AI Tutoring RAG System

from mcp_host.schemas.chats import UploadMetadata from fastapi import Form from typing import Dict, Any, Optional # NEW: Import Optional def parse_upload_metadata( student_id: str = Form(...), subject: str = Form(...), topic: str = Form(...), difficulty_level: int = Form(...), document_title: Optional[str] = Form(None), # NEW: Document title parameter ) -> UploadMetadata: return UploadMetadata( student_id=student_id, subject=subject, topic=topic, difficulty_level=difficulty_level, document_title=document_title, # NEW: Include document title ) async def call_mcp_server_tool( sessions: dict, server_name: str, tool_name: str, tool_args: dict[str, Any] ) -> Any: """ Call an MCP server tool with preprocessing for compatibility. Automatically converts query arrays to strings for knowledge_base_retrieval to handle cases where AI agents send arrays instead of strings. """ # Preprocess arguments for knowledge_base_retrieval tool if tool_name == "knowledge_base_retrieval" and "query" in tool_args: query = tool_args["query"] # Convert array to comma-separated string if needed if isinstance(query, list): tool_args["query"] = ", ".join(query) print(f"⚠️ Preprocessed query array to string: {tool_args['query'][:100]}...") session = sessions[server_name]["session"] return await session.call_tool(tool_name, tool_args)