"""H5 (exploratory): cost-aware signal evaluation from saved predictions (runs where pred_*.npz live).
Median of the cumulative return = sum of per-minute Student-t locations (symmetric marginals)."""
from __future__ import annotations

import argparse
import json
import sys
from pathlib import Path

import numpy as np
import pandas as pd

ROOT = Path(__file__).resolve().parents[1]
sys.path.insert(0, str(ROOT / "src"))
from chronos_x.eval.economic import signal_pnl  # noqa: E402


def main() -> None:
    ap = argparse.ArgumentParser(); ap.add_argument("--exp", default="main"); ap.add_argument("--tag", default="test")
    ap.add_argument("--models", default="mrpt_small,mrpt_base,mlp,tcn"); ap.add_argument("--horizons", default="5,15,30,60")
    a = ap.parse_args()
    d = ROOT / "results" / "metrics" / a.exp
    tab = pd.read_csv(d / f"origins_{a.tag}.csv")
    out = []
    for m in a.models.split(","):
        f = d / f"pred_{m}_{a.tag}.npz"
        if not f.exists():
            continue
        p = np.load(f)
        mu = p["mu"]; ym = p["y_mask"]
        for H in (int(h) for h in a.horizons.split(",")):
            med = (mu[:, :H] * ym[:, :H]).sum(1)
            real = tab[f"cum_{H}"].values
            assert np.allclose((p["y"][:, :H] * ym[:, :H]).sum(1), real, atol=1e-6)
            for cost in (0.0, 1.0, 2.5, 5.0, 10.0):
                for k in (1.0, 2.0):
                    r = signal_pnl(med, real, tab, cost_bps=cost, k=k)
                    out.append({"model": m, "H": H, **r})
            # information content without trading: correlation and regression slope of realised on median
            sl = np.polyfit(med, real, 1)[0] if med.std() > 0 else np.nan
            out.append({"model": m, "H": H, "cost_bps": -1, "k": 0, "corr": float(np.corrcoef(med, real)[0, 1]), "slope": float(sl), "n": int(len(med)),
                        "median_abs_bps": float(np.abs(med).mean() * 1e4)})
    df = pd.DataFrame(out); df.to_csv(d / f"economic_{a.tag}.csv", index=False)
    print(df.to_string(index=False, float_format=lambda x: f"{x:.4g}"))


if __name__ == "__main__":
    main()