SpringBoot3.0整合chatGPT的完整步骤

Java
485
0
0
2023-07-08
目录
  • 导读
  • 新建父项目
  • 1、快速新建父项目
  • 2、在pom.xml中引入SpringBoot3.0
  • 3、删除父项目的src文件夹
  • 新建openai-spring-boot-starter模块
  • 1、新增模块
  • 2、在模块中引入相关依赖
  • 3、定义模块外部属性有那些
  • 4、实现核心业务逻辑
  • 5、配置自动装配
  • 新建openai-starter-test模块
  • 新增模块
  • 导入依赖
  • 创建启动类
  • 配置属性
  • 编写测试类
  • 运行报错
  • 总结 

导读

12月总体来说互联网的技术圈是非常热闹的,chatGPT爆火,SpringBoot3.0发布等重磅陆消息续进入大家的视线,而本文作者将以技术整合的角度,带大家把最火的两个技术整合在一起。读完本文,你将熟悉SpringBoot3.0自定stater模块的操作流程,并熟悉OpenAi为chatGPT提供的49种场景。

项目项目我已经提交GITEE:https://gitee.com/miukoo/openai-spring 欢迎Star

新建父项目

我们这个项目分为starter和test两个模块,因此需要一个父项目来包裹。

1、快速新建父项目

2、在pom.xml中引入SpringBoot3.0

  • 项目的父工程设置成SpringBoot3.0
  • 在项目中定义openai的版本并导入(com.theokanning.openai-gpt3-java)
<project xmlns="http://maven.apache.org/POM/.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
         xsi:schemaLocation="http://maven.apache.org/POM/.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
    <modules>
        <module>openai-spring-boot-starter</module>
        <module>openai-starter-test</module>
    </modules>
    <packaging>pom</packaging>
    <modelVersion>.0.0</modelVersion>

    <groupId>cn.gjsm</groupId>
    <artifactId>openai-spring</artifactId>
    <version>.0-SNAPSHOT</version>

    <properties>
        <project.build.sourceEncoding>UTF-</project.build.sourceEncoding>
        <project.reporting.sourceEncoding>UTF-</project.reporting.sourceEncoding>
        <openai-version>.8.1</openai-version>
    </properties>

    <parent>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-parent</artifactId>
        <version>.0.0</version>
    </parent>
    <dependencyManagement>
        <dependencies>
            <dependency>
                <groupId>com.theokanning.openai-gpt-java</groupId>
                <artifactId>client</artifactId>
                <version>${openai-version}</version>
            </dependency>
        </dependencies>
    </dependencyManagement>
    <dependencies>
        <dependency>
            <groupId>org.projectlombok</groupId>
            <artifactId>lombok</artifactId>
        </dependency>
    </dependencies>
</project>

3、删除父项目的src文件夹

新建openai-spring-boot-starter模块

openai-spring-boot-starter 模块主要用来封装openai的核心api,该模块就是springboot自定starter的标准5步:
  • 新建模块
  • 在模块中引入相关依赖
  • 定义模块外部属性有那些
  • 实现核心业务逻辑
  • 配置自动装配

1、新增模块

注意模块名称的规范:非官方starter命名规则为 模块名称+'-spring-boot-starter’结尾

2、在模块中引入相关依赖

<project xmlns="http://maven.apache.org/POM/.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
         xsi:schemaLocation="http://maven.apache.org/POM/.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
    <packaging>pom</packaging>
    <parent>
        <artifactId>openai-spring</artifactId>
        <groupId>cn.gjsm</groupId>
        <version>.0-SNAPSHOT</version>
    </parent>
    <modelVersion>.0.0</modelVersion>

    <groupId>cn.gjsm</groupId>
    <artifactId>openai-spring-boot-starter</artifactId>
    <version>.0-SNAPSHOT</version>


    <dependencies>
        <!-- 自定义starter必须导入的依赖 -->
        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter</artifactId>
        </dependency>
        <!-- 这个包可以用来支持自定义属性的输入提示 -->
        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-configuration-processor</artifactId>
            <optional>true</optional>
        </dependency>
        <!-- 导入openai依赖,版本在父项目中已经约束 -->
        <dependency>
            <groupId>com.theokanning.openai-gpt-java</groupId>
            <artifactId>client</artifactId>
        </dependency>
    </dependencies>

</project>

3、定义模块外部属性有那些

通过@ConfigurationProperties配置一个类,这个类中的属性将从外部的application.yml中读取。在这里OpenAi需要两个属性需要配置,一是token秘钥,一是timeout超时时间。关于timeout可以配置时间长一点,因为OpenAi在国外有些慢。

package cn.gjsm.miukoo.properties;

import cn.gjsm.miukoo.utils.OpenAiUtils;
import lombok.Data;
import org.springframework.beans.factory.InitializingBean;
import org.springframework.boot.context.properties.ConfigurationProperties;

@Data
@ConfigurationProperties(prefix = "openai")
public class OpenAiProperties implements InitializingBean {
    // 秘钥
    String token;
    // 超时时间
    Integer timeout;

    // 设置属性时同时设置给OpenAiUtils
    @Override
    public void afterPropertiesSet() throws Exception {
        OpenAiUtils.OPENAPI_TOKEN = token;
        OpenAiUtils.TIMEOUT = timeout;
    }
}

4、实现核心业务逻辑

核心业务逻辑指的就是你自定义这个starter可以提供给其它模块那些api使用;在这里我们直接通过一个静态类工具OpenAiUtils,这样在引入该模块后,其它模块直接可调用该静态工具类,使用便捷一些。

同时在这个类中提供openai官方49种场景想对应的方法。

package cn.gjsm.miukoo.utils;

import cn.gjsm.miukoo.pojos.OpenAi;
import com.theokanning.openai.OpenAiService;
import com.theokanning.openai.completion.CompletionChoice;
import com.theokanning.openai.completion.CompletionRequest;
import org.springframework.util.StringUtils;

import java.util.*;

/**
 * 调用OpenAi的中方法
 */
public class OpenAiUtils {
    public static final Map<String, OpenAi> PARMS = new HashMap<>();

    static {
        PARMS.put("OpenAi", new OpenAi("OpenAi01", "问&答", "依据现有知识库问&答", "text-davinci-003", "Q: %s\nA:", 0.0, 1.0, 1.0, 0.0, 0.0, "\n"));
        PARMS.put("OpenAi", new OpenAi("OpenAi02", "语法纠正", "将句子转换成标准的英语,输出结果始终是英文", "text-davinci-003", "%s", 0.0, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi03", "内容概况", "将一段话,概况中心", "text-davinci-003", "Summarize this for a second-grade student:\n%s", 0.7, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi04", "生成OpenAi的代码", "一句话生成OpenAi的代码", "code-davinci-002", "\"\"\"\nUtil exposes the following:\nutil.openai() -> authenticates & returns the openai module, which has the following functions:\nopenai.Completion.create(\n    prompt=\"<my prompt>\", # The prompt to start completing from\n    max_tokens=123, # The max number of tokens to generate\n    temperature=1.0 # A measure of randomness\n    echo=True, # Whether to return the prompt in addition to the generated completion\n)\n\"\"\"\nimport util\n\"\"\"\n%s\n\"\"\"\n\n", 0.0, 1.0, 1.0, 0.0, 0.0, "\"\"\""));
        PARMS.put("OpenAi", new OpenAi("OpenAi05", "程序命令生成", "一句话生成程序的命令,目前支持操作系统指令比较多", "text-davinci-003", "Convert this text to a programmatic command:\n\nExample: Ask Constance if we need some bread\nOutput: send-msg `find constance` Do we need some bread?\n\n%s", 0.0, 1.0, 1.0, 0.2, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi06", "语言翻译", "把一种语法翻译成其它几种语言", "text-davinci-003", "Translate this into %s:\n%s", 0.3, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi07", "Stripe国际API生成", "一句话生成Stripe国际支付API", "code-davinci-002", "\"\"\"\nUtil exposes the following:\n\nutil.stripe() -> authenticates & returns the stripe module; usable as stripe.Charge.create etc\n\"\"\"\nimport util\n\"\"\"\n%s\n\"\"\"", 0.0, 1.0, 1.0, 0.0, 0.0, "\"\"\""));
        PARMS.put("OpenAi", new OpenAi("OpenAi08", "SQL语句生成", "依据上下文中的表信息,生成SQL语句", "code-davinci-002", "### %s SQL tables, 表字段信息如下:\n%s\n#\n### %s\n %s", 0.0, 1.0, 1.0, 0.0, 0.0, "# ;"));
        PARMS.put("OpenAi", new OpenAi("OpenAi09", "结构化生成", "对于非结构化的数据抽取其中的特征生成结构化的表格", "text-davinci-003", "A table summarizing, use Chinese:\n%s\n", 0.0, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi10", "信息分类", "把一段信息继续分类", "text-davinci-003", "%s\n分类:", 0.0, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi11", "Python代码解释", "把代码翻译成文字,用来解释程序的作用", "code-davinci-002", "# %s \n %s \n\n# 解释代码作用\n\n#", 0.0, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi12", "文字转表情符号", "将文本编码成表情服务", "text-davinci-003", "转换文字为表情。\n%s:", 0.8, 1.0, 1.0, 0.0, 0.0, "\n"));
        PARMS.put("OpenAi", new OpenAi("OpenAi13", "时间复杂度计算", "求一段代码的时间复杂度", "text-davinci-003", "%s\n\"\"\"\n函数的时间复杂度是", 0.0, 1.0, 1.0, 0.0, 0.0, "\n"));
        PARMS.put("OpenAi", new OpenAi("OpenAi14", "程序代码翻译", "把一种语言的代码翻译成另外一种语言的代码", "code-davinci-002", "##### 把这段代码从%s翻译成%s\n### %s\n    \n   %s\n    \n### %s", 0.0, 1.0, 1.0, 0.0, 0.0, "###"));
        PARMS.put("OpenAi", new OpenAi("OpenAi15", "高级情绪评分", "支持批量列表的方式检查情绪", "text-davinci-003", "对下面内容进行情感分类:\n%s\"\n情绪评级:", 0.0, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi16", "代码解释", "对一段代码进行解释", "code-davinci-002", "代码:\n%s\n\"\"\"\n上面的代码在做什么:\n1. ", 0.0, 1.0, 1.0, 0.0, 0.0, "\"\"\""));
        PARMS.put("OpenAi", new OpenAi("OpenAi17", "关键字提取", "提取一段文本中的关键字", "text-davinci-003", "抽取下面内容的关键字:\n%s", 0.5, 1.0, 1.0, 0.8, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi18", "问题解答", "类似解答题", "text-davinci-003", "Q: %s\nA: ?", 0.0, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi19", "广告设计", "给一个产品设计一个广告", "text-davinci-003", "为下面的产品创作一个创业广告,用于投放到抖音上:\n产品:%s.", 0.5, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi20", "产品取名", "依据产品描述和种子词语,给一个产品取一个好听的名字", "text-davinci-003", "产品描述: %s.\n种子词: %s.\n产品名称: ", 0.8, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi21", "句子简化", "把一个长句子简化成一个短句子", "text-davinci-003", "%s\nTl;dr: ", 0.7, 1.0, 1.0, 0.0, 1.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi22", "修复代码Bug", "自动修改代码中的bug", "code-davinci-002", "##### 修复下面代码的bug\n### %s\n %s\n###  %s\n", 0.0, 1.0, 1.0, 0.0, 0.0, "###"));
        PARMS.put("OpenAi", new OpenAi("OpenAi23", "表格填充数据", "自动为一个表格生成数据", "text-davinci-003", "spreadsheet ,%s rows:\n%s\n", 0.5, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi24", "语言聊天机器人", "各种开发语言的两天机器人", "code-davinci-002", "You: %s\n%s机器人:", 0.0, 1.0, 1.0, 0.5, 0.0, "You: "));
        PARMS.put("OpenAi", new OpenAi("OpenAi25", "机器学习机器人", "机器学习模型方面的机器人", "text-davinci-003", "You: %s\nML机器人:", 0.3, 1.0, 1.0, 0.5, 0.0, "You: "));
        PARMS.put("OpenAi", new OpenAi("OpenAi26", "清单制作", "可以列出各方面的分类列表,比如歌单", "text-davinci-003", "列出10%s:", 0.5, 1.0, 1.0, 0.52, 0.5, "11.0"));
        PARMS.put("OpenAi", new OpenAi("OpenAi27", "文本情绪分析", "对一段文字进行情绪分析", "text-davinci-003", "推断下面文本的情绪是积极的, 中立的, 还是消极的.\n文本: \"%s\"\n观点:", 0.0, 1.0, 1.0, 0.5, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi28", "航空代码抽取", "抽取文本中的航空diam信息", "text-davinci-003", "抽取下面文本中的航空代码:\n文本:\"%s\"\n航空代码:", 0.0, 1.0, 1.0, 0.0, 0.0, "\n"));
        PARMS.put("OpenAi", new OpenAi("OpenAi29", "生成SQL语句", "无上下文,语句描述生成SQL", "text-davinci-003", "%s", 0.3, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi30", "抽取联系信息", "从文本中抽取联系方式", "text-davinci-003", "从下面文本中抽取%s:\n%s", 0.0, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi31", "程序语言转换", "把一种语言转成另外一种语言", "code-davinci-002", "#%s to %s:\n%s:%s\n\n%s:", 0.0, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi32", "好友聊天", "模仿好友聊天", "text-davinci-003", "You: %s\n好友:", 0.5, 1.0, 1.0, 0.5, 0.0, "You:"));
        PARMS.put("OpenAi", new OpenAi("OpenAi33", "颜色生成", "依据描述生成对应颜色", "text-davinci-003", "%s:\nbackground-color: ", 0.0, 1.0, 1.0, 0.0, 0.0, ";"));
        PARMS.put("OpenAi", new OpenAi("OpenAi34", "程序文档生成", "自动为程序生成文档", "code-davinci-002", "# %s\n \n%s\n# 上述代码的详细、高质量文档字符串:\n\"\"\"", 0.0, 1.0, 1.0, 0.0, 0.0, "#\"\"\""));
        PARMS.put("OpenAi", new OpenAi("OpenAi35", "段落创作", "依据短语生成相关文短", "text-davinci-003", "为下面短语创建一个中文段:\n%s:\n", 0.5, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi36", "代码压缩", "把多行代码简单的压缩成一行", "code-davinci-002", "将下面%s代码转成一行:\n%s\n%s一行版本:", 0.0, 1.0, 1.0, 0.0, 0.0, ";"));
        PARMS.put("OpenAi", new OpenAi("OpenAi37", "故事创作", "依据一个主题创建一个故事", "text-davinci-003", "主题: %s\n故事创作:", 0.8, 1.0, 1.0, 0.5, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi38", "人称转换", "第一人称转第3人称", "text-davinci-003", "把下面内容从第一人称转为第三人称 (性别女):\n%s\n", 0.0, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi39", "摘要说明", "依据笔记生成摘要说明", "text-davinci-003", "将下面内容转换成将下%s摘要:\n%s", 0.0, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi40", "头脑风暴", "给定一个主题,让其生成一些主题相关的想法", "text-davinci-003", "头脑风暴一些关于%s的想法:", 0.6, 1.0, 1.0, 1.0, 1.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi41", "ESRB文本分类", "按照ESRB进行文本分类", "text-davinci-003", "Provide an ESRB rating for the following text:\\n\\n\\\"%s\"\\n\\nESRB rating:", 0.3, 1.0, 1.0, 0.0, 0.0, "\n"));
        PARMS.put("OpenAi", new OpenAi("OpenAi42", "提纲生成", "按照提示为相关内容生成提纲", "text-davinci-003", "为%s提纲:", 0.3, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi43", "美食制作(后果自负)", "依据美食名称和材料生成美食的制作步骤", "text-davinci-003", "依据下面成分和美食,生成制作方法:\n%s\n成分:\n%s\n制作方法:", 0.3, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi44", "AI聊天", "与AI机器进行聊天", "text-davinci-003", "Human: %s", 0.9, 1.0, 1.0, 0.0, 0.6, "Human:AI:"));
        PARMS.put("OpenAi", new OpenAi("OpenAi45", "摆烂聊天", "与讽刺机器进行聊天", "text-davinci-003", "Marv不情愿的回答问题.\nYou:%s\nMarv:", 0.5, 0.3, 1.0, 0.5, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi46", "分解步骤", "把一段文本分解成几步来完成", "text-davinci-003", "为下面文本生成次序列表,并增加列表数子: \n%s\n", 0.3, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi47", "点评生成", "依据文本内容自动生成点评", "text-davinci-003", "依据下面内容,进行点评:\n%s\n点评:", 0.5, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi48", "知识学习", "可以为学习知识自动解答", "text-davinci-003", "%s", 0.3, 1.0, 1.0, 0.0, 0.0, ""));
        PARMS.put("OpenAi", new OpenAi("OpenAi49", "面试", "生成面试题", "text-davinci-003", "创建10道%s相关的面试题(中文):\n", 0.5, 1.0, 10.0, 0.0, 0.0, ""));
    }

    public static String OPENAPI_TOKEN = "";
    public static Integer TIMEOUT = null;

    /**
     * 获取ai
     *
     * @param openAi
     * @param prompt
     * @return
     */
    public static List<CompletionChoice> getAiResult(OpenAi openAi, String prompt) {
        if (TIMEOUT == null || TIMEOUT <) {
            TIMEOUT =;
        }
        OpenAiService service = new OpenAiService(OPENAPI_TOKEN, TIMEOUT);
        CompletionRequest.CompletionRequestBuilder builder = CompletionRequest.builder()
                .model(openAi.getModel())
                .prompt(prompt)
                .temperature(openAi.getTemperature())
                .maxTokens()
                .topP(openAi.getTopP())
                .frequencyPenalty(openAi.getFrequencyPenalty())
                .presencePenalty(openAi.getPresencePenalty());
        if (!StringUtils.isEmpty(openAi.getStop())) {
            builder.stop(Arrays.asList(openAi.getStop().split(",")));
        }
        CompletionRequest completionRequest = builder.build();
        return service.createCompletion(completionRequest).getChoices();
    }

    /**
     * 问答
     *
     * @param question
     * @return
     */
    public static List<CompletionChoice> getQuestionAnswer(String question) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), question));
    }

    /**
     * 语法纠错
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getGrammarCorrection(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 将一段话,概况中心
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getSummarize(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 一句话生成OpenAi的代码
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getOpenAiApi(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 一句话生成程序的命令,目前支持操作系统指令比较多
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getTextToCommand(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 把一种语法翻译成其它几种语言
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getTranslatesLanguages(String text, String translatesLanguages) {
        if (StringUtils.isEmpty(translatesLanguages)) {
            translatesLanguages = ". French, 2. Spanish and 3. English";
        }
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), translatesLanguages, text));
    }

    /**
     * 一句话生成Stripe国际支付API
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getStripeApi(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }


    /**
     * 依据上下文中的表信息,生成SQL语句
     *
     * @param databaseType 数据库类型
     * @param tables       上午依赖的表和字段 Employee(id, name, department_id)
     * @param text         SQL描述
     * @param sqlType      sql类型,比如SELECT
     * @return
     */
    public static List<CompletionChoice> getStripeApi(String databaseType, List<String> tables, String text, String sqlType) {
        OpenAi openAi = PARMS.get("OpenAi");
        StringJoiner joiner = new StringJoiner("\n");
        for (int i =; i < tables.size(); i++) {
            joiner.add("# " + tables);
        }
        return getAiResult(openAi, String.format(openAi.getPrompt(), databaseType, joiner.toString(), text, sqlType));
    }

    /**
     * 对于非结构化的数据抽取其中的特征生成结构化的表格
     *
     * @param text 非结构化的数据
     * @return
     */
    public static List<CompletionChoice> getUnstructuredData(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 把一段信息继续分类
     *
     * @param text 要分类的文本
     * @return
     */
    public static List<CompletionChoice> getTextCategory(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 把一段信息继续分类
     *
     * @param codeType 代码类型,比如Python
     * @param code     要解释的代码
     * @return
     */
    public static List<CompletionChoice> getCodeExplain(String codeType, String code) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), codeType, code));
    }

    /**
     * 将文本编码成表情服务
     *
     * @param text 文本
     * @return
     */
    public static List<CompletionChoice> getTextEmoji(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 求一段代码的时间复杂度
     *
     * @param code 代码
     * @return
     */
    public static List<CompletionChoice> getTimeComplexity(String code) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), code));
    }


    /**
     * 把一种语言的代码翻译成另外一种语言的代码
     *
     * @param fromLanguage 要翻译的代码语言
     * @param toLanguage   要翻译成的代码语言
     * @param code         代码
     * @return
     */
    public static List<CompletionChoice> getTranslateProgramming(String fromLanguage, String toLanguage, String code) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), fromLanguage, toLanguage, fromLanguage, code, toLanguage));
    }

    /**
     * 支持批量列表的方式检查情绪
     *
     * @param texts 文本
     * @return
     */
    public static List<CompletionChoice> getBatchTweetClassifier(List<String> texts) {
        OpenAi openAi = PARMS.get("OpenAi");
        StringJoiner stringJoiner = new StringJoiner("\n");
        for (int i =; i < texts.size(); i++) {
            stringJoiner.add((i +) + ". " + texts.get(i));
        }
        return getAiResult(openAi, String.format(openAi.getPrompt(), stringJoiner.toString()));
    }

    /**
     * 对一段代码进行解释
     *
     * @param code 文本
     * @return
     */
    public static List<CompletionChoice> getExplainCOde(String code) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), code));
    }

    /**
     * 提取一段文本中的关键字
     *
     * @param text 文本
     * @return
     */
    public static List<CompletionChoice> getTextKeywords(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 事实回答答题
     *
     * @param text 文本
     * @return
     */
    public static List<CompletionChoice> getFactualAnswering(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 给一个产品设计一个广告
     *
     * @param text 文本
     * @return
     */
    public static List<CompletionChoice> getAd(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 依据产品描述和种子词语,给一个产品取一个好听的名字
     *
     * @param productDescription 产品描述
     * @param seedWords          种子词语
     * @return
     */
    public static List<CompletionChoice> getProductName(String productDescription, String seedWords) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), productDescription, seedWords));
    }

    /**
     * 把一个长句子简化成一个短句子
     *
     * @param text 长句子
     * @return
     */
    public static List<CompletionChoice> getProductName(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 自动修改代码中的bug
     *
     * @param codeType 语言类型
     * @param code     代码
     * @return
     */
    public static List<CompletionChoice> getBugFixer(String codeType, String code) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), codeType, code, codeType));
    }

    /**
     * 自动为一个表格生成数据
     *
     * @param rows    生成的行数
     * @param headers 数据表头,格式如:姓名| 年龄|性别|生日
     * @return
     */
    public static List<CompletionChoice> getFillData(int rows, String headers) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), rows, headers));
    }

    /**
     * 各种开发语言的两天机器人
     *
     * @param question             你的问题
     * @param programmingLanguages 语言 比如Java JavaScript
     * @return
     */
    public static List<CompletionChoice> getProgrammingLanguageChatbot(String question, String programmingLanguages) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), question, programmingLanguages));
    }

    /**
     * 机器学习模型方面的机器人
     *
     * @param question 你的问题
     * @return
     */
    public static List<CompletionChoice> getMLChatbot(String question) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), question));
    }

    /**
     * 可以列出各方面的分类列表,比如歌单
     *
     * @param text 清单描述
     * @return
     */
    public static List<CompletionChoice> getListMaker(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 对一段文字进行情绪分析
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getTweetClassifier(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 抽取文本中的航空代码信息
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getAirportCodeExtractor(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 无上下文,语句描述生成SQL
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getSQL(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 从文本中抽取联系方式
     *
     * @param extractContent 抽取内容描述
     * @param text
     * @return 从下面文本中抽取邮箱和电话:\n教育行业A股IPO第一股(股票代码)\n全国咨询/投诉热线:400-618-4000    举报邮箱:mc@itcast.cn
     */
    public static List<CompletionChoice> getExtractContactInformation(String extractContent, String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), extractContent, text));
    }

    /**
     * 把一种语言转成另外一种语言代码
     *
     * @param fromCodeType 当前代码类型
     * @param toCodeType   转换的代码类型
     * @param code
     * @return
     */
    public static List<CompletionChoice> getTransformationCode(String fromCodeType, String toCodeType, String code) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), fromCodeType, toCodeType, fromCodeType, code, toCodeType));
    }

    /**
     * 模仿好友聊天
     *
     * @param question
     * @return
     */
    public static List<CompletionChoice> getFriendChat(String question) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), question));
    }

    /**
     * 依据描述生成对应颜色
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getMoodToColor(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 自动为程序生成文档
     *
     * @param codeType 语言
     * @param code
     * @return
     */
    public static List<CompletionChoice> getCodeDocument(String codeType, String code) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), codeType, code));
    }

    /**
     * 依据短语生成相关文短
     *
     * @param text 短语
     * @return
     */
    public static List<CompletionChoice> getCreateAnalogies(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 把多行代码简单的压缩成一行
     *
     * @param codeType 语言
     * @param code
     * @return
     */
    public static List<CompletionChoice> getCodeLine(String codeType, String code) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), codeType, code, codeType));
    }

    /**
     * 依据一个主题创建一个故事
     *
     * @param topic 创作主题
     * @return
     */
    public static List<CompletionChoice> getStory(String topic) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), topic));
    }

    /**
     * 第一人称转第人称
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getStoryCreator(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 依据笔记生成摘要说明
     *
     * @param scene 生成的摘要场景
     * @param note  记录的笔记
     * @return
     */
    public static List<CompletionChoice> getNotesToSummary(String scene, String note) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), note));
    }

    /**
     * 给定一个主题,让其生成一些主题相关的想法
     *
     * @param topic 头脑风暴关键词
     * @return
     */
    public static List<CompletionChoice> getIdeaGenerator(String topic) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), topic));
    }

    /**
     * 按照ESRB进行文本分类
     *
     * @param text 文本
     * @return
     */
    public static List<CompletionChoice> getESRBRating(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 按照提示为相关内容生成提纲
     *
     * @param text 场景,比如 数据库软件生成大学毕业论文
     * @return
     */
    public static List<CompletionChoice> getEssayOutline(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 依据美食名称和材料生成美食的制作步骤
     *
     * @param name        美食名称
     * @param ingredients 美食食材
     * @return
     */
    public static List<CompletionChoice> getRecipeCreator(String name, List<String> ingredients) {
        OpenAi openAi = PARMS.get("OpenAi");
        StringJoiner joiner = new StringJoiner("\n");
        for (String ingredient : ingredients) {
            joiner.add(ingredient);
        }
        return getAiResult(openAi, String.format(openAi.getPrompt(), name, joiner.toString()));
    }

    /**
     * 与AI机器进行聊天
     *
     * @param question
     * @return
     */
    public static List<CompletionChoice> getAiChatbot(String question) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), question));
    }

    /**
     * 与讽刺机器进行聊天,聊天的机器人是一种消极情绪
     *
     * @param question
     * @return
     */
    public static List<CompletionChoice> getMarvChatbot(String question) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), question));
    }

    /**
     * 把一段文本分解成几步来完成
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getTurnDirection(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 依据文本内容自动生成点评
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getReviewCreator(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 可以为学习知识自动解答
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getStudyNote(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

    /**
     * 生成面试题
     *
     * @param text
     * @return
     */
    public static List<CompletionChoice> getInterviewQuestion(String text) {
        OpenAi openAi = PARMS.get("OpenAi");
        System.out.println(String.format(openAi.getPrompt(), text));
        return getAiResult(openAi, String.format(openAi.getPrompt(), text));
    }

}

5、配置自动装配

这一步是非常关键的,你的项目能在其他模块启动的时候就能够用,就必须配置这一步,而这一步有两小步:

  • 编写自动装配类
  • 配置自动装配类
编写自动装配类,参考代码:
package cn.gjsm.miukoo.config;

import cn.gjsm.miukoo.properties.OpenAiProperties;
import org.springframework.boot.context.properties.EnableConfigurationProperties;
import org.springframework.context.annotation.Configuration;

/**
 * 自动配置类
 */
@Configuration
@EnableConfigurationProperties(OpenAiProperties.class)
public class OpenAiAutoConfiguration {
}
配置自动装配类:

在resources文件夹下的META-INF/spring.factories文件中配置:

org.springframework.boot.autoconfigure.EnableAutoConfiguration=cn.gjsm.miukoo.config.OpenAiAutoConfiguration

新建openai-starter-test模块

经过上述五部我们就完成了chatGPT的stater的封装,接下来我们创建一个模块来测试。

新增模块

测试模块的名称最好是以test结尾

导入依赖

在测试模块中直接可以导入我们封装好的openai-spring-boot-starter,当然还有测试spring-boot-starter-test依赖。

<project xmlns="http://maven.apache.org/POM/.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
         xsi:schemaLocation="http://maven.apache.org/POM/.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
    <parent>
        <artifactId>openai-spring-boot-starter</artifactId>
        <groupId>cn.gjsm</groupId>
        <version>.0-SNAPSHOT</version>
        <relativePath>../openai-spring-boot-starter/pom.xml</relativePath>
    </parent>
    <modelVersion>.0.0</modelVersion>

    <groupId>cn.gjsm</groupId>
    <artifactId>openai-starter-test</artifactId>
    <version>.0-SNAPSHOT</version>


    <dependencies>
        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-test</artifactId>
        </dependency>
        <dependency>
            <groupId>cn.gjsm</groupId>
            <artifactId>openai-spring-boot-starter</artifactId>
            <version>.0-SNAPSHOT</version>
        </dependency>
    </dependencies>

</project>

创建启动类

我们计划使用SpringBoot去测试,因此需要创建一个启动类
package cn.gjsm.miukoo;

import org.springframework.boot.SpringApplication;
import org.springframework.boot.autoconfigure.SpringBootApplication;

@SpringBootApplication
public class OpenAiApplication {
    public static void main(String[] args) {
        SpringApplication.run(OpenAiApplication.class, args);
    }
}

配置属性

在测试模块的application.yml中,我们需要配置,我们在openai-spring-boot-starter中定义的两个属性

server:
  port:

openai:
  token: 你的token
  timeout:

编写测试类

我们在测试包下,新建一个测试类,即可直接调用我们在stater中封装的OpenAiUtils工具类,通过其来完成chatGPT功能调用。

package cn.gjsm.miukoo;

import cn.gjsm.miukoo.utils.OpenAiUtils;
import com.theokanning.openai.completion.CompletionChoice;
import org.junit.jupiter.api.Test;
import org.springframework.boot.test.context.SpringBootTest;

import java.util.List;

@SpringBootTest
public class OpenAiTest {

    /**
     * 测试问答
     */
    @Test
    public void testQA(){
        List<CompletionChoice> questionAnswer = OpenAiUtils.getQuestionAnswer("重庆今天的天气怎么样?");
        for (CompletionChoice completionChoice : questionAnswer) {
            System.out.println(completionChoice.getText());
        }
    }

    /**
     * 测试面试题生成
     */
    @Test
    public void testInterview(){
        List<CompletionChoice> results = OpenAiUtils.getInterviewQuestion("redis");
        for (CompletionChoice completionChoice : results) {
            System.out.println(completionChoice.getText());
        }
    }

}

tater中封装的OpenAiUtils工具类,通过其来完成chatGPT功能调用。

package cn.gjsm.miukoo;

import cn.gjsm.miukoo.utils.OpenAiUtils;
import com.theokanning.openai.completion.CompletionChoice;
import org.junit.jupiter.api.Test;
import org.springframework.boot.test.context.SpringBootTest;

import java.util.List;

@SpringBootTest
public class OpenAiTest {

    /**
     * 测试问答
     */
    @Test
    public void testQA(){
        List<CompletionChoice> questionAnswer = OpenAiUtils.getQuestionAnswer("重庆今天的天气怎么样?");
        for (CompletionChoice completionChoice : questionAnswer) {
            System.out.println(completionChoice.getText());
        }
    }

    /**
     * 测试面试题生成
     */
    @Test
    public void testInterview(){
        List<CompletionChoice> results = OpenAiUtils.getInterviewQuestion("redis");
        for (CompletionChoice completionChoice : results) {
            System.out.println(completionChoice.getText());
        }
    }

}

运行报错

如果你运行代码,出现下面错误,不应紧张,那是英文springboot3.0需要jdk17的版本

选中父项目右键打开项目配置创建,修改JDK为17版本即可,重新运行即可正常。