Sping boot 整合chatGPT

一、介绍

        Spring Boot 是一种用于创建独立的、基于 Spring 的应用程序的框架。它的主要目的是简化 Spring 应用程序的开发和部署过程。在本文中，我将向您展示如何使用 Spring Boot 框架与 ChatGPT 进行整合。

二、获取api密钥

OpenAI API

 注意，不会在生成秘密API密钥后再次显示密钥。

三、整合chatGPT

1、pom.xml中添加依赖

     <dependency>
            <groupId>com.theokanning.openai-gpt3-java</groupId>
            <artifactId>service</artifactId>
            <version>0.11.0</version>
     </dependency>
     <dependency>
            <groupId>com.theokanning.openai-gpt3-java</groupId>
            <artifactId>client</artifactId>
            <version>0.11.0</version>
     </dependency>

2、domain

package com.lxy.openai.domain;

import lombok.Data;

/**
 * @author lixingyu
 */
@Data
public class OpenAi {
    String model;
    Double temperature;
    Double topP;
    Double frequencyPenalty;
    Double presencePenalty;
    String stop;

    String role;

    public OpenAi(String model,Double temperature, Double topP, Double frequencyPenalty, Double presencePenalty) {
        this.model = model;
        this.temperature = temperature;
        this.topP = topP;
        this.frequencyPenalty = frequencyPenalty;
        this.presencePenalty = presencePenalty;
    }
}

3、工具类编写

package com.lxy.openai.utils;

import com.lxy.openai.domain.OpenAi;
import com.theokanning.openai.completion.chat.ChatCompletionChoice;
import com.theokanning.openai.completion.chat.ChatCompletionRequest;
import com.theokanning.openai.completion.chat.ChatMessage;
import com.theokanning.openai.service.OpenAiService;
import com.theokanning.openai.completion.CompletionChoice;
import com.theokanning.openai.completion.CompletionRequest;
import java.time.Duration;
import java.util.*;


/**
 * 调用OpenAi的gpt-3.5-turbo 与 text-davinci-003
 *
 * @author lixingyu
 */
public class OpenAiUtils {

    public static String API_KEY = "api密钥";
    public static Duration TIMEOUT = Duration.ofSeconds(300);

    public static final Map<String, OpenAi> DEFAULT_MODEL= new HashMap<>();

    static {
        DEFAULT_MODEL.put("turbo", new OpenAi("gpt-3.5-turbo", 0.2, 0.1, 2d, 2d));
        DEFAULT_MODEL.put("davinci-003", new OpenAi("text-davinci-003",  0.9, 0.1, 1.0, 1.0));
    }

    /**
     * 获取ai
     *
     * @param openAi
     * @param prompt
     * @return
     */
    public static List<CompletionChoice> getAiResult(OpenAi openAi, String prompt) {
        OpenAiService service = new OpenAiService(API_KEY, TIMEOUT);
        CompletionRequest.CompletionRequestBuilder builder = CompletionRequest.builder()
                .model(openAi.getModel())
                .prompt(prompt)
                .temperature(openAi.getTemperature())
                .maxTokens(1000)
                .topP(openAi.getTopP())
                .frequencyPenalty(openAi.getFrequencyPenalty())
                .presencePenalty(openAi.getPresencePenalty());
        CompletionRequest completionRequest = builder.build();
        return service.createCompletion(completionRequest).getChoices();

    }
    public static List<ChatCompletionChoice> getAiChatResult(OpenAi openAi, String prompt) {
        OpenAiService service = new OpenAiService(API_KEY, TIMEOUT);
        List<ChatMessage> messageList = new ArrayList<>();
        ChatMessage message = new ChatMessage();
        message.setContent(prompt);
        message.setRole("user");
        messageList.add(message);
        ChatCompletionRequest.ChatCompletionRequestBuilder requestBuilder = ChatCompletionRequest.builder() .model(openAi.getModel())
                .messages(messageList)
                .temperature(openAi.getTemperature())
                .maxTokens(3000)
                .topP(openAi.getTopP())
                .frequencyPenalty(openAi.getFrequencyPenalty())
                .presencePenalty(openAi.getPresencePenalty());
        ChatCompletionRequest build = requestBuilder.build();
        return service.createChatCompletion(build).getChoices();
    }


    /**
     * 文本
     * @param text 文本
     * @return
     */
    public static List<CompletionChoice> getRating(String text) {
        OpenAi openAi = DEFAULT_MODEL.get("davinci-003");
        return getAiResult(openAi, text);
    }

    /**
     * GPT-3.5型号
     * @param text 文本
     * @return
     */
    public static List<ChatCompletionChoice> getChatRating(String text) {
        OpenAi openAi = DEFAULT_MODEL.get("turbo");
        return getAiChatResult(openAi, text);
    }


}

四、总结

本文主要使用gpt-3.5-turbo与text-davinci-003，具体接口说明可参考官方说明OpenAI API。GPT-4是OpenAI最新推出的语言模型，但是它目前仍处于测试阶段，并且只向少数研究人员和企业客户授予了访问权限。此外，OpenAI也已经更新了API的定价和使用方式，需要访问他们的网站获取更多信息。