NCCN Guidelines MCP Server

Instructions

Implementation Reference

• server.py:146-166 (handler)

  The MCP tool handler for 'get_index'. This function reads and returns the raw YAML content of the NCCN guidelines index file generated by the scraper.
  @mcp.tool() async def get_index() -> str: """ Get the raw contents of the NCCN guidelines index YAML file. Returns: String containing the raw YAML content of the guidelines index """ try: index_path = current_dir / GUIDELINES_INDEX_FILE with open(index_path, 'r', encoding='utf-8') as f: content = f.read() logger.info(f"Successfully loaded guidelines index from {index_path}") return content except FileNotFoundError: logger.error(f"Guidelines index file not found: {index_path}") return "Error: Guidelines index file not found" except Exception as e: logger.error(f"Error reading guidelines index: {str(e)}") return f"Error reading guidelines index: {str(e)}"
• nccn_get_index.py:369-450 (helper)

  Helper function that scrapes the NCCN website to generate and maintain the guidelines index YAML file used by the get_index tool. Called during server initialization.
  async def ensure_nccn_index(output_file: str = DEFAULT_OUTPUT_FILE, max_age_days: int = CACHE_MAX_AGE_DAYS) -> dict: """ Ensure NCCN guideline index exists and is valid This is the main interface for MCP Server calls Args: output_file: Output file path max_age_days: Maximum cache file validity period (days) Returns: Parsed guideline index data """ import time # Check cache file cache_info = check_cache_file(output_file) # Determine if re-scraping is needed should_scrape = not cache_info['exists'] or not cache_info['is_valid'] if cache_info['exists']: if cache_info['is_valid']: logger.info(f"Using valid cache file: {output_file} (created at {cache_info['created_time'].strftime('%Y-%m-%d %H:%M:%S')}, {cache_info['age_days']} days ago)") else: logger.info(f"Cache file expired ({cache_info['age_days']} days > {max_age_days} days) or corrupted, starting re-scraping...") else: logger.info("Cache file not found, starting NCCN guideline index scraping...") if should_scrape: start_time = time.time() try: # Scrape all category data categories_data = await scrape_all_categories() if not categories_data: logger.error("Scraping failed, no data retrieved") # If scraping fails but old cache exists, try using old cache if cache_info['exists']: logger.info("Scraping failed, attempting to use existing cache file") return load_cached_data(output_file) return {} # Generate YAML document yaml_content = generate_yaml(categories_data) # Save to file with open(output_file, 'w', encoding='utf-8') as f: f.write(yaml_content) # Calculate statistics total_guidelines = sum(len(cat.get('items', [])) for cat in categories_data) successful_guidelines = sum( len([item for item in cat.get('items', []) if item.get('guideline_link')]) for cat in categories_data ) elapsed_time = time.time() - start_time logger.info(f"Scraping completed! Index saved to {output_file}") logger.info(f"Processed {len(categories_data)} categories, found {successful_guidelines}/{total_guidelines} valid guideline links") logger.info(f"Scraping time: {elapsed_time:.2f} seconds") except Exception as e: logger.error(f"Error during scraping process: {e}") # If scraping fails but cache exists, use cache if cache_info['exists']: logger.info("Scraping failed, using existing cache file") return load_cached_data(output_file) return {} # Load and return data cached_data = load_cached_data(output_file) if cached_data and 'nccn_guidelines' in cached_data: total_categories = len(cached_data['nccn_guidelines']) total_guidelines = sum(len(cat.get('guidelines', [])) for cat in cached_data['nccn_guidelines']) logger.info(f"NCCN guideline index ready: {total_categories} categories, {total_guidelines} total guidelines") else: logger.warning("Guideline index file format is abnormal") return cached_data
• server.py:146-146 (registration)

  The @mcp.tool() decorator registers the get_index function as an MCP tool.
  @mcp.tool()
• server.py:81-85 (helper)

  Server initialization call to ensure_nccn_index, which populates the index file before the tool is used.
  guidelines_data = await ensure_nccn_index( output_file=str(current_dir / GUIDELINES_INDEX_FILE), max_age_days=7 # Refresh index every 7 days )