Stata-MCP

#!/usr/bin/python3 # -*- coding: utf-8 -*- # # Copyright (C) 2025 - Present Sepine Tam, Inc. All Rights Reserved # # @Author : Sepine Tam (谭淞) # @Email : sepinetam@gmail.com # @File : langchain_agent.py from langchain.chat_models import init_chat_model from langchain_mcp_adapters.client import MultiServerMCPClient from langgraph.prebuilt import create_react_agent from .pmp.prompt_generator import PromptGenerator, TemplateName class StataAgent: # Set default MCP server mcp_client = MultiServerMCPClient( { "stata-mcp": { "command": "uv", "args": [ "--directory", "/Users/sepinetam/Documents/Github/stata-mcp", "run", 'main.py' ], "transport": "stdio", } } ) def __init__(self, model: str = "gpt-5", work_dir: str = None): """ Please set your OPENAI_API_KEY and OPENAI_BASE_URL (OPENAI_API_BASE) in your environment. Although OpenAI's ChatGPT is great enough, I recommend to use DeepSeek as its ability and score. If you want to know more about it, visit https://www.statamcp.com/reports to find more information. """ # Set model also could be changed self.model = init_chat_model(model=model, model_provider="openai") self.work_dir = work_dir # Manage prompt self.prompt_generator = PromptGenerator( template_name=TemplateName.ReAct, ROOT=self.work_dir, agent_provider="langchain" ) self.instructions = self.prompt_generator.instructions( root=self.work_dir ) async def set_agent(self): tools = await self.mcp_client.get_tools() agent = create_react_agent(self.model, tools, prompt=self.instructions) return agent async def run(self, data_source: str, task: str, is_display: bool = True): message = self.prompt_generator.tasks( datas=data_source, aims=task, root=self.work_dir ) task_message = {"messages": message} agent = await self.set_agent() result = await agent.ainvoke( task_message, config={"recursion_limit": 50} ) if is_display: print(result) return result