赞
踩
在maven的setting.xml
<mirror>
<id>spring-milestones</id>
<name>Spring Milestones</name>
<mirrorOf>spring-milestones</mirrorOf>
<url>https://repo.spring.io/milestone</url>
</mirror>
这里以调用GPT4o为例
后来为了测试JDK8是否可用 将版本调整成了2.7.2 结果不能使用
因国内无法直接访问 按了个nginx代理
- server {
- #HTTPS的默认访问端口443。
- #如果未在此处配置HTTPS的默认访问端口,可能会造成Nginx无法启动。
- listen 443 ssl;
-
- #填写证书绑定的域名
- server_name xxxx xxxxxx;
-
- #填写证书文件绝对路径
- ssl_certificate /etc/letsencrypt/live/xxx.com/fullchain.pem;
- #填写证书私钥文件绝对路径
- ssl_certificate_key /etc/letsencrypt/live/xxxx.com/privkey.pem;
-
- ssl_session_cache shared:SSL:1m;
- ssl_session_timeout 5m;
-
- #自定义设置使用的TLS协议的类型以及加密套件(以下为配置示例,请您自行评估是否需要配置)
- #TLS协议版本越高,HTTPS通信的安全性越高,但是相较于低版本TLS协议,高版本TLS协议对浏览器的兼容性较差。
- ssl_ciphers ECDHE-RSA-AES128-GCM-SHA256:ECDHE:ECDH:AES:HIGH:!NULL:!aNULL:!MD5:!ADH:!RC4;
- #ssl_protocols TLSv1.1 TLSv1.2 TLSv1.3;
-
- #表示优先使用服务端加密套件。默认开启
- ssl_prefer_server_ciphers on;
- location /v1/{
- chunked_transfer_encoding off;
- proxy_cache off;
- proxy_buffering off;
- proxy_redirect off;
- proxy_ssl_protocols TLSv1 TLSv1.1 TLSv1.2;
- proxy_ssl_server_name on;
- proxy_http_version 1.1;
- proxy_set_header Host api.openai.com;
- proxy_set_header X-Real-IP $server_addr;
- proxy_set_header X-Forwarded-For $server_addr;
- proxy_set_header X-Real-Port $server_port;
- proxy_set_header Connection '';
- proxy_pass https://api.openai.com/;
-
- }
配置ChatClient另外种方式
特定的对话风格或角色,通常建议详细定义你希望模型如何回应,然后在你的应用中相应地构建提示 其实就是对话之前
你也可以为每一个接口设置单独的预定义角色 例如
以流的方式返回
这个在postMan中不好提现
可以直接在浏览器
可以看到它是以流的方式返回的,但是乱码
除了使用nginx转发 还可以用本地代理 只要在应用启动前配置好就行
关于ChatClient和ChatModel
ChatClient:较为通用
ChatModel:设置模型独有功能
模型选择
下面使用ChatModel演示调用
可以参数中指定, 也可以application.properties中指定
流式
演示文生图功能
文生语音
下面做法是有问题的,因为你保存到resources目录下的话 项目是打包之后运行的 因此你第一次运行保存之后是读不到的 要读只能重新启动,这里只是演示 就先这样了
重启应用
关于语音转文本
关于多模态(意思就是你可以要发文本,要发图片,要发语音)
意思只能用GPT4或4o模型才能用多模态
以上的代码
- package com.example.springai.controller;
-
- import org.springframework.ai.chat.client.ChatClient;
- import org.springframework.ai.chat.messages.Media;
- import org.springframework.ai.chat.messages.UserMessage;
- import org.springframework.ai.chat.model.ChatResponse;
- import org.springframework.ai.chat.prompt.Prompt;
- import org.springframework.ai.image.ImagePrompt;
- import org.springframework.ai.image.ImageResponse;
- import org.springframework.ai.openai.*;
- import org.springframework.ai.openai.api.OpenAiApi;
- import org.springframework.ai.openai.api.OpenAiAudioApi;
- import org.springframework.ai.openai.audio.speech.SpeechPrompt;
- import org.springframework.ai.openai.audio.speech.SpeechResponse;
- import org.springframework.ai.openai.audio.transcription.AudioTranscriptionPrompt;
- import org.springframework.ai.openai.audio.transcription.AudioTranscriptionResponse;
- import org.springframework.beans.factory.annotation.Autowired;
- import org.springframework.core.io.ClassPathResource;
- import org.springframework.http.HttpStatus;
- import org.springframework.http.ResponseEntity;
- import org.springframework.util.MimeTypeUtils;
- import org.springframework.web.bind.annotation.GetMapping;
- import org.springframework.web.bind.annotation.RestController;
- import reactor.core.publisher.Flux;
-
- import java.io.IOException;
- import java.nio.file.Files;
- import java.nio.file.Path;
- import java.nio.file.Paths;
- import java.util.List;
-
- /**
- * @author hrui
- * @date 2024/6/8 2:19
- */
- @RestController
- public class HelloGPT {
- @Autowired
- private ChatClient chatClient;
-
- // public HelloGPT(ChatClient.Builder chatClientBuilder) {
- // this.chatClient=chatClientBuilder.build();
- // }
-
- @GetMapping("/helloai")
- public Object generate(String userInput) {
- System.out.println("userInput:"+userInput);
- return chatClient.prompt()//提示词
- .user(userInput)//用户输入
- //.system("You are a helpful assistant.")
- .call()//调用
- .content();//返回文本
- }
-
-
- @GetMapping(value = "/helloai2",produces = "text/html;charset=UTF-8")
- public Flux<String> generate2(String userInput) {
- Flux<String> output = chatClient.prompt()
- .user(userInput)
- .stream()
- .content();
- return output;
- }
-
-
- @Autowired
- private OpenAiChatModel chatModel;//ChatModel可以自动装配 不需要@Bean
-
- @GetMapping("/helloai3")
- public Object generate3(String userInput) {
- // ChatResponse response = chatModel.call(
- // new Prompt(
- // "Generate the names of 5 famous pirates.",//这个其实好比用户消息
- // OpenAiChatOptions.builder()
- // .withModel("gpt-4-32k")
- // .withTemperature(0.8F)
- // .build()
- // ));
- ChatResponse response = chatModel.call(
- new Prompt(
- userInput,//底层封装成new UserMessage(userInput)
- OpenAiChatOptions.builder()
- .withModel("gpt-4-turbo")
- .withTemperature(0.8F)
- .build()
- ));
- return response.getResult().getOutput().getContent();
- }
-
-
- @GetMapping("/helloai4")
- public Flux<ChatResponse> generate4(String userInput) {
- System.out.println("userInput:"+userInput);
- Flux<ChatResponse> stream = chatModel.stream(
- new Prompt(
- userInput//底层封装成new UserMessage(userInput)
-
- ));
-
- return stream;
- }
-
-
- @Autowired
- private OpenAiImageModel openAiImageModel;
-
- @GetMapping("/helloai6")
- public Object generate6(String userInput) {
- ImageResponse response = openAiImageModel.call(
- new ImagePrompt(userInput,
- OpenAiImageOptions.builder()
- //设置图片清晰度
- .withQuality("hd")
- .withModel("dall-e-3")//默认就是这个
- .withN(1)//生成几张图片
- //默认高度和宽度
- .withHeight(1024)
- .withWidth(1024).build())
-
- );
-
- return response.getResult().getOutput().getUrl();
- }
-
-
- // @Autowired
- // private OpenAiAudioTranscriptionModel openAiAudioTranscriptionModel;
- @Autowired
- private OpenAiAudioSpeechModel openAiAudioSpeechModel;
-
- @GetMapping("/helloai7")
- public Object generate7(String userInput) {
- OpenAiAudioSpeechOptions speechOptions = OpenAiAudioSpeechOptions.builder()
- //用的模型
- .withModel(OpenAiAudioApi.TtsModel.TTS_1.value)
- //设置人声
- .withVoice(OpenAiAudioApi.SpeechRequest.Voice.ALLOY)
- .withResponseFormat(OpenAiAudioApi.SpeechRequest.AudioResponseFormat.MP3)
- .withSpeed(1.0f)//合成语音的速度 0.0最慢 1.0最快
- .build();
-
- SpeechPrompt speechPrompt = new SpeechPrompt(userInput, speechOptions);
- SpeechResponse response = openAiAudioSpeechModel.call(speechPrompt);
-
- byte[] output = response.getResult().getOutput();
-
- try {
- // 指定文件名,这里以当前时间戳命名以避免重名
- String filename = "audio_" + System.currentTimeMillis() + ".mp3";
- // 指定保存路径
- Path path = Paths.get("src/main/resources/static/" + filename);
- // 写入文件
- Files.write(path, output);
-
- // 获取可访问的URL,假设你的服务运行在 localhost:8080
- String fileUrl = "http://localhost:8082/" + filename;
- return ResponseEntity.ok(fileUrl);
- } catch (Exception e) {
- e.printStackTrace();
- return ResponseEntity.status(HttpStatus.INTERNAL_SERVER_ERROR).body("Error saving file");
- }
-
- }
-
-
-
- @Autowired
- private OpenAiAudioTranscriptionModel openAiTranscriptionModel;
-
- @GetMapping("/helloai8")
- public Object generate8() {
- //语音翻译的可选配置
- var transcriptionOptions = OpenAiAudioTranscriptionOptions.builder()
- .withResponseFormat(OpenAiAudioApi.TranscriptResponseFormat.TEXT)
- //温度 0f不需要创造力,语音是什么,翻译什么
- .withTemperature(0f)
- .build();
- var audioFile=new ClassPathResource("hello.m4a");
- //var audioFile = new FileSystemResource("/path/to/your/resource/speech/jfk.flac");
-
- AudioTranscriptionPrompt transcriptionRequest = new AudioTranscriptionPrompt(audioFile, transcriptionOptions);
- AudioTranscriptionResponse response = openAiTranscriptionModel.call(transcriptionRequest);
-
- return response.getResult().getOutput();
- }
-
-
-
- @GetMapping("/helloai9")
- public Object generate9() throws IOException {
- //图片二进制流
- byte[] imageData=new ClassPathResource("1717865484569.png").getContentAsByteArray();
-
- //用户信息
- var userMessage=new UserMessage("这是什么", List.of(new Media(MimeTypeUtils.IMAGE_PNG,imageData)));
-
- OpenAiChatOptions build = OpenAiChatOptions.builder()
- .withModel(OpenAiApi.ChatModel.GPT_4_O.getValue()).build();
-
- ChatResponse response=chatModel.call(new Prompt(userMessage,build));
-
-
- return response.getResult().getOutput().getContent();
- }
- }
-
关于Function call 应对大模型无法获取实时信息的弊端
比如说,我现在问 今天杭州火车东站的客流量是多少,这样GPT肯定无法回答
那么需要怎么办呢 我们可以调用第三方接口得到信息 再告知GPT 然后GPT回答问题
大概解释
例如 我问 杭州有多少人口
这类问题,GPT是无法回答的,当然现在GPT会查阅相关资料回答,假设
这句话里有 location和count两个关键
Function Call的作用是 当问GPT一个类似问题之后,GPT用Function Call来回调我们的应用并携带关键信息 location和count信息,我们的应用去查数据库也好,去调用第三方接口也好,再告诉GPT 那么GPT就可以回答这个问题了
当GPT携带参数过来的时候会调用Function.apply(){}这个方法,那么我们在这个方法里写我们自己的逻辑 可以查数据库,可以调用第三方接口
创建一个实现Function接口的类
如果报错
所有代码
- package com.example.springai.controller;
-
- import org.springframework.ai.chat.model.ChatResponse;
- import org.springframework.ai.chat.prompt.Prompt;
- import org.springframework.ai.openai.OpenAiChatModel;
- import org.springframework.ai.openai.OpenAiChatOptions;
- import org.springframework.beans.factory.annotation.Autowired;
- import org.springframework.web.bind.annotation.GetMapping;
- import org.springframework.web.bind.annotation.RequestParam;
- import org.springframework.web.bind.annotation.RestController;
-
-
- /**
- * @author hrui
- * @date 2024/6/9 1:16
- */
- @RestController
- public class FunctionCallController {
-
- @Autowired
- private OpenAiChatModel chatModel;
-
-
- @GetMapping("/functionCall")
- public Object functionCall(@RequestParam(value = "message",defaultValue = "杭州有多少个名字叫韩妹妹的人") String message) {
-
- OpenAiChatOptions aiChatOptions=OpenAiChatOptions.builder()
- //设置实现了Function接口的beanName
- .withFunction("locationCount")
- .withModel("gpt-3.5-turbo")
- .build();
-
- ChatResponse response=chatModel.call(new Prompt(message,aiChatOptions));
- //Flux<ChatResponse> stream = chatModel.stream(new Prompt(message, aiChatOptions));
- return response.getResult().getOutput().getContent();
- }
- }
- package com.example.springai.controller;
-
- import org.springframework.ai.chat.client.ChatClient;
- import org.springframework.ai.chat.messages.Media;
- import org.springframework.ai.chat.messages.UserMessage;
- import org.springframework.ai.chat.model.ChatResponse;
- import org.springframework.ai.chat.prompt.Prompt;
- import org.springframework.ai.image.ImagePrompt;
- import org.springframework.ai.image.ImageResponse;
- import org.springframework.ai.openai.*;
- import org.springframework.ai.openai.api.OpenAiApi;
- import org.springframework.ai.openai.api.OpenAiAudioApi;
- import org.springframework.ai.openai.audio.speech.SpeechPrompt;
- import org.springframework.ai.openai.audio.speech.SpeechResponse;
- import org.springframework.ai.openai.audio.transcription.AudioTranscriptionPrompt;
- import org.springframework.ai.openai.audio.transcription.AudioTranscriptionResponse;
- import org.springframework.beans.factory.annotation.Autowired;
- import org.springframework.core.io.ClassPathResource;
- import org.springframework.http.HttpStatus;
- import org.springframework.http.ResponseEntity;
- import org.springframework.util.MimeTypeUtils;
- import org.springframework.web.bind.annotation.GetMapping;
- import org.springframework.web.bind.annotation.RestController;
- import reactor.core.publisher.Flux;
-
- import java.io.IOException;
- import java.nio.file.Files;
- import java.nio.file.Path;
- import java.nio.file.Paths;
- import java.util.List;
-
- /**
- * @author hrui
- * @date 2024/6/8 2:19
- */
- @RestController
- public class HelloGPT {
- @Autowired
- private ChatClient chatClient;
-
- // public HelloGPT(ChatClient.Builder chatClientBuilder) {
- // this.chatClient=chatClientBuilder.build();
- // }
-
- @GetMapping("/helloai")
- public Object generate(String userInput) {
- System.out.println("userInput:"+userInput);
- return chatClient.prompt()//提示词
- .user(userInput)//用户输入
- //.system("You are a helpful assistant.")
- .call()//调用
- .content();//返回文本
- }
-
-
- @GetMapping(value = "/helloai2",produces = "text/html;charset=UTF-8")
- public Flux<String> generate2(String userInput) {
- Flux<String> output = chatClient.prompt()
- .user(userInput)
- .stream()
- .content();
- return output;
- }
-
-
- @Autowired
- private OpenAiChatModel chatModel;//ChatModel可以自动装配 不需要@Bean
-
- @GetMapping("/helloai3")
- public Object generate3(String userInput) {
- // ChatResponse response = chatModel.call(
- // new Prompt(
- // "Generate the names of 5 famous pirates.",//这个其实好比用户消息
- // OpenAiChatOptions.builder()
- // .withModel("gpt-4-32k")
- // .withTemperature(0.8F)
- // .build()
- // ));
- ChatResponse response = chatModel.call(
- new Prompt(
- userInput,//底层封装成new UserMessage(userInput)
- OpenAiChatOptions.builder()
- .withModel("gpt-4-turbo")
- .withTemperature(0.8F)
- .build()
- ));
- return response.getResult().getOutput().getContent();
- }
-
-
- @GetMapping("/helloai4")
- public Flux<ChatResponse> generate4(String userInput) {
- System.out.println("userInput:"+userInput);
- Flux<ChatResponse> stream = chatModel.stream(
- new Prompt(
- userInput//底层封装成new UserMessage(userInput)
-
- ));
-
- return stream;
- }
-
-
- @Autowired
- private OpenAiImageModel openAiImageModel;
-
- @GetMapping("/helloai6")
- public Object generate6(String userInput) {
- ImageResponse response = openAiImageModel.call(
- new ImagePrompt(userInput,
- OpenAiImageOptions.builder()
- //设置图片清晰度
- .withQuality("hd")
- .withModel("dall-e-3")//默认就是这个
- .withN(1)//生成几张图片
- //默认高度和宽度
- .withHeight(1024)
- .withWidth(1024).build())
-
- );
-
- return response.getResult().getOutput().getUrl();
- }
-
-
- // @Autowired
- // private OpenAiAudioTranscriptionModel openAiAudioTranscriptionModel;
- @Autowired
- private OpenAiAudioSpeechModel openAiAudioSpeechModel;
-
- @GetMapping("/helloai7")
- public Object generate7(String userInput) {
- OpenAiAudioSpeechOptions speechOptions = OpenAiAudioSpeechOptions.builder()
- //用的模型
- .withModel(OpenAiAudioApi.TtsModel.TTS_1.value)
- //设置人声
- .withVoice(OpenAiAudioApi.SpeechRequest.Voice.ALLOY)
- .withResponseFormat(OpenAiAudioApi.SpeechRequest.AudioResponseFormat.MP3)
- .withSpeed(1.0f)//合成语音的速度 0.0最慢 1.0最快
- .build();
-
- SpeechPrompt speechPrompt = new SpeechPrompt(userInput, speechOptions);
- SpeechResponse response = openAiAudioSpeechModel.call(speechPrompt);
-
- byte[] output = response.getResult().getOutput();
-
- try {
- // 指定文件名,这里以当前时间戳命名以避免重名
- String filename = "audio_" + System.currentTimeMillis() + ".mp3";
- // 指定保存路径
- Path path = Paths.get("src/main/resources/static/" + filename);
- // 写入文件
- Files.write(path, output);
-
- // 获取可访问的URL,假设你的服务运行在 localhost:8080
- String fileUrl = "http://localhost:8082/" + filename;
- return ResponseEntity.ok(fileUrl);
- } catch (Exception e) {
- e.printStackTrace();
- return ResponseEntity.status(HttpStatus.INTERNAL_SERVER_ERROR).body("Error saving file");
- }
-
- }
-
-
-
- @Autowired
- private OpenAiAudioTranscriptionModel openAiTranscriptionModel;
-
- @GetMapping("/helloai8")
- public Object generate8() {
- //语音翻译的可选配置
- var transcriptionOptions = OpenAiAudioTranscriptionOptions.builder()
- .withResponseFormat(OpenAiAudioApi.TranscriptResponseFormat.TEXT)
- //温度 0f不需要创造力,语音是什么,翻译什么
- .withTemperature(0f)
- .build();
- var audioFile=new ClassPathResource("hello.m4a");
- //var audioFile = new FileSystemResource("/path/to/your/resource/speech/jfk.flac");
-
- AudioTranscriptionPrompt transcriptionRequest = new AudioTranscriptionPrompt(audioFile, transcriptionOptions);
- AudioTranscriptionResponse response = openAiTranscriptionModel.call(transcriptionRequest);
-
- return response.getResult().getOutput();
- }
-
-
-
- @GetMapping("/helloai9")
- public Object generate9() throws IOException {
- //图片二进制流
- byte[] imageData=new ClassPathResource("1717865484569.png").getContentAsByteArray();
-
- //用户信息
- var userMessage=new UserMessage("这是什么", List.of(new Media(MimeTypeUtils.IMAGE_PNG,imageData)));
-
- OpenAiChatOptions build = OpenAiChatOptions.builder()
- .withModel(OpenAiApi.ChatModel.GPT_4_O.getValue()).build();
-
- ChatResponse response=chatModel.call(new Prompt(userMessage,build));
-
-
- return response.getResult().getOutput().getContent();
- }
- }
-
- package com.example.springai.function;
-
- import org.springframework.context.annotation.Description;
- import org.springframework.stereotype.Component;
- import org.springframework.util.StringUtils;
-
- import java.util.function.Function;
-
- /**
- * @author hrui
- * @date 2024/6/9 1:46
- */
- @Component("locationCount")
- @Description("某个地方有多少个名字某某的人")
- public class LocationCount implements Function<LocationCount.Request, LocationCount.Response> {
-
- @Override//接收GPT提取后的信息(自动调用该方法)
- public Response apply(Request request) {
- //好的(聪明的模型)当location和name确实时候根本不会来调用该方法
- if(!StringUtils.hasLength(request.location)||!StringUtils.hasLength(request.name)){
- return new Response("参数缺失,无需Function-Call,正常响应就可以");
- }
- //如果有值 调用数据库或者第三方接口
-
- return new Response("只有唯一强调唯一1个");
- }
-
-
- //密封类(好比有了@Data作用) 接收GPT信息
- public record Request(String name, String location) {}
- //密封类,最终响应GPT信息
- public record Response(String message) {}
-
- }
-
-
- package com.example.springai;
-
- import org.springframework.ai.chat.client.ChatClient;
- import org.springframework.ai.chat.model.ChatModel;
- import org.springframework.boot.SpringApplication;
- import org.springframework.boot.autoconfigure.SpringBootApplication;
- import org.springframework.context.annotation.Bean;
-
- @SpringBootApplication
- public class SpringaiApplication {
-
-
- @Bean
- public ChatClient chatClient(ChatClient.Builder chatClientBuilder){
- return chatClientBuilder.defaultSystem("你现在不是ChatGPT,也不是人工智能AI,现在开始你是一个正常人类,你的名字叫hrui")
- .build();
- }
- public static void main(String[] args) {
- //System.setProperty("proxyType","4");
- System.setProperty("proxyHost","127.0.0.1");
- System.setProperty("proxyPort","7897");
- //System.setProperty("proxySet","true");
- SpringApplication.run(SpringaiApplication.class, args);
- }
-
- }
Copyright © 2003-2013 www.wpsshop.cn 版权所有,并保留所有权利。