LangChain4j 入门教程

1. LangChain4j 介绍

LangChain4j 是一个用于简化 Java 应用程序集成大型语言模型（LLMs）的库。它提供了统一的 API、丰富的工具箱和众多示例，帮助你快速构建各种 LLM 驱动的应用。

支持的模型

LangChain4j 支持多种大型语言模型，包括但不限于：

• OpenAI 的 GPT-3
• Anthropic 的 Claude
• 腾讯的 Mixie
• 智谱AI 的 Qwen

2. LangChain4j 的使用

单次对话示例

以下是一个简单的单次对话代码示例：

import dev.langchain4j.model.chat.ChatLanguageModel;
import dev.langchain4j.model.openai.OpenAiChatModel;

public class SingleChatDemo {
    public static void main(String[] args) {
        ChatLanguageModel model = OpenAiChatModel.withApiKey("your-api-key");
        String response = model.generate("Hello, how are you?");
        System.out.println(response);
    }
}

多轮对话示例

多轮对话需要使用 ChatMemory 来维持上下文：

import dev.langchain4j.model.chat.ChatLanguageModel;
import dev.langchain4j.model.chat.MessageWindowChatMemory;
import dev.langchain4j.model.openai.OpenAiChatModel;
import dev.langchain4j.chains.conversation.ConversationalChain;

public class MultiChatDemo {
    public static void main(String[] args) {
        ChatLanguageModel model = OpenAiChatModel.withApiKey("your-api-key");
        ChatMemory memory = MessageWindowChatMemory.withMaxMessages(5);
        ConversationalChain chain = ConversationalChain.builder()
                .chatLanguageModel(model)
                .chatMemory(memory)
                .build();

        String response = chain.execute("Hello, how are you?");
        System.out.println(response);

        response = chain.execute("What's the weather like today?");
        System.out.println(response);
    }
}

流式响应示例

流式响应可以实时获取模型的输出：

import dev.langchain4j.model.chat.ChatLanguageModel;
import dev.langchain4j.model.openai.OpenAiChatModel;
import dev.langchain4j.chains.conversation.TokenStream;

public class StreamingChatDemo {
    public static void main(String[] args) {
        ChatLanguageModel model = OpenAiChatModel.withApiKey("your-api-key");
        TokenStream tokenStream = model.stream("Tell me a story.");

        tokenStream.onNext(token -> System.out.print(token))
                .onError(error -> error.printStackTrace())
                .start();
    }
}

3. 整合 Spring Boot

添加依赖

在 Spring Boot 项目的 pom.xml 文件中添加 LangChain4j Spring Boot Starter 依赖：

<dependency>
    <groupId>dev.langchain4j</groupId>
    <artifactId>langchain4j-open-ai-spring-boot-starter</artifactId>
    <version>0.34.0</version>
</dependency>

配置模型参数

在 application.properties 或 application.yml 文件中配置模型参数：

langchain4j.open-ai.chat-model.api-key=your-api-key
langchain4j.open-ai.chat-model.model-name=gpt-4o

创建控制器

创建一个 Spring Boot 控制器，并注入 ChatLanguageModel：

import dev.langchain4j.model.chat.ChatLanguageModel;
import org.springframework.web.bind.annotation.GetMapping;
import org.springframework.web.bind.annotation.RequestMapping;
import org.springframework.web.bind.annotation.RestController;

@RestController
@RequestMapping("/chat")
public class ChatController {
    private final ChatLanguageModel chatLanguageModel;

    @Autowired
    public ChatController(ChatLanguageModel chatLanguageModel) {
        this.chatLanguageModel = chatLanguageModel;
    }

    @GetMapping("/single")
    public String singleChat(String message) {
        return chatLanguageModel.generate(message);
    }

    @GetMapping("/multi")
    public String multiChat(String message) {
        // 实现多轮对话逻辑
        return "response";
    }
}

流式响应控制器

import dev.langchain4j.model.chat.ChatLanguageModel;
import dev.langchain4j.model.chat.StreamingChatLanguageModel;
import org.springframework.web.bind.annotation.GetMapping;
import org.springframework.web.bind.annotation.RestController;
import reactor.core.publisher.Flux;

@RestController
public class StreamingChatController {
    private final StreamingChatLanguageModel streamingChatLanguageModel;

    @Autowired
    public StreamingChatController(StreamingChatLanguageModel streamingChatLanguageModel) {
        this.streamingChatLanguageModel = streamingChatLanguageModel;
    }

    @GetMapping("/streaming")
    public Flux<String> streamingChat(String message) {
        return streamingChatLanguageModel.stream(message);
    }
}

总结

通过以上步骤，你可以在 Spring Boot 项目中成功整合 LangChain4j，并实现单次对话、多轮对话和流式输出。这使得你的应用程序能够利用大型语言模型的强大功能，同时提供流畅的用户体验。