api/config.py

"""
API configuration module.
"""
import os
import logging
import yaml
from typing import Dict, Any, Optional
import base64

logger = logging.getLogger(__name__)

# Cache config for performance
_config_cache = None

# Store the list of supported base models
_supported_base_models = []

def get_api_config() -> Dict[str, Any]:
    """
    Returns the API configuration.
    
    Returns:
        Dict[str, Any]: API configuration
    """
    global _config_cache
    
    # Use cache if available
    if _config_cache is not None:
        return _config_cache
    
    # Try to load config from file
    config_path = os.environ.get("MEZURA_CONFIG_PATH", "config/api_config.yaml")
    if os.path.exists(config_path):
        try:
            with open(config_path, 'r') as f:
                config = yaml.safe_load(f)
                _config_cache = config
                logger.info(f"Loaded API configuration from {config_path}")
                return config
        except Exception as e:
            logger.error(f"Error loading config from {config_path}: {e}")
            raise RuntimeError(f"Failed to load configuration from {config_path}: {e}")
    else:
        # If config file not found, raise error
        error_msg = f"Configuration file not found at {config_path}"
        logger.error(error_msg)
        raise FileNotFoundError(error_msg)

def get_api_config_for_type(evaluation_type: str) -> Dict[str, Any]:
    """
    Get API configuration for a specific evaluation type.
    
    Args:
        evaluation_type: Evaluation type (e.g., "evalmix")
        
    Returns:
        Dict[str, Any]: API configuration for the specified type
    """
    config = get_api_config()
    
    # Convert evaluation type to config key
    api_type = evaluation_type.lower().replace("-", "_")
    
    # Get API configuration
    if "apis" in config and api_type in config["apis"]:
        type_config = config["apis"][api_type]
        logger.debug(f"Using config for {evaluation_type}: {type_config}")
        return type_config
    
    # Get default configuration if not found
    if "default" in config:
        logger.warning(f"No configuration found for {evaluation_type}, using default")
        return config["default"]
    
    # If no default config either, return empty dict
    logger.warning(f"No configuration found for {evaluation_type} and no default config")
    return {}

def get_airflow_config() -> Dict[str, Any]:
    """
    Get Airflow API configuration.
    
    Returns:
        Dict[str, Any]: Airflow API configuration
    """
    config = get_api_config()
    
    # Get Airflow config from the loaded yaml configuration
    if "apis" in config and "airflow" in config["apis"]:
        airflow_config = config["apis"]["airflow"]
        logger.debug(f"Using Airflow config from YAML: {airflow_config}")
        
        # --- Load base_url from environment if available ---
        env_base_url = os.environ.get("AIRFLOW_URL")
        if env_base_url:
            airflow_config["base_url"] = env_base_url
            logger.info(f"Loaded Airflow base_url from environment variable AIRFLOW_URL: {env_base_url}")
        else:
            logger.info(f"Using Airflow base_url from YAML config: {airflow_config.get('base_url')}")
        # --- END base_url env logic ---
        
        # Check if credentials should be loaded from environment
        auth_config = airflow_config.get("auth", {})
        if auth_config.get("use_env", False):
            # Get environment variable names
            username_env = auth_config.get("env_username", "MEZURA_API_USERNAME")
            password_env = auth_config.get("env_password", "MEZURA_API_PASSWORD")
            
            # Log environment variable names
            logger.info(f"Looking for credentials in environment variables: {username_env}, {password_env}")
            
            # Check if environment variables are set
            username = os.environ.get(username_env)
            password = os.environ.get(password_env)
            
            # SECURITY: Commented out to prevent potential credential exposure
            # Directly access environment variables for better logging
            # all_env_vars = os.environ.keys()
            # logger.info(f"Available environment variables: {', '.join(all_env_vars)}")
            
            # Log results of environment variable check
            logger.info(f"Username variable '{username_env}' found: {username is not None}")
            logger.info(f"Password variable '{password_env}' found: {password is not None}")
            
            # Update auth config with credentials from environment
            if username and password:
                auth_config["username"] = username
                auth_config["password"] = password
                # Update the auth config in airflow_config
                airflow_config["auth"] = auth_config
        
        return airflow_config
    
    # If not found in config, log warning and return empty dict
    logger.warning("Airflow configuration not found in config file")
    return {} 

def update_base_model_list(models):
    """
    Updates the list of supported base models.
    
    Args:
        models (list): List of supported model names
    """
    global _supported_base_models
    _supported_base_models = models

def get_base_model_list():
    """
    Returns the current list of supported base models.
    
    Returns:
        list: List of supported model names
    """
    return _supported_base_models