本文是 2026面试六大前沿热点速查 的分章阅读版。
答案要点:
ChatClient:统一调用GPT/Claude/Qwen/DeepSeek (类似JDBC)EmbeddingModel:向量模型统一接口VectorStore:向量数据库统一接口 (Chroma/Milvus/PgVector/Redis)FunctionCallback:工具调用 (类似Function Calling)答案要点:
@Service
public class RagService {
private final ChatClient chatClient;
private final VectorStore vectorStore;
public RagService(ChatClient.Builder builder, VectorStore vectorStore) {
this.chatClient = builder
.defaultSystem("基于提供的Context回答用户问题,不要编造")
.build();
this.vectorStore = vectorStore;
}
public String ask(String question) {
// 1. 检索相关文档
List<Document> docs = vectorStore.similaritySearch(
SearchRequest.query(question).withTopK(5)
);
// 2. 拼接Context
String context = docs.stream()
.map(Document::getText)
.collect(Collectors.joining("\n\n"));
// 3. 调用LLM
return chatClient.prompt()
.user(u -> u.text("Context:\n{context}\n\nQuestion: {question}")
.param("context", context)
.param("question", question))
.call()
.content();
}
}
// 配置 (application.yml)
spring:
ai:
openai:
api-key: ${OPENAI_API_KEY}
chat.options.model: gpt-4o
vectorstore:
pgvector:
dimensions: 1536
distance-type: cosine-distance
index-type: hnsw
答案要点:
@Component
public class WeatherTools {
@Bean
@Description("获取指定城市的天气")
public Function<WeatherRequest, WeatherResponse> getWeather() {
return request -> weatherApi.query(request.getCity());
}
}
// 调用
String response = chatClient.prompt()
.user("深圳今天天气怎么样?")
.functions("getWeather") // 自动调用
.call()
.content();
追问:与原生Function Calling区别?
答案要点:
@Configuration
public class McpConfig {
@Bean
public McpSyncClient githubMcpClient() {
return McpClient.sync()
.stdio("npx -y @modelcontextprotocol/server-github")
.build();
}
}
// 使用MCP工具
@Autowired
private McpSyncClient mcpClient;
public String analyzeRepo() {
List<Tool> tools = mcpClient.listTools();
CallToolResult result = mcpClient.callTool("search_repos",
Map.of("query", "spring-ai"));
// 把工具能力暴露给ChatClient
return chatClient.prompt()
.tools(tools)
.user("搜索SpringAI仓库并分析star数")
.call()
.content();
}
答案要点:
spring:
ai:
openai:
chat:
options:
model: gpt-4o-mini
temperature: 0.7
max-tokens: 2000
retry:
max-attempts: 3
backoff: exponential
# 速率限制 (Resilience4j)
resilience4j.ratelimiter:
instances:
llm-api:
limit-for-period: 100
limit-refresh-period: 1m