Implement error handling for model_tuning_core()

[ETA444]

Error Handling in model_tuning_core()

Type Validations

• Data Type Validation
  • Ensures x_train and y_train are either pandas DataFrames or NumPy ndarrays. This check ensures that the data inputs are compatible with the models being tuned.

if not isinstance(x_train, (pd.DataFrame, np.ndarray)):
    raise TypeError("model_tuning_core(): 'x_train' must be a pandas DataFrame or NumPy ndarray.")
if not isinstance(y_train, (pd.Series, np.ndarray)):
    raise TypeError("model_tuning_core(): 'y_train' must be a pandas Series or NumPy ndarray.")

• Parameter Type Checks
  • Validates types for all parameters to ensure they match expected types, preventing errors during processing. For example, ensuring task_type is a string, models is a dictionary, and integers are used where required.

if not isinstance(task_type, str):
    raise TypeError("model_tuning_core(): 'task_type' must be a string.")
if not isinstance(models, dict):
    raise TypeError("model_tuning_core(): 'models' must be a dictionary with model names as keys and model instances as values.")
if priority_metrics is not None and not isinstance(priority_metrics, list):
    raise TypeError("model_tuning_core(): 'priority_metrics' must be a list of strings.")
if refit_metric is not None and not isinstance(refit_metric, (str, Callable)):
    raise TypeError("model_tuning_core(): 'refit_metric' must be either a string or a callable object.")

Value Validations

• Logical Consistency Checks
  • Checks for logical consistency, such as ensuring task_type is one of the allowed types ('classification' or 'regression') and the n_top_models parameter is greater than zero and less than the total available models.

if task_type not in ['classification', 'regression']:
    raise ValueError("model_tuning_core(): 'task_type' must be either 'classification' or 'regression'.")
if n_top_models <= 0:
    raise ValueError("model_tuning_core(): 'n_top_models' must be an integer greater than 0.")

• Data Integrity Checks
  • Ensures that x_train and y_train have the same length and are not empty, which are fundamental requirements for fitting models.

if x_train.shape[0] != y_train.shape[0]:
    raise ValueError("model_tuning_core(): 'x_train' and 'y_train' must have the same number of rows.")
if x_train.size == 0 or y_train.size == 0:
    raise ValueError("model_tuning_core(): 'x_train' and 'y_train' cannot be empty.")

• Specific Parameter Checks
  • Ensures all elements in priority_metrics and priority_tuners are strings and that the metrics and tuners specified are valid. Also checks the range of cv, n_iter_random, and n_iter_bayesian.

if any(not isinstance(metric, str) for metric in priority_metrics):
    raise ValueError("model_tuning_core(): All elements in 'priority_metrics' must be strings.")
if priority_tuners and any(tuner not in ['grid', 'random', 'bayesian'] for tuner in priority_tuners):
    raise ValueError("model_tuning_core(): 'priority_tuners' contains unrecognized tuners. Valid tuners are 'grid', 'random', 'bayesian'.")
if cv < 1:
    raise ValueError("model_tuning_core(): 'cv' must be an integer greater than 0.")
if n_iter_random is not None and n_iter_random < 1:
    raise ValueError("model_tuning_core(): 'n_iter_random' must be a non-negative integer greater than 0.")
if n_iter_bayesian is not None and n_iter_bayesian < 1:
    raise ValueError("model_tuning_core(): 'n_iter_bayesian' must be a non-negative integer greater than 0.")