#!/usr/bin/env python3
# Author: Simon-Pierre Boucher — simon-pierre.boucher@uqo.ca
"""Leakage audit on the built data layer → results/data/leakage_audit.json (exit code 1 if any check fails).

    python3 scripts/05_leakage_audit.py [--synthetic] [--subjects 300] [--asof-rows 20000]
"""
from __future__ import annotations

import argparse
import sys
import time

import polars as pl

from agentic_valuation.config import PROCESSED, RESULTS, SYNTHETIC, load_config
from agentic_valuation.data.leakage import audit, to_json


def main() -> None:
    ap = argparse.ArgumentParser(description=__doc__)
    ap.add_argument("--synthetic", action="store_true")
    ap.add_argument("--subjects", type=int, default=300)
    ap.add_argument("--asof-rows", type=int, default=20_000, help="rows (whole municipalities) for the feature-matrix perturbation test")
    args = ap.parse_args()
    base = SYNTHETIC if args.synthetic else PROCESSED
    t0 = time.perf_counter()
    analytic = pl.read_parquet(base / "analytic.parquet")
    splits = pl.read_parquet(base / "splits.parquet")
    subs = pl.read_parquet(base / "llm_subsamples.parquet")
    matrix = pl.read_parquet(base / "feature_matrix.parquet") if (base / "feature_matrix.parquet").exists() else None
    geo = pl.read_parquet(base / "geo_features.parquet") if (base / "geo_features.parquet").exists() else None
    n_by = {"validation": 40, "test": 60, "holdout": 30} if args.synthetic else None
    report = audit(analytic, splits, subs, matrix, geo, load_config("data"), n_subjects=args.subjects, n_by_split=n_by, asof_sample=args.asof_rows)
    report["synthetic"] = args.synthetic
    report["seconds"] = round(time.perf_counter() - t0, 1)
    out = RESULTS / "data" / ("leakage_audit_synthetic.json" if args.synthetic else "leakage_audit.json")
    out.parent.mkdir(parents=True, exist_ok=True)
    out.write_text(to_json(report))
    for c in report["checks"]:
        print(f"[{'PASS' if c['passed'] else 'FAIL'}] {c['name']}: checked={c['n_checked']} failed={c['n_failed']}")
    print(f"all_passed={report['all_passed']} seconds={report['seconds']} → {out}")
    sys.exit(0 if report["all_passed"] else 1)


if __name__ == "__main__":
    main()