RAG MCP Tool

import yaml import os from pydantic import BaseModel, Field from typing import Optional class LLMConfig(BaseModel): service_type: str = Field(default="openai", description="Service type: openai, local, etc.") base_url: str = Field(default="http://localhost:1234/v1", description="LLM API Base URL") api_key: Optional[str] = Field(default=None, description="API Key") timeout: int = Field(default=60, description="Request timeout in seconds") class ModelConfig(BaseModel): name: str = Field(default="text-embedding-qwen3-embedding-4b", description="Model name") context_window: int = Field(default=4096, description="Context window size") temperature: float = Field(default=0.7, description="Generation temperature") class ProcessingConfig(BaseModel): chunk_count: int = Field(default=5, description="Number of chunks to split the file into") class AppConfig(BaseModel): llm: LLMConfig = Field(default_factory=LLMConfig) model: ModelConfig = Field(default_factory=ModelConfig) processing: ProcessingConfig = Field(default_factory=ProcessingConfig) def load_config(config_path: str) -> AppConfig: if not os.path.exists(config_path): # If config file doesn't exist, return default return AppConfig() with open(config_path, 'r', encoding='utf-8') as f: data = yaml.safe_load(f) if not data: return AppConfig() return AppConfig(**data)