Java实现调用openAI接口返回流式数据

一，添加依赖

<dependency>
    <groupId>com.github.plexpt</groupId>
    <artifactId>chatgpt</artifactId>
    <version>4.0.7</version>
</dependency>

二，重写SseEmitter 改为UTF-8编码

import org.springframework.http.HttpHeaders;
import org.springframework.http.MediaType;
import org.springframework.http.server.ServerHttpResponse;
import org.springframework.web.servlet.mvc.method.annotation.SseEmitter;

import java.nio.charset.Charset;

/**
 * @ClassName: SseEmitterUTF8
 * @Description: 重写SseEmitter 改为UTF-8编码
 * @Authror: XQD
 * @Date: 2023/11/6 18:51
 */
public class SseEmitterUTF8 extends SseEmitter {

    public SseEmitterUTF8(Long timeout) {
        super(timeout);
    }

    @Override
    protected void extendResponse(ServerHttpResponse outputMessage) {
        super.extendResponse(outputMessage);

        HttpHeaders headers = outputMessage.getHeaders();
        headers.setContentType( new MediaType("text", "event-stream", Charset.forName("UTF-8")));
    }
}

三，核心方法

private static final String OPENAI_API_HOST = "chatgpt的接口";


    @GetMapping(value = "/v1/stream")
    @PermitAll
    public SseEmitter streamEvents(@RequestParam String query) {
        SseEmitterUTF8 sseEmitter = new SseEmitterUTF8(-1L);

        SseStreamListener listener = new SseStreamListener(sseEmitter);
        //ConsoleStreamListener listener = new ConsoleStreamListener();
        Message message = Message.of(query);
        ChatCompletion chatCompletion = ChatCompletion.builder()
                .model("llama2")
                .messages(Arrays.asList(message))
                .stream(true)
                .temperature(1)
                .topP(1)
                .presencePenalty(0)
                .frequencyPenalty(0)
                .maxTokens(250)
                .build();

        // 不需要代理的话，注销此行
        //Proxy proxy = Proxys.http("192.168.1.98", 7890);
        ChatGPTStream chatGPTStream = ChatGPTStream.builder()
                .timeout(600)
                .apiKey("empty")
                //.proxy(proxy)
                .apiHost(OPENAI_API_HOST)
                .build()
                .init();

        chatGPTStream.streamChatCompletion(chatCompletion, listener);
        listener.setOnComplate(msg -> {
            //回答完成，可以做一些事情
            //sseEmitter.complete();
            System.out.println(msg);
        });
        return sseEmitter;
    }

可以在浏览器测试