SpringBoot 集成Ollama 本地大模型

本文主要简述springBoot如何结合spring ai集成本地大模型实现智能对话

1.引入依赖

<dependencies>
    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-web</artifactId>
    </dependency>
    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-test</artifactId>
        <scope>test</scope>
    </dependency>
    <!--ollama依赖，最新稳定版0.8.0-->
    <dependency>
        <groupId>org.springframework.ai</groupId>
        <artifactId>spring-ai-ollama</artifactId>
        <version>0.8.0</version>
    </dependency>
    <dependency>
        <groupId>org.springframework.ai</groupId>
        <artifactId>spring-ai-ollama-spring-boot-starter</artifactId>
        <version>0.8.0</version>
        <exclusions>
            <exclusion>
                <groupId>org.springframework.cloud</groupId>
                <artifactId>spring-cloud-function-context</artifactId>
            </exclusion>
        </exclusions>
    </dependency>
</dependencies>
<!--配置spring的仓库-->
<repositories>
    <repository>
        <id>spring-milestones</id>
        <url>https://repo.spring.io/milestone</url>
        <snapshots><enabled>false</enabled></snapshots>
    </repository>
</repositories>、
<!--spring ai的bom管理-->
<dependencyManagement>
    <dependencies>
        <dependency>
            <groupId>com.alibaba.cloud.ai</groupId>
            <artifactId>spring-ai-alibaba-bom</artifactId>
            <version>1.0.0.2</version>
            <type>pom</type>
            <scope>import</scope>
        </dependency>
    </dependencies>
</dependencyManagement>    

2.修改配置文件

spring.application.name=OllamaProject
spring.ai.ollama.base-url=http://localhost:11434  #ollama地址
spring.ai.ollama.chat.model=qwen3:4b  #大模型名称

3.编写Contoller

spring ai 中集成了OllamaChatClient，直接使用ollamaChatClient即可

import org.springframework.ai.ollama.OllamaChatClient;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.web.bind.annotation.GetMapping;
import org.springframework.web.bind.annotation.RequestMapping;
import org.springframework.web.bind.annotation.RequestParam;
import org.springframework.web.bind.annotation.RestController;

@RestController
@RequestMapping("/ai")
public class OllamaTestController {
    @Autowired
    private OllamaChatClient ollamaChatClient;
    @GetMapping("/chat")
    public String generate(@RequestParam String message) {
        return ollamaChatClient.call(message);
    }
    @GetMapping(value = "/chatStream",produces = MediaType.TEXT_MARKDOWN_VALUE + ";charset=UTF-8" )
    public Flux<String> chatStream(@RequestParam String message) {
        Flux<ChatResponse> stream = ollamaChatClient.stream(new Prompt(message));
        return stream.map(chatResponse -> chatResponse.getResult().getOutput().getContent()) ;
    }
    public Flux<String> chatStream(@RequestParam String message) {
        Flux<ChatResponse> stream = ollamaChatClient.stream(new Prompt(message));
        return stream.map(chatResponse -> chatResponse.getResult().getOutput().getContent()) ;
    }


}

4.修改启动类

package com.example.ollamaproject;

import org.springframework.boot.SpringApplication;
import org.springframework.boot.autoconfigure.SpringBootApplication;
import org.springframework.cloud.function.context.config.ContextFunctionCatalogAutoConfiguration;
@SpringBootApplication(exclude = {
        ContextFunctionCatalogAutoConfiguration.class
})
public class OllamaProjectApplication {

    public static void main(String[] args) {
        SpringApplication.run(OllamaProjectApplication.class, args);
    }

}

5.启动验证

调用chat接口，一次性返回

调用chatStream接口，会采用流式输出方式：

6.多模态大模型

如果要完成多模态的大模型集成，可以拉取

代码如下：

  @Test
    public void ollamaImageTest(){
        ClassPathResource resource = new ClassPathResource("/files/fengj.jpg");
        OllamaApi ollamaApi = new OllamaApi("http://localhost:11434");
        OllamaOptions ollamaOptions = OllamaOptions.create();
        ollamaOptions.setModel("gemma3:1b");
        ChatResponse response = new OllamaChatClient(ollamaApi)
                .withDefaultOptions(ollamaOptions)
                .call(new Prompt(new UserMessage(resource)));
        System.out.println(response.getResult().getOutput().getContent());


    }

fengj,jpg是一张风景照片

返回结果：

以上就是集成ollama的所有内容，使用时由于版本更新，语法上可能存在差异，仅供参考！