标签Ollama、Qwen3、SpringAI Alibaba、ChatClient、流式响应、Java AI一、前言上一篇介绍了阿里云百炼云端模型和 OpenAI 接口规范。本篇进入实操环节使用 Ollama 在本地部署 Qwen3 大模型在 Spring Boot 中集成 SpringAI Alibaba同时配置云端DashScope和本地Ollama双模型编写同步对话与流式对话接口二、Ollama本地大模型管理神器2.1 什么是 OllamaOllama 是一个用于在本地运行、管理和发布大模型的开源框架。它让下载和运行 Qwen 等模型变得像docker run一样简单。2.2 安装 Ollama访问官网https://ollama.com下载对应操作系统的安装包一键安装2.3 验证安装打开终端执行ollama--version2.4 拉取并运行 Qwen3ollama run qwen3:1.7b模型大小约 1.4GB下载完成后会自动运行模型服务服务地址http://localhost:11434也可以打开 Ollama 客户端界面搜索模型后直接下载启动。2.5 测试本地模型方式一交互界面在 ollama run 的交互界面中直接输入问题查看回复。方式二浏览器访问http://localhost:11434/api/tags能看到已拉取的模型列表说明本地服务启动成功。恭喜你拥有了一个完全在本地运行、可对话的大模型三、SpringAI Alibaba 简介3.1 什么是 SpringAI AlibabaSpringAI Alibaba 是 Spring AI 项目的一个实现由阿里云贡献深度集成阿里云百炼DashScope支持 OpenAI 兼容接口提供统一的 ChatClient、AiStreamClient 等接口通过配置即可切换模型极大降低集成复杂度官方文档https://java2ai.com3.2 SpringAI 与 SpringAI Alibaba 的区别特性SpringAISpringAI Alibaba定位国外 AI 接口标准国内环境优化版本国内模型兼容性百炼等国内模型不完全兼容阿里云百炼全系列完美兼容使用体验访问国内模型较麻烦方便快捷开箱即用四、Spring Boot 项目配置双模型4.1 版本对应关系SpringAI AlibabaSpring AISpring Boot1.0.0.21.0.03.4.54.2 Maven 依赖propertiesjava.version21/java.versionspring-ai.version1.0.0/spring-ai.versionspring-ai-alibaba.version1.0.0.2/spring-ai-alibaba.versionspring-boot.version3.4.5/spring-boot.version/propertiesdependencyManagementdependenciesdependencygroupIdcom.alibaba.cloud.ai/groupIdartifactIdspring-ai-alibaba-bom/artifactIdversion${spring-ai-alibaba.version}/versiontypepom/typescopeimport/scope/dependencydependencygroupIdorg.springframework.boot/groupIdartifactIdspring-boot-dependencies/artifactIdversion${spring-boot.version}/versiontypepom/typescopeimport/scope/dependencydependencygroupIdorg.springframework.ai/groupIdartifactIdspring-ai-bom/artifactIdversion${spring-ai.version}/versiontypepom/typescopeimport/scope/dependency/dependencies/dependencyManagementdependenciesdependencygroupIdorg.springframework.boot/groupIdartifactIdspring-boot-starter-web/artifactId/dependencydependencygroupIdcom.alibaba.cloud.ai/groupIdartifactIdspring-ai-alibaba-starter-dashscope/artifactId/dependencydependencygroupIdorg.springframework.ai/groupIdartifactIdspring-ai-starter-model-ollama/artifactId/dependencydependencygroupIdcom.alibaba.cloud.ai/groupIdartifactIdspring-ai-alibaba-starter-memory-redis/artifactId/dependencydependencygroupIdredis.clients/groupIdartifactIdjedis/artifactIdversion5.2.0/version/dependency/dependencies4.3 application.yml 配置spring:application:name:SpringAIAlibaba-RAG-Milvusai:dashscope:api-key:${dashscope.api.key}# 请替换为自己的 API Keychat:model:qwen-plusoptions:temperature:0.7ollama:enabled:truebase-url:http://localhost:11434chat:model:qwen3:1.7boptions:temperature:0.7server:port:80804.4 配置 ChatClientpackagenet.xdclass.config;importorg.springframework.ai.chat.client.ChatClient;importorg.springframework.ai.chat.model.ChatModel;importorg.springframework.beans.factory.annotation.Qualifier;importorg.springframework.context.annotation.Bean;importorg.springframework.context.annotation.Configuration;ConfigurationpublicclassChatConfig{BeanpublicChatClientdashscopeChatClient(Qualifier(dashscopeChatModel)ChatModeldashscopeChatModel){returnChatClient.builder(dashscopeChatModel).build();}BeanpublicChatClientollamaChatClient(Qualifier(ollamaChatModel)ChatModelollamaChatModel){returnChatClient.builder(ollamaChatModel).build();}}五、编写同步 AI 对话接口5.1 控制器代码packagenet.xdclass.controller;importlombok.RequiredArgsConstructor;importorg.springframework.ai.chat.client.ChatClient;importorg.springframework.web.bind.annotation.*;RestControllerRequestMapping(/api/ai)RequiredArgsConstructorpublicclassSimpleAiController{privatefinalChatClientdashscopeChatClient;privatefinalChatClientollamaChatClient;/** * 同步响应 - 云端模型 */GetMapping(value/simple/chat,producestext/html;charsetutf-8)publicStringsimpleChat(RequestParam(question)Stringquestion){returndashscopeChatClient.prompt(question).call().content();}/** * 同步响应 - 本地 Ollama 模型 */GetMapping(value/ollamaSimpleChat/chat,producestext/html;charsetutf-8)publicStringollamaSimpleChat(RequestParam(question)Stringquestion){returnollamaChatClient.prompt(question).call().content();}}5.2 添加日志 AdvisorBeanpublicChatClientdashscopeChatClient(ChatModeldashscopeChatModel){returnChatClient.builder(dashscopeChatModel).defaultAdvisors(newSimpleLoggerAdvisor()).build();}并在 application.yml 中开启 debug 日志logging:level:org.springframework.ai:debug访问http://localhost:8080/api/ai/simple/chat?question你是谁六、流式 AI 对话接口流式响应可以让 AI一个字一个字地输出体验更接近 ChatGPT。GetMapping(value/stream/chat,producestext/html;charsetutf-8)publicFluxStringstreamChat(RequestParam(question)Stringquestion){returndashscopeChatClient.prompt(question).stream().content();}GetMapping(value/ollamaStreamChat/chat,producestext/html;charsetutf-8)publicFluxStringollamaStreamChat(RequestParam(question)Stringquestion){returnollamaChatClient.prompt(question).stream().content();}七、小结本篇完成了Ollama 本地部署 Qwen3SpringAI Alibaba 项目搭建云端 本地双模型配置同步对话与流式对话接口实现下一篇将介绍提示词工程、系统提示词配置以及使用 Redis 让 AI 记住上下文。
