DeepSeek API 调用 - Spring Boot 实现

DeepSeek API 调用 - Spring Boot 实现

1. 项目依赖

在 pom.xml 中添加以下依赖：

<dependencies><dependency><groupId>org.springframework.boot</groupId><artifactId>spring-boot-starter-webflux</artifactId></dependency><dependency><groupId>org.projectlombok</groupId><artifactId>lombok</artifactId></dependency><dependency><groupId>com.fasterxml.jackson.core</groupId><artifactId>jackson-databind</artifactId></dependency>
</dependencies>

2. 项目结构

deepseek-project/
├── src/main/java/com/example/deepseek/
│   ├── DeepSeekApplication.java
│   ├── config/
│   │   └── DeepSeekConfig.java
│   ├── model/
│   │   ├── ChatRequest.java
│   │   ├── ChatResponse.java
│   │   └── Message.java
│   └── service/
│       └── DeepSeekService.java
└── conversation.txt

3. 完整代码实现

3.1 配置类 DeepSeekConfig.java

package com.example.deepseek.config;import org.springframework.beans.factory.annotation.Value;
import org.springframework.context.annotation.Configuration;@Configuration
@Getter
public class DeepSeekConfig {@Value("${deepseek.api.url}")private String apiUrl;@Value("${deepseek.api.key}")private String apiKey;
}

3.2 请求/响应模型

Message.java:

package com.example.deepseek.model;import lombok.Data;@Data
public class Message {private String role;private String content;
}

ChatRequest.java:

package com.example.deepseek.model;import lombok.Data;
import java.util.List;@Data
public class ChatRequest {private String model = "deepseek-ai/DeepSeek-V3";private List<Message> messages;private boolean stream = true;private int max_tokens = 2048;private double temperature = 0.7;private double top_p = 0.7;private int top_k = 50;private double frequency_penalty = 0.5;private int n = 1;private ResponseFormat response_format = new ResponseFormat("text");@Datapublic static class ResponseFormat {private String type;public ResponseFormat(String type) {this.type = type;}}
}

ChatResponse.java:

package com.example.deepseek.model;import lombok.Data;
import java.util.List;@Data
public class ChatResponse {private List<Choice> choices;@Datapublic static class Choice {private Delta delta;}@Datapublic static class Delta {private String content;}
}

3.3 服务类 DeepSeekService.java

package com.example.deepseek.service;import com.example.deepseek.config.DeepSeekConfig;
import com.example.deepseek.model.ChatRequest;
import com.example.deepseek.model.ChatResponse;
import com.example.deepseek.model.Message;
import com.fasterxml.jackson.databind.ObjectMapper;
import lombok.RequiredArgsConstructor;
import org.springframework.stereotype.Service;
import org.springframework.web.reactive.function.client.WebClient;
import reactor.core.publisher.Flux;import java.io.FileWriter;
import java.io.IOException;
import java.io.PrintWriter;
import java.time.LocalDateTime;
import java.time.format.DateTimeFormatter;
import java.util.Collections;
import java.util.Scanner;@Service
@RequiredArgsConstructor
public class DeepSeekService {private final DeepSeekConfig config;private final WebClient.Builder webClientBuilder;private final ObjectMapper objectMapper = new ObjectMapper();public void startInteractiveChat() {try (Scanner scanner = new Scanner(System.in);PrintWriter fileWriter = new PrintWriter(new FileWriter("conversation.txt", true))) {while (true) {System.out.print("\n请输入您的问题 (输入 q 退出): ");String question = scanner.nextLine().trim();if ("q".equalsIgnoreCase(question)) {System.out.println("程序已退出");break;}// 保存问题saveToFile(fileWriter, question, true);// 发起对话请求Flux<String> responseFlux = sendChatRequest(question);StringBuilder fullResponse = new StringBuilder();responseFlux.doOnNext(chunk -> {System.out.print(chunk);fullResponse.append(chunk);}).doOnComplete(() -> {// 保存完整回复saveToFile(fileWriter, fullResponse.toString(), false);System.out.println("\n----------------------------------------");fileWriter.println("\n----------------------------------------");fileWriter.flush();}).blockLast();}} catch (IOException e) {e.printStackTrace();}}private Flux<String> sendChatRequest(String question) {ChatRequest request = new ChatRequest();Message userMessage = new Message();userMessage.setRole("user");userMessage.setContent(question);request.setMessages(Collections.singletonList(userMessage));return webClientBuilder.build().post().uri(config.getApiUrl()).header("Authorization", "Bearer " + config.getApiKey()).header("Content-Type", "application/json").bodyValue(request).retrieve().bodyToFlux(String.class).filter(line -> line.startsWith("data: ") && !line.equals("data: [DONE]")).map(line -> {try {String jsonStr = line.substring(6);ChatResponse response = objectMapper.readValue(jsonStr, ChatResponse.class);return response.getChoices().get(0).getDelta().getContent();} catch (Exception e) {return "";}}).filter(content -> !content.isEmpty());}private void saveToFile(PrintWriter fileWriter, String content, boolean isQuestion) {String timestamp = LocalDateTime.now().format(DateTimeFormatter.ofPattern("yyyy-MM-dd HH:mm:ss"));if (isQuestion) {fileWriter.printf("\n[%s] Question:\n%s\n\n[%s] Answer:\n", timestamp, content, timestamp);} else {fileWriter.print(content);}fileWriter.flush();}
}

3.4 主应用类 DeepSeekApplication.java

package com.example.deepseek;import com.example.deepseek.service.DeepSeekService;
import org.springframework.boot.SpringApplication;
import org.springframework.boot.autoconfigure.SpringBootApplication;
import org.springframework.context.ConfigurableApplicationContext;@SpringBootApplication
public class DeepSeekApplication {public static void main(String[] args) {ConfigurableApplicationContext context = SpringApplication.run(DeepSeekApplication.class, args);DeepSeekService deepSeekService = context.getBean(DeepSeekService.class);deepSeekService.startInteractiveChat();}
}

3.5 配置文件 application.properties

deepseek.api.url=https://api.siliconflow.cn/v1/chat/completions
deepseek.api.key=YOUR_API_KEY

4. 代码详解

4.1 关键特性

1. 使用 Spring WebFlux 的响应式编程模型
2. 流式处理 API 响应
3. 文件记录对话
4. 错误处理和异常管理

4.2 主要组件

• DeepSeekConfig: 管理 API 配置
• DeepSeekService: 处理对话逻辑和 API 交互
• 模型类: 定义请求和响应结构

5. 使用方法

1. 替换 application.properties 中的 YOUR_API_KEY
2. 运行 DeepSeekApplication
3. 在控制台输入问题
4. 输入 ‘q’ 退出程序
5. 查看 conversation.txt 获取对话记录

6. 性能和可扩展性

• 使用响应式编程提高并发性能
• 灵活的配置管理
• 易于扩展和定制

7. 注意事项

• 确保正确配置 API Key
• 处理网络异常
• 注意内存使用

总结

Spring Boot 实现提供了一个健壮、可扩展的 DeepSeek API 调用方案，利用响应式编程提供高效的流式对话体验。

立即体验

快来体验 DeepSeek：https://cloud.siliconflow.cn/i/vnCCfVaQ