#!/usr/bin/env python3
import argparse
import json
from pathlib import Path

import numpy as np
import pandas as pd


def sample_sequence(length: int, gc: float, rng: np.random.Generator) -> str:
    p_gc = max(0.0, min(1.0, gc)) / 2.0
    p_at = (1.0 - max(0.0, min(1.0, gc))) / 2.0
    return "".join(rng.choice(list("ACGT"), size=length, p=[p_at, p_gc, p_gc, p_at]).tolist())


def main():
    parser = argparse.ArgumentParser(description="Build a GC-matched random DNA negative baseline.")
    parser.add_argument("--dataset_dir", required=True)
    parser.add_argument("--split", default="valid")
    parser.add_argument("--output_jsonl", required=True)
    parser.add_argument("--num_samples", type=int, default=384)
    parser.add_argument("--seed", type=int, default=42)
    args = parser.parse_args()

    rng = np.random.default_rng(args.seed)
    df = pd.read_parquet(Path(args.dataset_dir) / f"{args.split}.parquet")
    df = df.sample(n=min(args.num_samples, len(df)), random_state=args.seed).reset_index(drop=True)
    output = Path(args.output_jsonl)
    output.parent.mkdir(parents=True, exist_ok=True)

    with open(output, "w", encoding="utf-8") as f:
        for _, row in df.iterrows():
            ref = str(row["sequence"]).upper()
            gc = (ref.count("G") + ref.count("C")) / len(ref)
            seq = sample_sequence(len(ref), gc, rng)
            f.write(
                json.dumps(
                    {
                        "source": "generated",
                        "activity_bucket": "random_gc_matched",
                        "generated_sequence": seq,
                        "reference_sequence": ref,
                        "valid_dna": True,
                        "generated_bp_length": len(seq),
                    }
                )
                + "\n"
            )
    print(output)


if __name__ == "__main__":
    main()