Springboot 如何调用本地部署Ollama中的多个模型

下面我将介绍如何在Spring Boot应用中调用本地部署的Ollama中的多个模型，包括设置依赖、配置和调用代码。

1、添加依赖

首先，在 pom.xml 中添加必要的依赖：

<dependencies>
    <!-- Spring Boot Web -->
    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-web</artifactId>
    </dependency>
    
    <!-- HTTP Client -->
    <dependency>
        <groupId>org.apache.httpcomponents</groupId>
        <artifactId>httpclient</artifactId>
        <version>4.5.13</version>
    </dependency>
    
    <!-- JSON Processing -->
    <dependency>
        <groupId>com.fasterxml.jackson.core</groupId>
        <artifactId>jackson-databind</artifactId>
    </dependency>
    
    <!-- Lombok (可选) -->
    <dependency>
        <groupId>org.projectlombok</groupId>
        <artifactId>lombok</artifactId>
        <optional>true</optional>
    </dependency>
</dependencies>

2. 配置Ollama连接

在application.properties或application.yml中添加Ollama配置：

# application.properties
ollama.base-url=http://localhost:11434
ollama.models=llama2,gemma,mistral
ollama.timeout=30000

或者YAML格式：

# application.yml
ollama:
  base-url: http://localhost:11434
  models: llama2,gemma,mistral
  timeout: 30000

3. 创建配置类

import org.springframework.beans.factory.annotation.Value;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;
import org.apache.http.impl.client.CloseableHttpClient;
import org.apache.http.impl.client.HttpClients;

@Configuration
public class OllamaConfig {
    
    @Value("${ollama.base-url}")
    private String baseUrl;
    
    @Value("${ollama.models}")
    private String[] models;
    
    @Value("${ollama.timeout}")
    private int timeout;
    
    @Bean
    public CloseableHttpClient httpClient() {
        return HttpClients.createDefault();
    }
    
    // Getter methods
    public String getBaseUrl() {
        return baseUrl;
    }
    
    public String[] getModels() {
        return models;
    }
    
    public int getTimeout() {
        return timeout;
    }
}

4. 创建请求和响应DTO

import lombok.Data;
@Data
public class OllamaRequest {
    private String model;
    private String prompt;
    private boolean stream = false;
    // 可以根据需要添加其他参数
}

@Data
public class OllamaResponse {
    private String model;
    private String response;
    // 可以根据实际响应结构添加更多字段
}

5. 创建服务类调用Ollama API

import org.apache.http.client.methods.CloseableHttpResponse;
import org.apache.http.client.methods.HttpPost;
import org.apache.http.entity.StringEntity;
import org.apache.http.impl.client.CloseableHttpClient;
import org.apache.http.util.EntityUtils;
import com.fasterxml.jackson.databind.ObjectMapper;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Service;
import java.io.IOException;
import java.util.HashMap;
import java.util.Map;

@Service
public class OllamaService {
    
    private final CloseableHttpClient httpClient;
    private final OllamaConfig ollamaConfig;
    private final ObjectMapper objectMapper;
    
    @Autowired
    public OllamaService(CloseableHttpClient httpClient, OllamaConfig ollamaConfig, ObjectMapper objectMapper) {
        this.httpClient = httpClient;
        this.ollamaConfig = ollamaConfig;
        this.objectMapper = objectMapper;
    }
    
    public String generateText(String model, String prompt) throws IOException {
        String url = ollamaConfig.getBaseUrl() + "/api/generate";
        
        OllamaRequest request = new OllamaRequest();
        request.setModel(model);
        request.setPrompt(prompt);
        
        HttpPost httpPost = new HttpPost(url);
        httpPost.setHeader("Content-Type", "application/json");
        httpPost.setEntity(new StringEntity(objectMapper.writeValueAsString(request)));
        
        try (CloseableHttpResponse response = httpClient.execute(httpPost)) {
            String responseBody = EntityUtils.toString(response.getEntity());
            OllamaResponse ollamaResponse = objectMapper.readValue(responseBody, OllamaResponse.class);
            return ollamaResponse.getResponse();
        }
    }
    
    public Map<String, String> generateWithAllModels(String prompt) {
        Map<String, String> responses = new HashMap<>();
        
        for (String model : ollamaConfig.getModels()) {
            try {
                String response = generateText(model, prompt);
                responses.put(model, response);
            } catch (IOException e) {
                responses.put(model, "Error calling model: " + e.getMessage());
            }
        }
        
        return responses;
    }
}

6. 创建控制器

import org.springframework.web.bind.annotation.*;
import java.util.Map;

@RestController
@RequestMapping("/api/ollama")
public class OllamaController {
    
    private final OllamaService ollamaService;
    
    public OllamaController(OllamaService ollamaService) {
        this.ollamaService = ollamaService;
    }
    
    @PostMapping("/generate")
    public String generateText(@RequestParam String model, @RequestParam String prompt) {
        try {
            return ollamaService.generateText(model, prompt);
        } catch (Exception e) {
            return "Error: " + e.getMessage();
        }
    }
    
    @PostMapping("/generate-all")
    public Map<String, String> generateWithAllModels(@RequestParam String prompt) {
        return ollamaService.generateWithAllModels(prompt);
    }
}

7. 使用示例

启动Spring Boot应用后，你可以通过以下方式调用：

（1）调用单个模型：

POST /api/ollama/generate?model=llama2&prompt=Tell me about AI

（2）调用所有配置的模型：

POST /api/ollama/generate-all?prompt=Explain quantum computing

8. 高级功能扩展

如果需要更复杂的功能，可以考虑：

（1）添加模型切换策略

（2）实现流式响应

（3）添加模型健康检查

（4）实现模型性能监控

例如，添加模型健康检查：

public boolean isModelAvailable(String model) throws IOException {
    String url = ollamaConfig.getBaseUrl() + "/api/tags";
    HttpGet httpGet = new HttpGet(url);
    
    try (CloseableHttpResponse response = httpClient.execute(httpGet)) {
        String responseBody = EntityUtils.toString(response.getEntity());
        // 解析响应检查模型是否存在
        return responseBody.contains(model);
    }
}

注意事项

1、确保本地Ollama服务已启动并运行在默认端口(11434)

2、根据实际Ollama API版本调整请求和响应结构

3、在生产环境中添加适当的错误处理和日志记录

4、考虑添加请求超时和重试机制

 以上代码提供了一个完整的Spring Boot集成Ollama多个模型的实现方案，大家可以根据实际需求进行调整和扩展。