规划一个聊天机器人
智能化完全于依托于GPT, 而产品化是我们需要考虑的事情 比如，如何去构建一个聊天机器人 聊天机器人它的处理逻辑其实非常的清晰 我们输入问题调用 GPT 然后，GPT 给我们生成回答就可以了 需要注意的是，聊天机器人不同于调用API进行一个简单的测试 我们和聊天机器人的对话，可能是多轮的一个对话 在这时候，我们去调用API的时候，就需要将我们多轮的问答都传递给GPT才行新增一些实现类，结构如下
以下Java版代码来源于网络，可基于此逻辑，改造成其他编程语言
src
main java com.xxx.gpt.client util ChatContextHolder.java ChatBotClient.java … test java com.xxx.gpt.client.test FunctionCallTest.java …ChatContextHolder.java
package com.xxx.gpt.client.util; import com.xxx.gpt.client.entity.Message; import java.util.ArrayList; import java.util.HashMap; import java.util.List; import java.util.Map; public class ChatContextHolder { private static Map<String, List<Message>> context = new HashMap<>(); public static List<Message> get(String id) { // TODO 限制轮数，或者限制token数量 List<Message> messages = context.get(id); if (messages == null) { messages = new ArrayList<>(); context.put(id, messages); } return messages; } public static void add(String id, String msg) { Message message = Message.builder().content(msg).build(); add(id, message); } public static void add(String id, Message message) { List<Message> messages = context.get(id); if (messages == null) { messages = new ArrayList<>(); context.put(id, messages); } messages.add(message); } public static void remove(String id) { context.remove(id); } } ChatBotClient.java
package com.xxx.gpt.client; import com.xxx.gpt.client.entity.Message; import com.xxx.gpt.client.listener.ConsoleStreamListener; import com.xxx.gpt.client.util.ChatContextHolder; import com.xxx.gpt.client.util.Proxys; import lombok.SneakyThrows; import lombok.extern.slf4j.Slf4j; import java.io.BufferedReader; import java.io.IOException; import java.io.InputStreamReader; import java.net.Proxy; import java.util.ArrayList; import java.util.Arrays; import java.util.List; import java.util.UUID; import java.util.concurrent.CountDownLatch; import java.util.stream.Collectors; @Slf4j public class ChatBotClient { public static Proxy proxy = Proxy.NO_PROXY; public static void main(String[] args) { System.out.println("ChatGPT - Java command-line interface"); System.out.println("Press enter twice to submit your question."); System.out.println(); System.out.println("按两次回车以提交您的问题！！！"); String chatUuid = UUID.randomUUID().toString(); String key = "sk-adfas"; proxy = Proxys.http("127.0.0.1", 7890); while (true) { String prompt = getInput("\nYou:\n"); ChatGPTStreamClient chatGPT = ChatGPTStreamClient.builder() .apiKey(key) .proxy(proxy) .build() .init(); System.out.println("AI: "); // 卡住 CountDownLatch countDownLatch = new CountDownLatch(1); Message message = Message.of(prompt); ChatContextHolder.add(chatUuid, message); ConsoleStreamListener listener = new ConsoleStreamListener() { @Override public void onError(Throwable throwable, String response) { throwable.printStackTrace(); countDownLatch.countDown(); } }; listener.setOnComplate(msg -> { ChatContextHolder.add(chatUuid, Message.ofAssistant(msg)); countDownLatch.countDown(); }); chatGPT.streamChatCompletion(ChatContextHolder.get(chatUuid), listener); try { countDownLatch.await(); } catch (InterruptedException e) { e.printStackTrace(); } } } @SneakyThrows public static String getInput(String prompt) { System.out.print(prompt); BufferedReader reader = new BufferedReader(new InputStreamReader(System.in)); List<String> lines = new ArrayList<>(); String line; try { while ((line = reader.readLine()) != null && !line.isEmpty()) { lines.add(line); } } catch (IOException e) { e.printStackTrace(); } return lines.stream().collect(Collectors.joining("\n")); } } 
