所有的大模型本身是不进行信息存储的,也不提供连续对话功能,所以想要实现连续对话功能需要开发者自己写代码才能实现。那怎么才能实现大模型的连续对话功能呢?
大模型连续对话功能不同的框架实现也是不同的,以行业使用最多的 Java AI 框架 Spring AI 和 Spring AI Alibaba 为例,给大家演示一下它们连续对话是如何实现的。
1.SpringAI连续对话实现
Spring AI 以 MySQL 数据库为例,我们来实现一下它的连续对话功能。
PS:我们只有先讲对话存储起来,才能实现连续对话功能,所以我们需要借助数据库存储来连续对话。
1.1 准备工作
1.创建表
CREATE TABLE chat_message ( id BIGINT AUTO_INCREMENT PRIMARY KEY, conversation_id VARCHAR(255) NOT NULL, role VARCHAR(50) NOT NULL, context TEXT NOT NULL, created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP ) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci;
2.添加数据库和 MyBatisPlus 依赖:
<dependency> <groupId>com.baomidou</groupId> <artifactId>mybatis-plus-spring-boot3-starter</artifactId> <version>3.5.11</version> </dependency> <dependency> <groupId>com.mysql</groupId> <artifactId>mysql-connector-j</artifactId> <scope>runtime</scope> </dependency>
3.设置配置文件:
spring: datasource: url: jdbc:mysql://127.0.0.1:3306/testdb?characterEncoding=utf8 username: root password: 12345678 driver-class-name: com.mysql.cj.jdbc.Driver # 配置打印 MyBatis 执行的 SQL mybatis-plus: configuration: log-impl: org.apache.ibatis.logging.stdout.StdOutImpl # 配置打印 MyBatis 执行的 SQL logging: level: com: ai: deepseek: debug
4.编写实体类
import com.baomidou.mybatisplus.annotation.IdType; import com.baomidou.mybatisplus.annotation.TableId; import com.baomidou.mybatisplus.annotation.TableName; import lombok.Getter; import lombok.Setter; import java.io.Serializable; import java.util.Date; @Getter @Setter @TableName("chat_message") public class ChatMessageDO implements Serializable { private static final long serialVersionUID = 1L; @TableId(value = "id", type = IdType.AUTO) private Long id; private String conversationId; private String role; private String context; private Date createdAt; }
5.编写 Mapper:
import com.ai.chat.entity.ChatMessageDO; import com.baomidou.mybatisplus.core.mapper.BaseMapper; import org.apache.ibatis.annotations.Mapper; @Mapper public interface ChatMessageMapper extends BaseMapper<ChatMessageDO> { }
1.2 自定义ChatMemory类
自定义的 ChatMemory 实现类,将对话记录存储到 MySQL:
import com.ai.deepseek.entity.ChatMessageDO; import com.ai.deepseek.mapper.ChatMessageMapper; import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper; import org.springframework.ai.chat.memory.ChatMemory; import org.springframework.ai.chat.messages.Message; import org.springframework.ai.chat.messages.UserMessage; import org.springframework.beans.factory.annotation.Autowired; import org.springframework.stereotype.Component; import java.util.List; import java.util.stream.Collectors; @Component public class MySQLChatMemory implements ChatMemory { @Autowired private ChatMessageMapper repository; @Override public void add(String conversationId, Message message) { ChatMessageDO entity = new ChatMessageDO(); entity.setConversationId(conversationId); entity.setRole(message.getMessageType().name()); entity.setContext(message.getText()); repository.insert(entity); } @Override public void add(String conversationId, List<Message> messages) { messages.forEach(message -> add(conversationId, message)); } @Override public List<Message> get(String conversationId, int lastN) { LambdaQueryWrapper<ChatMessageDO> queryWrapper = new LambdaQueryWrapper<>(); queryWrapper.eq(ChatMessageDO::getConversationId, conversationId); // queryWrapper.orderByDesc(ChatMessageDO::getId); return repository.selectList(queryWrapper) .stream() .limit(lastN) .map(e -> new UserMessage(e.getContext())) .collect(Collectors.toList()); } @Override public void clear(String conversationId) { LambdaQueryWrapper<ChatMessageDO> queryWrapper = new LambdaQueryWrapper<>(); queryWrapper.eq(ChatMessageDO::getConversationId, conversationId); repository.delete(queryWrapper); } }
1.3 代码调用
编写代码测试历史对话保存到 MySQL 的功能:
import com.ai.deepseek.component.MySQLChatMemory; import org.springframework.ai.chat.client.ChatClient; import org.springframework.ai.chat.client.advisor.MessageChatMemoryAdvisor; import org.springframework.beans.factory.annotation.Autowired; import org.springframework.web.bind.annotation.RequestMapping; import org.springframework.web.bind.annotation.RequestParam; import org.springframework.web.bind.annotation.RestController; import reactor.core.publisher.Flux; @RestController @RequestMapping("/multi") public class MultiChatController { @Autowired private ChatClient chatClient; @Autowired private MySQLChatMemory chatMemory; @RequestMapping("/chat") public Flux<String> chat(@RequestParam("msg") String msg, @RequestParam(defaultValue = "default") String sessionId) { // 添加MessageChatMemoryAdvisor,自动管理上下文 MessageChatMemoryAdvisor advisor = new MessageChatMemoryAdvisor(chatMemory, sessionId, 10); // 保留最近5条历史 return chatClient.prompt() .user(msg) .advisors(advisor) // 关键:注入记忆管理 .stream() .content(); } }
以上程序执行结果如下:

2.SpringAIAlibaba实现连续对话
Spring AI Alibaba 连续对话的实现就简单很多了,因为它内置了 MySQL 和 Redis 的连续对话存储方式,接下来以 Redis 为例演示 SAA 的连续对话实现,它的实现步骤如下:
- 添加依赖。
- 设置配置文件,配置 Redis 连接信息。
- 添加 Redis 配置类,注入 RedisChatMemoryRepository 对象。
- 配置 ChatClient 实现连续对话。
具体实现如下。
2.1 添加依赖
<dependency> <groupId>com.alibaba.cloud.ai</groupId> <artifactId>spring-ai-alibaba-starter-memory-redis</artifactId> </dependency>
2.2 设置配置文件
设置配置文件,配置 Redis 连接信息:
spring: ai: memory: redis: host: localhost port: 6379 timeout: 5000
2.3 添加Redis配置类
添加 Redis 配置类,注入 RedisChatMemoryRepository 对象,实现 Redis 自定义存储器注入:
import com.alibaba.cloud.ai.memory.redis.RedisChatMemoryRepository; import org.springframework.beans.factory.annotation.Value; import org.springframework.context.annotation.Bean; import org.springframework.context.annotation.Configuration; @Configuration public class RedisMemoryConfig { @Value("${spring.ai.memory.redis.host}") private String redisHost; @Value("${spring.ai.memory.redis.port}") private int redisPort; // @Value("${spring.ai.memory.redis.password}") // private String redisPassword; @Value("${spring.ai.memory.redis.timeout}") private int redisTimeout; @Bean public RedisChatMemoryRepository redisChatMemoryRepository() { return RedisChatMemoryRepository.builder() .host(redisHost) .port(redisPort) // 若没有设置密码则注释该项 // .password(redisPassword) .timeout(redisTimeout) .build(); } }
2.4 配置ChatClient实现连续对话
import com.alibaba.cloud.ai.memory.redis.RedisChatMemoryRepository; import org.springframework.ai.chat.client.ChatClient; import org.springframework.ai.chat.client.advisor.MessageChatMemoryAdvisor; import org.springframework.ai.chat.memory.MessageWindowChatMemory; import org.springframework.ai.chat.model.ChatModel; import org.springframework.web.bind.annotation.GetMapping; import org.springframework.web.bind.annotation.RequestMapping; import org.springframework.web.bind.annotation.RestController; import static org.springframework.ai.chat.memory.ChatMemory.CONVERSATION_ID; @RestController @RequestMapping("/redis") public class RedisMemoryController { private final ChatClient chatClient; private final int MAXMESSAGES = 10; private final MessageWindowChatMemory messageWindowChatMemory; public RedisMemoryController(ChatModel dashscopeChatModel, RedisChatMemoryRepository redisChatMemoryRepository) { this.messageWindowChatMemory = MessageWindowChatMemory.builder() .chatMemoryRepository(redisChatMemoryRepository) .maxMessages(MAXMESSAGES) .build(); this.chatClient = ChatClient.builder(dashscopeChatModel) .defaultAdvisors( MessageChatMemoryAdvisor.builder(messageWindowChatMemory) .build() ) .build(); } @GetMapping("/call") public String call(String msg, String cid) { return chatClient.prompt(msg) .advisors( a -> a.param(CONVERSATION_ID, cid) ) .call().content(); } }
小结
通过以上代码大家也可以看出来,使用 Spring AI 实现连续对话是比较复杂的,需要自己实现数据库增删改查的代码,并且重写 ChatMemory 才能实现连续对话功能;而 Spring AI Alibaba 因为内置了连续对话的多种实现(Redis 和其他数据库),所以只需要简单配置就可以实现了。
本文已收录到我的面试小站 www.javacn.site,其中包含的内容有:场景题、SpringAI、SpringAIAlibaba、并发编程、MySQL、Redis、Spring、Spring MVC、Spring Boot、Spring Cloud、MyBatis、JVM、设计模式、消息队列、AI常见面试题等。