"""
Run a dataset condition benchmark
=================================

Classify AI4I or Raptor Maps samples from explicitly supplied local data.
See the :doc:`dataset guide </_generated/guides/dataset-use-cases>` for download
and installation instructions. Run this script from the repository root with
the dataset name, input path and ``--output`` directory.
"""

import argparse
import hashlib
import json
from pathlib import Path

from energy_analysis.dataset_benchmarks import evaluate, load_ai4i, load_infrared


def main():
    parser = argparse.ArgumentParser(description=__doc__)
    parser.add_argument("dataset", choices=["ai4i", "infrared"])
    parser.add_argument("input", type=Path)
    parser.add_argument("--output", type=Path, required=True)
    args = parser.parse_args()
    loader = load_ai4i if args.dataset == "ai4i" else load_infrared
    x, y, ids = loader(args.input)
    result = evaluate(x, y, ids, dataset=args.dataset, output=args.output)
    paths = [args.input] if args.input.is_file() else sorted(args.input.rglob("*"))
    manifest = {
        str(p.relative_to(args.input) if args.input.is_dir() else p.name): hashlib.sha256(
            p.read_bytes()
        ).hexdigest()
        for p in paths
        if p.is_file()
    }
    (args.output / "input_sha256.json").write_text(json.dumps(manifest, indent=2) + "\n")
    print(
        json.dumps(
            {
                "rows": len(y),
                "macro_f1": result["classification"]["macro avg"]["f1-score"],
                "baseline_macro_f1": result["baseline"]["macro avg"]["f1-score"],
            },
            indent=2,
        )
    )


if __name__ == "__main__":
    main()
