        return scores
+
+    def _score_single_document(self, query: str, document: str) -> float:
+        """
+        让 LLM 为单个文档的相关性打分 (0.0-1.0)
+        """
+        prompt = f"""你是一个文档相关性评分专家。请评估以下文档与查询的相关性，返回一个0到1之间的分数：
+- 1.0表示完全相关
+- 0.0表示完全不相关
+
+查询: {query}
+
+文档: {document}
+
+请只返回一个数字，不要解释。"""
+
+        try:
+            result = self.llm.invoke(prompt)
+            content = result.content if hasattr(result, 'content') else str(result)
+            # 尝试提取数字
+            import re
+            match = re.search(r'(\d+\.?\d*)', content)
+            if match:
+                score = float(match.group(1))
+                # 确保在 0-1 之间
+                return max(0.0, min(1.0, score))
+            # 如果没有找到数字，返回0.5作为默认值
+            return 0.5
+        except Exception as e:
+            logger.warning(f"LLM 打分失败，返回默认分数 0.5: {e}")
+            return 0.5
+
+
+class LLMFallbackRerankProvider(BaseServiceProvider[BaseRerankService]):
+    """
+    LLM 降级重排服务提供者
+    """
+
+    def __init__(self, llm=None):
+        super().__init__("llm_fallback_rerank")
+        self._llm = llm
+
+    def is_available(self) -> bool:
+        """
+        LLM 降级方案总是可用（只要 LLM 服务可用）
+        """
+        try:
+            from .chat_services import get_chat_service
+            get_chat_service()
+            logger.info("LLM 降级重排服务可用")
+            return True
+        except Exception as e:
+            logger.warning(f"LLM 降级重排服务不可用: {e}")
+            return False
+
+    def get_service(self) -> BaseRerankService:
+        """
+        获取 LLM 降级重排服务
+        """
+        if self._service_instance is None:
+            self._service_instance = LLMFallbackRerankService(self._llm)
+        return self._service_instance
+
+
 class LocalLlamaCppRerankProvider(BaseServiceProvider[BaseRerankService]):
     """
     本地 llama.cpp 重排服务提供者
@@ -221,13 +307,15 @@ def get_rerank_service() -> BaseRerankService:
     """
     获取重排服务（带自动降级）- 纯服务层
 
+    降级链: Local llama.cpp -> Zhipu Rerank -> LLM Fallback
+
     Returns:
         BaseRerankService: 重排服务实例
     """
     def _create_chain():
         primary = LocalLlamaCppRerankProvider()
-        fallback = ZhipuRerankProvider()
-        return FallbackServiceChain(primary, [fallback])
+        fallbacks = [ZhipuRerankProvider(), LLMFallbackRerankProvider()]
+        return FallbackServiceChain(primary, fallbacks)
 
     chain = SingletonServiceManager.get_or_create("rerank_service_chain", _create_chain)
     return chain.get_available_service()

scripts/start.sh

@@ -1,7 +1,7 @@
 #!/bin/bash
 # =============================================================================
-# AI Agent 启动与管理脚本
-# 用法: ./backend/scripts/start.sh both [check|backend|frontend|both|docker-up|docker-down]
+# AI Agent 启动与管理脚本 - 简化版
+# 用法: ./scripts/start.sh [check|backend|frontend|both]
 # =============================================================================
 
 set -e
@@ -14,284 +14,18 @@ YELLOW='\033[1;33m'
 NC='\033[0m' # No Color
 
 # 项目根目录
-PROJECT_DIR="/home/huang/Study/AIProject/Agent1"
+PROJECT_DIR="/root/projects/ailine"
 
 echo -e "${BLUE}========================================${NC}"
 echo -e "${BLUE}   AI Agent - 个人生活助手${NC}"
 echo -e "${BLUE}========================================${NC}"
 echo ""
 
-# =============================================================================
-# 配置检查函数
-# =============================================================================
-check_config() {
-    echo -e "${BLUE}📋 开始环境配置检查...${NC}"
-    echo ""
-    
-    # 加载 .env 文件
-    set -a
-    source "$PROJECT_DIR/.env" 2>/dev/null || true
-    set +a
-    
-    PASS=0
-    FAIL=0
-    WARN=0
-    
-    # 辅助函数
-    check_pass() {
-        echo -e "${GREEN}✓${NC} $1"
-        ((PASS++))
-    }
-    
-    check_fail() {
-        echo -e "${RED}✗${NC} $1"
-        ((FAIL++))
-    }
-    
-    check_warn() {
-        echo -e "${YELLOW}⚠${NC} $1"
-        ((WARN++))
-    }
-    
-    # 1. 检查 .env 文件
-    echo "🔍 检查配置文件..."
-    if [ -f "$PROJECT_DIR/.env" ]; then
-        check_pass ".env 文件存在"
-        
-        # 检查文件格式
-        if grep -q "^EOF" .env 2>/dev/null; then
-            check_fail ".env 文件格式错误：发现多余的 EOF 标记"
-        else
-            check_pass ".env 文件格式正确"
-        fi
-    else
-        check_fail ".env 文件不存在"
-        echo "   提示: 请创建 .env 文件并配置环境变量"
-        return 1
-    fi
-    
-    # 2. 检查必需的环境变量
-    echo ""
-    echo "🔑 检查环境变量..."
-    
-    # 检查 ZHIPUAI_API_KEY
-    if grep -q "^ZHIPUAI_API_KEY=" "$PROJECT_DIR/.env" 2>/dev/null; then
-        API_KEY=$(grep "^ZHIPUAI_API_KEY=" "$PROJECT_DIR/.env" | head -1 | cut -d'=' -f2- | tr -d '[:space:]')
-        if [ ${#API_KEY} -gt 10 ]; then
-            check_pass "ZHIPUAI_API_KEY 已配置（长度: ${#API_KEY}）"
-        else
-            check_fail "ZHIPUAI_API_KEY 配置可能无效（过短）"
-        fi
-    else
-        check_fail "ZHIPUAI_API_KEY 未配置或格式错误"
-    fi
-    
-    # 检查 LLAMACPP_API_KEY
-    if grep -q "^LLAMACPP_API_KEY=" "$PROJECT_DIR/.env" 2>/dev/null; then
-        check_pass "LLAMACPP_API_KEY 已配置"
-    else
-        check_warn "LLAMACPP_API_KEY 未配置（如不使用本地模型可忽略）"
-    fi
-    
-    # 检查 DB_URI (远程服务器)
-    if grep -q "^DB_URI=.*115.190.121.151" "$PROJECT_DIR/.env" 2>/dev/null; then
-        check_pass "DB_URI 已配置（远程服务器）"
-    elif grep -q "^DB_URI=" "$PROJECT_DIR/.env" 2>/dev/null; then
-        check_warn "DB_URI 已配置（非远程服务器地址）"
-    else
-        check_fail "DB_URI 未配置"
-    fi
-    
-    # 检查 QDRANT_URL (远程服务器)
-    if grep -q "^QDRANT_URL=.*115.190.121.151" "$PROJECT_DIR/.env" 2>/dev/null; then
-        check_pass "QDRANT_URL 已配置（远程服务器）"
-    elif grep -q "^QDRANT_URL=" "$PROJECT_DIR/.env" 2>/dev/null; then
-        check_warn "QDRANT_URL 已配置（非远程服务器地址）"
-    else
-        check_fail "QDRANT_URL 未配置"
-    fi
-    
-    # 3. 检查 Docker 环境
-    echo ""
-    echo "🐳 检查 Docker 环境..."
-    
-    if command -v docker &> /dev/null; then
-        check_pass "Docker 已安装"
-        
-        if docker info &> /dev/null; then
-            check_pass "Docker 守护进程正在运行"
-        else
-            check_fail "Docker 守护进程未运行"
-            echo "   提示: sudo systemctl start docker"
-        fi
-    else
-        check_fail "Docker 未安装"
-    fi
-    
-    if command -v docker compose version &> /dev/null || command -v docker-compose &> /dev/null; then
-        check_pass "Docker Compose 已安装"
-    else
-        check_fail "Docker Compose 未安装"
-    fi
-    
-    # 4. 检查端口占用
-    echo ""
-    echo "🔌 检查端口占用..."
-    
-    for port in 8081 8082 8083 8501; do
-        if lsof -i :$port &> /dev/null; then
-            check_warn "端口 $port 已被占用"
-        else
-            check_pass "端口 $port 可用"
-        fi
-    done
-    
-    # 5. 检查远程服务连接
-    echo ""
-    echo "🌐 检查远程服务连接..."
-    
-    # 测试 PostgreSQL 连接（从环境变量读取密码）
-    if command -v psql &> /dev/null; then
-        if [ -n "$DB_PASSWORD" ]; then
-            if PGPASSWORD="$DB_PASSWORD" psql -h 115.190.121.151 -U postgres -d langgraph_db -c "SELECT 1;" &> /dev/null; then
-                check_pass "PostgreSQL 远程连接正常 (115.190.121.151:5432)"
-            else
-                check_fail "PostgreSQL 远程连接失败"
-                echo "   提示: 检查网络连接和防火墙设置"
-            fi
-        else
-            check_warn "DB_PASSWORD 未配置，跳过 PostgreSQL 连接测试"
-        fi
-    else
-        check_warn "psql 客户端未安装，跳过 PostgreSQL 连接测试"
-    fi
-    
-    # 测试 Qdrant 连接
-    if curl -s http://115.190.121.151:6333/collections &> /dev/null; then
-        check_pass "Qdrant 远程连接正常 (115.190.121.151:6333)"
-    else
-        check_fail "Qdrant 远程连接失败"
-        echo "   提示: 检查网络连接和防火墙设置"
-    fi
-    
-    # 总结
-    echo ""
-    echo "=========================================="
-    echo "  检查结果汇总"
-    echo "=========================================="
-    echo -e "${GREEN}通过: $PASS${NC}"
-    echo -e "${RED}失败: $FAIL${NC}"
-    echo -e "${YELLOW}警告: $WARN${NC}"
-    echo ""
-    
-    if [ $FAIL -eq 0 ]; then
-        echo -e "${GREEN}✅ 配置检查通过！${NC}"
-        return 0
-    else
-        echo -e "${RED}❌ 发现 $FAIL 个错误，请修复后重试${NC}"
-        return 1
-    fi
-}
-
-# =============================================================================
-# Docker 容器检查函数（仅检查 llama.cpp 服务）
-# =============================================================================
-check_llamacpp() {
-    echo -e "${BLUE}🔍 检查 llama.cpp LLM 容器...${NC}"
-    if ! docker ps --format '{{.Names}}' | grep -q "^gemma4-llamacpp-server$"; then
-        echo -e "${YELLOW}⚠️  llama.cpp LLM 容器未运行${NC}"
-        return 1
-    else
-        echo -e "${GREEN}✓ llama.cpp LLM 容器正在运行 (gemma4-llamacpp-server)${NC}"
-        return 0
-    fi
-}
-
-check_embedding() {
-    echo -e "${BLUE}🔍 检查 llama.cpp Embedding 容器...${NC}"
-    if ! docker ps --format '{{.Names}}' | grep -q "^embedding-server$"; then
-        echo -e "${YELLOW}⚠️  llama.cpp Embedding 容器未运行${NC}"
-        return 1
-    else
-        echo -e "${GREEN}✓ llama.cpp Embedding 容器正在运行 (embedding-server)${NC}"
-        return 0
-    fi
-}
-
-# =============================================================================
-# 启动 Docker 依赖服务（llama.cpp）
-# =============================================================================
-start_llamacpp() {
-    echo -e "${BLUE}🚀 启动 llama.cpp LLM 容器...${NC}"
-    
-    # 检查模型文件
-    if [ ! -f "/home/huang/Study/AIModel/GGUF/Gemma-4-E2B-Uncensored-HauhauCS-Aggressive-Q6_K_P.gguf" ]; then
-        echo -e "${RED}✗ 错误：LLM 模型文件不存在${NC}"
-        exit 1
-    fi
-    
-    if [ ! -f "/home/huang/Study/AIModel/GGUF/mmproj-Gemma-4-E2B-Uncensored-HauhauCS-Aggressive-f16.gguf" ]; then
-        echo -e "${RED}✗ 错误：多模态投影文件不存在${NC}"
-        exit 1
-    fi
-    
-    docker run -d \
-        --name gemma4-llamacpp-server \
-        --restart=unless-stopped \
-        --group-add=video \
-        --device=/dev/kfd \
-        --device=/dev/dri \
-        -v /home/huang/Study/AIModel/GGUF:/models \
-        -p 8081:8080 \
-        ghcr.io/ggml-org/llama.cpp:server-rocm \
-        -m /models/Gemma-4-E2B-Uncensored-HauhauCS-Aggressive-Q6_K_P.gguf \
-        --mmproj /models/mmproj-Gemma-4-E2B-Uncensored-HauhauCS-Aggressive-f16.gguf \
-        --host 0.0.0.0 \
-        --port 8080 \
-        -ngl 99
-    
-    echo -e "${GREEN}✓ llama.cpp LLM 容器已启动 (端口 8081)${NC}"
-    echo -e "${YELLOW}⏳ 等待模型加载（可能需要几分钟）...${NC}"
-    sleep 15
-}
-
-start_embedding() {
-    echo -e "${BLUE}🚀 启动 llama.cpp Embedding 容器...${NC}"
-    
-    # 检查模型文件
-    if [ ! -f "/home/huang/Study/AIModel/GGUF/Qwen3-Embedding-0.6B-Q8_0.gguf" ]; then
-        echo -e "${RED}✗ 错误：Embedding 模型文件不存在${NC}"
-        exit 1
-    fi
-    
-    docker run -d \
-        --name embedding-server \
-        --restart=unless-stopped \
-        --group-add=video \
-        --device=/dev/kfd \
-        --device=/dev/dri \
-        -v /home/huang/Study/AIModel/GGUF:/models \
-        -p 8082:8080 \
-        -e LLAMA_ARG_CTX_SIZE=16384 \
-        -e LLAMA_ARG_N_PARALLEL=1 \
-        -e LLAMA_ARG_BATCH=512 \
-        -e LLAMA_ARG_N_GPU_LAYERS=99 \
-        -e LLAMA_ARG_API_KEY="$LLAMACPP_API_KEY" \
-        ghcr.io/ggml-org/llama.cpp:server-rocm \
-        -m /models/Qwen3-Embedding-0.6B-Q8_0.gguf \
-        --host 0.0.0.0 \
-        --port 8080 \
-        --embeddings
-    
-    echo -e "${GREEN}✓ llama.cpp Embedding 容器已启动 (端口 8082)${NC}"
-    sleep 5
-}
-
 # =============================================================================
 # 启动 Python 服务
 # =============================================================================
 start_backend() {
-    echo -e "\n${BLUE}🚀 启动后端服务 (端口 8079)...${NC}"
+    echo -e "\n${BLUE}🚀 启动后端服务 (端口 10079)...${NC}"
     cd "$PROJECT_DIR"
     
     # 加载 .env 文件中的环境变量
@@ -300,6 +34,7 @@ start_backend() {
     set +a
     
     export PYTHONPATH="$PROJECT_DIR/backend"
+    export BACKEND_PORT=10079
     python -m app.backend &
     BACKEND_PID=$!
     echo -e "${GREEN}✓ 后端服务已启动 (PID: $BACKEND_PID)${NC}"
@@ -307,7 +42,7 @@ start_backend() {
 }
 
 start_frontend() {
-    echo -e "\n${BLUE}🎨 启动前端界面 (端口 8501)...${NC}"
+    echo -e "\n${BLUE}🎨 启动前端界面 (端口 10501)...${NC}"
     cd "$PROJECT_DIR"
     
     # 加载 .env 文件中的环境变量
@@ -316,41 +51,12 @@ start_frontend() {
     set +a
     
     export PYTHONPATH="$PROJECT_DIR/frontend/src"
-    streamlit run frontend/src/frontend_main.py --server.port 8501 --server.address 0.0.0.0 &
+    export API_URL="http://127.0.0.1:10079/chat"
+    streamlit run frontend/src/frontend_main.py --server.port 10501 --server.address 0.0.0.0 &
     FRONTEND_PID=$!
     echo -e "${GREEN}✓ 前端服务已启动 (PID: $FRONTEND_PID)${NC}"
     echo -e "${GREEN}✓ 访问地址:${NC}"
-    echo -e "   本地开发: http://127.0.0.1:8501"
-}
-
-# =============================================================================
-# Docker Compose 管理
-# =============================================================================
-docker_up() {
-    echo -e "${BLUE}🐳 使用 Docker Compose 启动所有服务...${NC}"
-    cd "$PROJECT_DIR/docker"
-    
-    # 检查 .env 文件
-    if [ ! -f "../.env" ]; then
-        echo -e "${RED}✗ 错误：.env 文件不存在${NC}"
-        echo "   请先复制配置文件:"
-        echo "   cp .env.docker .env  # 服务器部署"
-        exit 1
-    fi
-    
-    docker compose up -d --build
-    
-    echo -e "\n${GREEN}✓ Docker Compose 服务已启动${NC}"
-    echo -e "${BLUE}📊 查看服务状态:${NC} docker compose ps"
-    echo -e "${BLUE}📝 查看日志:${NC} docker compose logs -f"
-    echo -e "${BLUE}🌐 访问应用:${NC} http://127.0.0.1:8501"
-}
-
-docker_down() {
-    echo -e "${BLUE}🛑 停止 Docker Compose 服务...${NC}"
-    cd "$PROJECT_DIR/docker"
-    docker compose down
-    echo -e "${GREEN}✓ 服务已停止${NC}"
+    echo -e "   本地开发: http://127.0.0.1:10501"
 }
 
 # =============================================================================
@@ -366,10 +72,6 @@ cleanup() {
         kill $FRONTEND_PID 2>/dev/null || true
         echo -e "${GREEN}✓ 前端服务已停止${NC}"
     fi
-    echo -e "${YELLOW}💡 提示：Docker 容器需要手动停止${NC}"
-    echo -e "   停止 llama.cpp LLM: docker stop gemma4-llamacpp-server"
-    echo -e "   停止 llama.cpp Embedding: docker stop embedding-server"
-    echo -e "   或使用: $0 docker-down"
     exit 0
 }
 
@@ -380,62 +82,36 @@ trap cleanup SIGINT SIGTERM
 # 主逻辑
 # =============================================================================
 case "${1:-help}" in
-    check)
-        check_config
-        ;;
-    
     backend)
-        check_config || exit 1
-        check_llamacpp || start_llamacpp
-        check_embedding || start_embedding
         start_backend
         echo -e "\n${GREEN}后端服务正在运行，按 Ctrl+C 停止${NC}"
         wait $BACKEND_PID
         ;;
     
     frontend)
-        check_config || exit 1
         start_frontend
         echo -e "\n${GREEN}前端服务正在运行，按 Ctrl+C 停止${NC}"
         wait $FRONTEND_PID
         ;;
     
     both)
-        check_config || exit 1
-        check_llamacpp || start_llamacpp
-        check_embedding || start_embedding
         start_backend
         sleep 3
         start_frontend
-        echo -e "\n${GREEN}所有服务正在运行，按 Ctrl+C 停止 Python 服务${NC}"
-        echo -e "${YELLOW}注意：Docker 容器会在后台继续运行${NC}"
+        echo -e "\n${GREEN}所有服务正在运行，按 Ctrl+C 停止${NC}"
         wait
         ;;
     
-    docker-up)
-        check_config || exit 1
-        docker_up
-        ;;
-    
-    docker-down)
-        docker_down
-        ;;
-    
     help|*)
         echo -e "${BLUE}用法:${NC} $0 [command]"
         echo ""
         echo -e "${BLUE}命令:${NC}"
-        echo "  check       检查环境配置"
         echo "  backend     仅启动后端服务"
         echo "  frontend    仅启动前端服务"
         echo "  both        启动前后端服务（默认）"
-        echo "  docker-up   使用 Docker Compose 启动所有服务"
-        echo "  docker-down 停止 Docker Compose 服务"
         echo "  help        显示此帮助信息"
         echo ""
         echo -e "${BLUE}示例:${NC}"
-        echo "  $0 check              # 检查配置"
         echo "  $0 both               # 启动本地开发环境"
-        echo "  $0 docker-up          # 启动 Docker 部署环境"
         ;;
 esac