Spring AI 实现 RAG 检索增强生成详解与实践

2. 配置向量存储和模型

在 application.yml 中配置：

spring:
  ai:
    openai:
      api-key: ${OPENAI_API_KEY}
      chat:
        options:
          model: gpt-4
          temperature: 0.7
    vectorstore:
      pgvector:
        index-type: HNSW
        distance-type: COSINE_DISTANCE
        dimensions: 1536

3. 创建向量存储服务

package com.example.rag.service;

import org.springframework.ai.document.Document;
import org.springframework.ai.vectorstore.VectorStore;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Service;

import java.util.List;

@Service
public class VectorStoreService {
    @Autowired
    private VectorStore vectorStore;

    /**
     * 添加文档到向量存储
     */
    public void addDocuments(List<Document> documents) {
        vectorStore.add(documents);
    }

    /**
     * 根据查询检索相似文档
     */
    public List<Document> search(String query, int topK) {
        return vectorStore.similaritySearch(query, topK);
    }
}

4. 实现 RAG 服务

package com.example.rag.service;

import org.springframework.ai.chat.client.ChatClient;
import org.springframework.ai.chat.messages.UserMessage;
import org.springframework.ai.document.Document;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Service;

import java.util.List;

@Service
public class RagService {
    @Autowired
    private ChatClient chatClient;

    @Autowired
    private VectorStoreService vectorStoreService;

    /**
     * 执行 RAG 查询
     */
    public String ragQuery(String query) {
        // 检索相关文档
        List<Document> relevantDocs = vectorStoreService.search(query, 3);
        
        // 构建上下文
        StringBuilder contextBuilder = new StringBuilder();
        for (Document doc : relevantDocs) {
            contextBuilder.append(doc.getContent()).append("\n");
        }

        // 生成回答
        UserMessage userMessage = new UserMessage(
            "基于以下上下文回答问题：" + contextBuilder.toString() + "\n问题：" + query
        );
        
        return chatClient.prompt().user(userMessage).call().content();
    }
}