LangChain4J入门实战：快速集成Chat API及常用配置详解-CSDN博客

本文链接：https://blog.csdn.net/qq_37064763/article/details/145997541

LangChain4J是什么？

官网地址： https://docs.langchain4j.dev/intro
在这里插入图片描述

快速开始：创建你的第一个Chat Demo

导入依赖

        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-web</artifactId>
        </dependency>
        <dependency>
            <groupId>dev.langchain4j</groupId>
            <artifactId>langchain4j-open-ai</artifactId>
        </dependency>
        <dependency>
            <groupId>dev.langchain4j</groupId>
            <artifactId>langchain4j</artifactId>
        </dependency>

配置类

@Configuration(proxyBeanMethods = false)
public class ChatLanguageModelConfig {
    @Bean
    public ChatLanguageModel chatLanguageModel() {
        return OpenAiChatModel.builder()
                .apiKey(System.getenv("API-KEY")) //配置在环境变变量
                .modelName("model-name")
                .baseUrl("base-url")
                .logRequests(true)
                .logResponses(true)
                .maxRetries(3) // 默认三次
                .maxTokens(500)// 限制响应长度
                .timeout(Duration.ofSeconds(15))
                .temperature(0.7)         // 控制生成随机性
                .build();
    }
}

控制层

@RestController
@Slf4j
public class ChatLanguageModelController {
    @Resource
    private ChatLanguageModel chatLanguageModel;

    @GetMapping("/v1/prompt")
    public String generateV1(@RequestParam(value = "prompt", defaultValue = "你是谁？") String prompt){
        String result = chatLanguageModel.generate(prompt);
        return result;
    }

    @GetMapping("/v1/userMessage")
    public String generateViaUseMessageV1(@RequestParam(value = "prompt")String prompt){
        UserMessage userMessage = UserMessage.from(prompt);
        Response<AiMessage> result = chatLanguageModel.generate(userMessage);
        return result.toString();
    }
}

测试结果

测试地址： http://localhost/v1/userMessage
返回结果

Response { 
content = AiMessage { text = "xxxxxxx" toolExecutionRequests = null },
tokenUsage = TokenUsage { inputTokenCount = 11, outputTokenCount = 17,totalTokenCount = 28 },
finishReason = STOP,
metadata = {} 
}

注意：tokenUsage 计算消耗的token，包含input和output的

其他参数

参考链接： https://github.com/langchain4j/langchain4j-spring/blob/main/langchain4j-open-ai-spring-boot-starter/src/main/java/dev/langchain4j/openai/spring/AutoConfig.java
在这里插入图片描述