main.py

"""
Entry point — runs both RL pipelines on bank-full.csv.

Usage:
    python main.py
"""

import sys
from pathlib import Path

# Make project root importable regardless of working directory
sys.path.insert(0, str(Path(__file__).parent))

from data.loader import load_and_preprocess, check_edge_cases
from ground_truth.generator import generate_ground_truth
from pipeline1.pipeline import run_pipeline1
from pipeline2.pipeline import run_pipeline2
from config import OUTPUT_DIR


def _print_final_summary(p1: dict, p2: list[dict]) -> None:
    print("\n" + "=" * 60)
    print("FINAL SUMMARY")
    print("=" * 60)

    print("\n[Pipeline 1 — Imputation]  Best method per scenario (by MAE):")
    for scenario, metrics in p1.items():
        best = min(metrics, key=lambda k: metrics[k]["NRMSE"])
        m    = metrics[best]
        print(f"  {scenario:<22} -> {best:<8}  MAE={m['MAE']:.4f}  RMSE={m['RMSE']:.4f}")

    print("\n[Pipeline 2 — Classification]  Ranked by F1:")
    for m in sorted(p2, key=lambda x: x["F1"], reverse=True):
        print(f"  {m['name']:<35}  F1={m['F1']:.4f}  "
              f"MCC={m['MCC']:.4f}  G-Mean={m['G-Mean']:.4f}")

    print(f"\nAll outputs -> {OUTPUT_DIR.resolve()}")
    print("Files:")
    for f in sorted(OUTPUT_DIR.iterdir()):
        print(f"  {f.name}")


def main() -> None:
    print("\n" + "#" * 60)
    print("  RL PIPELINES -- BANK MARKETING DATASET")
    print("#" * 60)

    df, _encoders = load_and_preprocess()
    check_edge_cases(df)

    generate_ground_truth(df)

    p1_metrics = run_pipeline1(df)
    p2_metrics = run_pipeline2(df)

    _print_final_summary(p1_metrics, p2_metrics)


if __name__ == "__main__":
    main()