当前位置：首页 > article >正文

SpringAI的使用

article 2025/9/12 16:26:25

1. 项目依赖配置

首先需要在 pom.xml 中添加 SpringAI 相关依赖。以下是关键依赖项：

xml

<!-- SpringAI 核心依赖 -->
<dependency><groupId>org.springframework.ai</groupId><artifactId>spring-ai-core</artifactId><version>0.8.0</version>
</dependency><!-- SpringAI 智谱 AI 支持 -->
<dependency><groupId>org.springframework.ai</groupId><artifactId>spring-ai-zhipuai</artifactId><version>0.8.0</version>
</dependency><!-- SpringAI Redis 向量存储支持 -->
<dependency><groupId>org.springframework.ai</groupId><artifactId>spring-ai-vectorstore-redis</artifactId><version>0.8.0</version>
</dependency><!-- 其他必要依赖 -->
<dependency><groupId>org.springframework.boot</groupId><artifactId>spring-boot-starter-web</artifactId>
</dependency>
<dependency><groupId>org.springframework.boot</groupId><artifactId>spring-boot-starter-data-redis</artifactId>
</dependency>
<dependency><groupId>org.springframework.cloud</groupId><artifactId>spring-cloud-starter-alibaba-nacos-discovery</artifactId>
</dependency>

知识总结：这些依赖提供了 SpringAI 的核心功能、智谱 AI 模型支持、Redis 向量存储以及项目所需的基础组件（如 Web 服务和服务发现）。

2. 应用配置

在 application.yml 中配置 SpringAI 和相关服务：

yaml

server:port: 8081spring:application:name: railway-Ai# SpringAI 配置ai:zhipuai:chat:enabled: trueoptions:model: GLM-4-Air  # 使用的智谱 AI 模型base-url: http://open.bigmodel.cn/api/paasapi-key:   # 智谱 AI API 密钥embedding:enabled: truebase-url: http://open.bigmodel.cn/api/paasapi-key:   # 智谱 AI API 密钥options:model: embedding-2  # 使用的智谱 AI 嵌入模型# Redis 向量存储配置vectorstore:redis:index: my_vector_index  # 向量索引名称prefix: chatbot  # 键前缀initialize-schema: true  # 自动初始化模式# Redis 配置data:redis:host: 121.43.138.23database: 0port: 6379# Nacos 服务发现配置cloud:nacos:discovery:server-addr: 121.43.138.23:8848namespace: publicgroup: DEFAULT_GROUP

知识总结：

spring.ai.zhipuai 部分配置了智谱 AI 的连接信息和使用的模型
spring.ai.vectorstore.redis 配置了 Redis 向量存储的参数
Redis 和 Nacos 是项目运行的基础设施，分别用于向量存储和服务注册发现

3. 核心组件配置

配置聊天内存和文本分割器：

java

// ChatClientConfig.java
import org.springframework.ai.chat.ChatMemory;
import org.springframework.ai.chat.inmemory.InMemoryChatMemory;
import org.springframework.ai.text.splitter.TokenTextSplitter;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;@Configuration
public class ChatClientConfig {@BeanChatMemory chatMemory() {return new InMemoryChatMemory();}@Beanpublic TokenTextSplitter tokenTextSplitter() {return new TokenTextSplitter();}
}

知识总结：

ChatMemory 用于存储对话历史，支持对话上下文感知
TokenTextSplitter 将文本分割为合适的 token 块，便于向量嵌入处理

4. 数据处理与向量存储

实现 PDF 文件内容提取并存储到向量数据库：

java

// PdfStoreServiceImpl.java
import org.springframework.ai.document.Document;
import org.springframework.ai.document.DocumentReader;
import org.springframework.ai.document.PagePdfDocumentReader;
import org.springframework.ai.document.TextSplitter;
import org.springframework.ai.vectorstore.VectorStore;
import org.springframework.core.io.Resource;
import org.springframework.core.io.ResourceLoader;
import org.springframework.stereotype.Service;import java.util.List;
import java.util.function.Function;@Service
public class PdfStoreServiceImpl {private final ResourceLoader resourceLoader;private final VectorStore vectorStore;private final Function<List<Document>, List<Document>> tokenTextSplitter;public PdfStoreServiceImpl(ResourceLoader resourceLoader, VectorStore vectorStore,Function<List<Document>, List<Document>> tokenTextSplitter) {this.resourceLoader = resourceLoader;this.vectorStore = vectorStore;this.tokenTextSplitter = tokenTextSplitter;}// 按页分割并存储 PDF 内容public void saveSourceByPage(String url) {Resource resource = resourceLoader.getResource(url);// 配置 PDF 读取器PdfDocumentReaderConfig loadConfig = PdfDocumentReaderConfig.builder().build();PagePdfDocumentReader pagePdfDocumentReader = new PagePdfDocumentReader(resource, loadConfig);// 分割文档并存储到向量库vectorStore.accept(tokenTextSplitter.apply(pagePdfDocumentReader.get()));}// 其他存储方法...
}

知识总结：

DocumentReader 负责从不同来源读取内容
TextSplitter 将文档分割为可管理的块
VectorStore 将文本转换为向量并存储，支持相似性检索

5. 聊天服务实现

实现基于向量检索和通用聊天的服务：

java

// ChatServiceImpl.java
import org.springframework.ai.chat.ChatClient;
import org.springframework.ai.chat.ChatResponse;
import org.springframework.ai.chat.adapter.zhipuai.ZhiPuAiChatModel;
import org.springframework.ai.chat.memory.MessageChatMemoryAdvisor;
import org.springframework.ai.prompt.Prompt;
import org.springframework.ai.prompt.messages.UserMessage;
import org.springframework.ai.retriever.QuestionAnswerAdvisor;
import org.springframework.ai.vectorstore.VectorStore;
import org.springframework.stereotype.Service;@Service
public class ChatServiceImpl implements ChatService {private static final String PROFESSIONAL_SYSTEM_PROMPT = "You are a professional railway maintenance advisor. " +"Provide accurate, concise, and actionable advice based on the context.";private final ZhiPuAiChatModel chatModel;private final VectorStore vectorStore;private final ToolConfig toolConfig;private final ChatMemory chatMemory;public ChatServiceImpl(ZhiPuAiChatModel chatModel, VectorStore vectorStore,ToolConfig toolConfig,ChatMemory chatMemory) {this.chatModel = chatModel;this.vectorStore = vectorStore;this.toolConfig = toolConfig;this.chatMemory = chatMemory;}@Overridepublic String chatByVectorStore(String message) {// 创建聊天客户端并设置系统提示ChatResponse response = ChatClient.builder(chatModel).defaultSystem(PROFESSIONAL_SYSTEM_PROMPT).build().prompt().tools(toolConfig)  // 注册工具.advisors(new MessageChatMemoryAdvisor(chatMemory),  // 聊天历史顾问new QuestionAnswerAdvisor(vectorStore)  // 向量检索顾问).user(message)  // 用户消息.call();  // 调用模型生成回复return response.getContent();}@Overridepublic String chatByGeneral(String message) {// 通用聊天，不使用向量检索return ChatClient.create(chatModel).prompt().advisors(new MessageChatMemoryAdvisor(chatMemory)).user(message).tools(toolConfig).call().content();}
}

知识总结：

ChatClient 是与 AI 模型交互的核心接口
MessageChatMemoryAdvisor 提供对话历史功能
QuestionAnswerAdvisor 利用向量存储进行相似内容检索
ToolConfig 注册可在对话中调用的工具

6. 控制器实现

提供 REST API 接口：

java

// ChatController.java
import org.springframework.web.bind.annotation.GetMapping;
import org.springframework.web.bind.annotation.RequestParam;
import org.springframework.web.bind.annotation.RestController;@RestController
@RequestMapping("/chat/")
public class ChatController {private final ChatService chatService;public ChatController(ChatService chatService) {this.chatService = chatService;}@GetMapping("think")public String think(@RequestParam("msg") String msg) {return chatService.chatByGeneral(msg);}@GetMapping("advisorThink")public String advisorThink(@RequestParam("msg") String msg) {return chatService.chatByVectorStore(msg);}
}

知识总结：

/chat/think 提供通用聊天功能
/chat/advisorThink 提供基于向量检索的专业问答功能

7. 工具配置

配置可在对话中调用的工具：

java

// ToolConfig.java
import org.springframework.ai.tool.Tool;
import org.springframework.stereotype.Service;import java.time.LocalDateTime;
import java.time.format.DateTimeFormatter;
import java.util.List;
import java.util.Map;@Service
public class ToolConfig {@Autowiredprivate RailwayFeignClient railwayFeignClient;@Tool(name = "拿到当前时间", description = "拿到用户当前的日期时间")public String getCurrentDateTime() {LocalDateTime now = LocalDateTime.now();DateTimeFormatter formatter = DateTimeFormatter.ISO_LOCAL_DATE_TIME;return now.format(formatter);}@Tool(name = "发送邮件", description = "帮助用户发送邮件")public void sendMail(@ToolParam(description = "需要维修的设备的相关信息name、model、station") List<Map<String, String>> maintInfoList,@ToolParam(description = "距离维修日期的天数") Integer day,@ToolParam(description = "要发送的对象的工号") List<String> emailList) {// 实现邮件发送逻辑}@Tool(name = "获取维修信息", description = "拿到需要维修的设备")public String getExpire() {return railwayFeignClient.expire(TokenConstant.getToken()).toString();}
}

知识总结：