scripts/report_eval.py

#!/usr/bin/env python
from __future__ import annotations

import argparse
import sys
from pathlib import Path

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

from dovla_cil.experiments.reports import generate_eval_report  # noqa: E402


def main(argv: list[str] | None = None) -> int:
    parser = argparse.ArgumentParser(
        description="Aggregate DoVLA-CIL evaluation metrics into CSV, Markdown, and plots."
    )
    parser.add_argument(
        "--inputs",
        nargs="+",
        required=True,
        help="One or more metrics.json paths or shell/glob patterns.",
    )
    parser.add_argument("--out", type=Path, required=True, help="Report output directory.")
    parser.add_argument("--name", default="evaluation_report", help="Experiment name for report.md.")
    args = parser.parse_args(argv)

    summary = generate_eval_report(args.inputs, args.out, experiment_name=args.name)
    print(f"report: {summary['markdown_report']}")
    print(f"aggregate_csv: {summary['aggregate_csv']}")
    print(f"num runs: {summary['num_runs']}")
    return 0


if __name__ == "__main__":
    raise SystemExit(main())