xieyinlu/ruoyi-ai
1
0
Fork 0
mirror of https://gitcode.com/ageerle/ruoyi-ai.git synced 2026-04-13 03:45:14 +00:00
Code Issues Packages Projects Releases Wiki Activity

修改数据库读取工具

Browse Source
This commit is contained in:
zhang
2026-02-24 16:07:18 +08:00
parent f25ebdf9ec
commit 26bcfbba8a
9 changed files with 214 additions and 211 deletions
Download Patch File Download Diff File Expand all files Collapse all files

75
ruoyi-modules/ruoyi-chat/src/main/java/org/ruoyi/service/chat/impl/AbstractStreamingChatService.java
View File

@@ -3,6 +3,7 @@ package org.ruoyi.service.chat.impl;
import dev.langchain4j.agentic.AgenticServices;
import dev.langchain4j.agentic.supervisor.SupervisorAgent;
import dev.langchain4j.agentic.supervisor.SupervisorResponseStrategy;
import dev.langchain4j.community.model.dashscope.QwenChatModel;
import dev.langchain4j.data.message.ChatMessage;
import dev.langchain4j.data.message.UserMessage;
import dev.langchain4j.mcp.McpToolProvider;
@@ -325,26 +326,26 @@ public abstract class AbstractStreamingChatService implements IChatService {
protected String doAgent(String userMessage, ChatModelVo chatModelVo) {
// 步骤1: 配置MCP传输层 - 连接到bing-cn-mcp服务器
// 该服务提供两个工具: bing_search (必应搜索) 和 crawl_webpage (网页抓取)
McpTransport transport = new StdioMcpTransport.Builder()
.command(List.of("C:\\Program Files\\nodejs\\npx.cmd", "-y",
"bing-cn-mcp"
))
.logEvents(true)
.build();
// McpTransport transport = new StdioMcpTransport.Builder()
// .command(List.of("C:\\Program Files\\nodejs\\npx.cmd", "-y",
// "bing-cn-mcp"
// ))
// .logEvents(true)
// .build();
// 步骤2: 创建MCP客户端
McpClient mcpClient = new DefaultMcpClient.Builder()
.transport(transport)
.build();
// // 步骤2: 创建MCP客户端
// McpClient mcpClient = new DefaultMcpClient.Builder()
// .transport(transport)
// .build();
// 步骤3: 配置工具提供者
ToolProvider toolProvider = McpToolProvider.builder()
.mcpClients(List.of(mcpClient))
.build();
// // 步骤3: 配置工具提供者
// ToolProvider toolProvider = McpToolProvider.builder()
// .mcpClients(List.of(mcpClient))
// .build();
McpTransport transport1 = new StdioMcpTransport.Builder()
.command(List.of("C:\\Program Files\\nodejs\\npx.cmd", "-y",
.command(List.of("npx", "-y",
"mcp-echarts"
))
.logEvents(true)
@@ -361,40 +362,52 @@ public abstract class AbstractStreamingChatService implements IChatService {
.build();
// 步骤4: 配置OpenAI模型
OpenAiChatModel PLANNER_MODEL = OpenAiChatModel.builder()
.baseUrl(chatModelVo.getApiHost())
// OpenAiChatModel PLANNER_MODEL = OpenAiChatModel.builder()
// .baseUrl(chatModelVo.getApiHost())
// .apiKey(chatModelVo.getApiKey())
// .modelName(chatModelVo.getModelName())
// .build();
QwenChatModel qwenChatModel = QwenChatModel.builder()
// .baseUrl(chatModelVo.getApiHost())
.apiKey(chatModelVo.getApiKey())
.modelName(chatModelVo.getModelName())
.build();
.build();
SqlAgent sqlAgent = AgenticServices.agentBuilder(SqlAgent.class)
.chatModel(PLANNER_MODEL)
.chatModel(
qwenChatModel)
.tools(
new QueryAllTablesTool(),
new QueryTableSchemaTool(),
new ExecuteSqlQueryTool()
SpringUtils.getBean(QueryAllTablesTool.class), // 必须通过 getBean 获取
SpringUtils.getBean(QueryTableSchemaTool.class),
SpringUtils.getBean(ExecuteSqlQueryTool.class)
)
.build();
WebSearchAgent searchAgent = AgenticServices.agentBuilder(WebSearchAgent.class)
.chatModel(PLANNER_MODEL)
.toolProvider(toolProvider)
.build();
// WebSearchAgent searchAgent = AgenticServices.agentBuilder(WebSearchAgent.class)
// .chatModel(PLANNER_MODEL)
// .toolProvider(toolProvider)
// .build();
ChartGenerationAgent chartGenerationAgent = AgenticServices.agentBuilder(ChartGenerationAgent.class)
.chatModel(PLANNER_MODEL)
.chatModel(
qwenChatModel)
.toolProvider(toolProvider1)
.build();
String res = sqlAgent.getData(userMessage);
String res1 = chartGenerationAgent.generateChart(res);
System.out.println(res1);
System.out.println(res);
SupervisorAgent supervisor = AgenticServices
.supervisorBuilder()
.chatModel(PLANNER_MODEL)
.chatModel(qwenChatModel)
.subAgents(sqlAgent, chartGenerationAgent)
.responseStrategy(SupervisorResponseStrategy.LAST)
.build();
String invoke = supervisor.invoke(userMessage);
System.out.println(invoke);
return invoke;
return res1;
}
}