Refactor main.py for improved logging and modularity

[bhushankhopkarr]

Key optimizations:

1. Error Handling: Added error handling in critical parts like dataset loading and report generation to avoid abrupt crashes.
2. Function Decomposition: Split large blocks of logic into reusable functions (load_dataset, split_data, create_models, generate_predictions) for better readability and maintainability.
3. Logger Setup: Used logging.basicConfig for logger configuration to ensure consistent logging.
4. User Input Logic: Extracted the prediction loop into generate_predictions to reduce redundancy and improve clarity.
5. Code Structure: Moved the XAIWrapper initialization and fit logic into a more streamlined flow.

This approach makes the code easier to follow and maintain, while ensuring errors are handled gracefully.

[ombhojane]

Thanks for contributing