main.py

"""
Main Entry Point for ML Hyperparameter Tuning Project
Orchestrates all model training and tuning workflows.
"""

import os
import sys
import argparse


def main(data_source='synthetic', csv_path=None, target_column=None):
    """Main execution function.
    
    Args:
        data_source: Source of data ('synthetic', 'csv', 'breast_cancer', 'iris', 'wine')
        csv_path: Path to CSV file (required if data_source='csv')
        target_column: Target column name in CSV (optional, uses last column if not specified)
    """
    print("=" * 70)
    print("ML HYPERPARAMETER TUNING EVALUATION PROJECT")
    print("=" * 70)
    print(f"Data Source: {data_source}")
    if csv_path:
        print(f"CSV Path: {csv_path}")
    print()
    
    # Create results directory
    os.makedirs('results', exist_ok=True)
    
    # Import modules
    from src.utils.data_loader import DataLoader
    from src.models.baseline_model import BaselineModel
    from src.tuning.grid_search_tuning import GridSearchTuner
    from src.tuning.random_search_tuning import RandomSearchTuner
    from src.tuning.bayesian_tuning import BayesianTuner
    from src.tuning.manual_tuning import ManualTuner
    from src.utils.evaluate_models import print_comparison_table, analyze_tradeoffs, load_results
    
    # Load and prepare data
    print("\n[1/7] Loading and preparing data...")
    print("-" * 70)
    loader = DataLoader()
    
    # Load data based on source
    if data_source == 'csv':
        if csv_path is None:
            print("ERROR: csv_path required for CSV data source")
            sys.exit(1)
        X_train, X_test, y_train, y_test = loader.load_data(
            source='csv',
            filepath=csv_path,
            target_column=target_column
        )
    else:
        X_train, X_test, y_train, y_test = loader.load_data(source=data_source)
    
    loader.get_data_info()
    
    # Baseline model
    print("\n[2/7] Training baseline model...")
    print("-" * 70)
    baseline = BaselineModel()
    baseline.create_baseline_model()
    baseline.train(X_train, y_train)
    baseline.evaluate(X_test, y_test)
    baseline.save_results()
    
    # Grid Search
    print("\n[3/7] Running Grid Search tuning...")
    print("-" * 70)
    try:
        grid_tuner = GridSearchTuner()
        grid_tuner.create_param_grid()
        grid_tuner.run_grid_search(X_train, y_train)
        grid_acc = grid_tuner.evaluate(X_test, y_test)
        if grid_acc:
            grid_tuner.save_results(grid_acc)
    except Exception as e:
        print(f"Grid Search failed: {e}")
    
    # Random Search
    print("\n[4/7] Running Random Search tuning...")
    print("-" * 70)
    try:
        random_tuner = RandomSearchTuner()
        random_tuner.create_param_distribution()
        random_tuner.run_random_search(X_train, y_train)
        random_acc = random_tuner.evaluate(X_test, y_test)
        if random_acc:
            random_tuner.save_results(random_acc)
    except Exception as e:
        print(f"Random Search failed: {e}")
    
    # Bayesian Optimization
    print("\n[5/7] Running Bayesian Optimization tuning...")
    print("-" * 70)
    try:
        bayes_tuner = BayesianTuner()
        bayes_tuner.create_search_spaces()
        bayes_tuner.run_bayesian_search(X_train, y_train)
        bayes_acc = bayes_tuner.evaluate(X_test, y_test)
        if bayes_acc:
            bayes_tuner.save_results(bayes_acc)
    except Exception as e:
        print(f"Bayesian Optimization failed: {e}")
    
    # Manual Tuning
    print("\n[6/7] Running manual parameter exploration...")
    print("-" * 70)
    try:
        manual_tuner = ManualTuner()
        manual_tuner.tune_n_estimators(X_train, y_train, X_test, y_test)
        manual_tuner.tune_max_depth(X_train, y_train, X_test, y_test)
        manual_tuner.tune_min_samples_split(X_train, y_train, X_test, y_test)
        
        print("\n" + "=" * 50)
        print("Manual Tuning Summary")
        print("=" * 50)
        manual_tuner.get_best_params()
        manual_tuner.save_results()
    except Exception as e:
        print(f"Manual tuning failed: {e}")
    
    # Comparison and Analysis
    print("\n[7/7] Generating comparison report...")
    print("-" * 70)
    results = load_results()
    
    if results.get('baseline') is not None:
        print_comparison_table(results)
        analyze_tradeoffs(results)
        print("\n✓ Project execution completed successfully!")
        print(f"✓ Results saved to 'results/' directory")
    else:
        print("Could not generate comparison - baseline results not found")


if __name__ == "__main__":
    parser = argparse.ArgumentParser(
        description='ML Hyperparameter Tuning Evaluation',
        formatter_class=argparse.RawDescriptionHelpFormatter,
        epilog="""
Examples:
  python main.py                                    # Use synthetic data
  python main.py --source breast_cancer            # Use breast_cancer dataset
  python main.py --source iris                     # Use iris dataset
  python main.py --source csv --csv-path data.csv  # Load from CSV
  python main.py --source csv --csv-path data.csv --target churn  # CSV with target column
        """
    )
    
    parser.add_argument(
        '--source',
        type=str,
        default='synthetic',
        choices=['synthetic', 'breast_cancer', 'iris', 'wine', 'csv'],
        help='Data source (default: synthetic)'
    )
    parser.add_argument(
        '--csv-path',
        type=str,
        help='Path to CSV file (required if --source csv)'
    )
    parser.add_argument(
        '--target',
        type=str,
        help='Target column name in CSV (optional, uses last column if not specified)'
    )
    
    args = parser.parse_args()
    
    try:
        main(
            data_source=args.source,
            csv_path=args.csv_path,
            target_column=args.target
        )
    except KeyboardInterrupt:
        print("\n\nExecution interrupted by user.")
        sys.exit(0)
    except Exception as e:
        print(f"\nFatal error: {e}")
        import traceback
        traceback.print_exc()
        sys.exit(1)
    
    # Baseline model
    print("\n[2/7] Training baseline model...")
    print("-" * 70)
    baseline = BaselineModel()
    baseline.create_baseline_model()
    baseline.train(X_train, y_train)
    baseline.evaluate(X_test, y_test)
    baseline.save_results()
    
    # Grid Search
    print("\n[3/7] Running Grid Search tuning...")
    print("-" * 70)
    try:
        grid_tuner = GridSearchTuner()
        grid_tuner.create_param_grid()
        grid_tuner.run_grid_search(X_train, y_train)
        grid_acc = grid_tuner.evaluate(X_test, y_test)
        if grid_acc:
            grid_tuner.save_results(grid_acc)
    except Exception as e:
        print(f"Grid Search failed: {e}")
    
    # Random Search
    print("\n[4/7] Running Random Search tuning...")
    print("-" * 70)
    try:
        random_tuner = RandomSearchTuner()
        random_tuner.create_param_distribution()
        random_tuner.run_random_search(X_train, y_train)
        random_acc = random_tuner.evaluate(X_test, y_test)
        if random_acc:
            random_tuner.save_results(random_acc)
    except Exception as e:
        print(f"Random Search failed: {e}")
    
    # Bayesian Optimization
    print("\n[5/7] Running Bayesian Optimization tuning...")
    print("-" * 70)
    try:
        bayes_tuner = BayesianTuner()
        bayes_tuner.create_search_spaces()
        bayes_tuner.run_bayesian_search(X_train, y_train)
        bayes_acc = bayes_tuner.evaluate(X_test, y_test)
        if bayes_acc:
            bayes_tuner.save_results(bayes_acc)
    except Exception as e:
        print(f"Bayesian Optimization failed: {e}")
    
    # Manual Tuning
    print("\n[6/7] Running manual parameter exploration...")
    print("-" * 70)
    try:
        manual_tuner = ManualTuner()
        manual_tuner.tune_n_estimators(X_train, y_train, X_test, y_test)
        manual_tuner.tune_max_depth(X_train, y_train, X_test, y_test)
        manual_tuner.tune_min_samples_split(X_train, y_train, X_test, y_test)
        
        print("\n" + "=" * 50)
        print("Manual Tuning Summary")
        print("=" * 50)
        manual_tuner.get_best_params()
        manual_tuner.save_results()
    except Exception as e:
        print(f"Manual tuning failed: {e}")
    
    # Comparison and Analysis
    print("\n[7/7] Generating comparison report...")
    print("-" * 70)
    results = load_results()
    
    if results.get('baseline') is not None:
        print_comparison_table(results)
        analyze_tradeoffs(results)
        print("\n✓ Project execution completed successfully!")
        print(f"✓ Results saved to 'results/' directory")
    else:
        print("Could not generate comparison - baseline results not found")


if __name__ == "__main__":
    try:
        main()
    except KeyboardInterrupt:
        print("\n\nExecution interrupted by user.")
        sys.exit(0)
    except Exception as e:
        print(f"\nFatal error: {e}")
        import traceback
        traceback.print_exc()
        sys.exit(1)