CT/lung2/scripts/run_fusion_benchmark.py

from __future__ import annotations

import argparse
import sys
from pathlib import Path

ROOT = Path(__file__).resolve().parents[1]
if str(ROOT) not in sys.path:
    sys.path.insert(0, str(ROOT))

import numpy as np
import pandas as pd

from src.models.complementarity import save_json
from src.models.fusion import concat_features, cross_validated_benchmark, feature_matrix, stacking_fusion
from src.utils import ensure_dir, load_config, seed_everything


def _load_feature(path: Path) -> pd.DataFrame:
    if not path.exists():
        raise FileNotFoundError(path)
    return pd.read_csv(path)


def main() -> None:
    parser = argparse.ArgumentParser()
    parser.add_argument("--config", default="configs/fusion_baseline.yaml")
    args = parser.parse_args()
    cfg = load_config(args.config)
    seed_everything(cfg.get("seed", 17))
    features_dir = Path(cfg["paths"]["features_dir"])
    out_dir = ensure_dir(cfg["paths"]["output_dir"])

    labels = pd.read_csv(features_dir / "labels.csv")
    y = labels["label"].to_numpy()
    rad = _load_feature(features_dir / "radiomics_clean.csv")
    dl_bbox = _load_feature(features_dir / "dl_bbox_medicalnet.csv")
    dl_patch = _load_feature(features_dir / "dl_patch_triregion.csv")

    results = {}
    x_rad, _ = feature_matrix(rad)
    x_bbox, _ = feature_matrix(dl_bbox)
    x_patch, _ = feature_matrix(dl_patch)
    fusion_x = concat_features([rad, dl_patch]).drop(columns=["patient_id"]).to_numpy()

    results["radiomics"] = cross_validated_benchmark(x_rad, y, n_splits=cfg["classifier"]["cv_folds"], seed=cfg.get("seed", 17))
    results["dl_bbox"] = cross_validated_benchmark(x_bbox, y, n_splits=cfg["classifier"]["cv_folds"], seed=cfg.get("seed", 17))
    results["dl_patch"] = cross_validated_benchmark(x_patch, y, n_splits=cfg["classifier"]["cv_folds"], seed=cfg.get("seed", 17))
    results["fusion"] = cross_validated_benchmark(fusion_x, y, n_splits=cfg["classifier"]["cv_folds"], seed=cfg.get("seed", 17))
    results["stacking"] = stacking_fusion(x_rad, x_patch, y, n_splits=cfg["classifier"]["cv_folds"], seed=cfg.get("seed", 17))

    summary = {k: {"auc": v["auc"], "accuracy": v["accuracy"]} for k, v in results.items()}
    save_json(summary, out_dir / "benchmark_summary.json")
    print(summary)


if __name__ == "__main__":
    main()