from __future__ import annotations

import argparse
import json
import sys
from datetime import UTC, datetime
from pathlib import Path


DEFAULT_THRESHOLDS = {
    "golden_source_recall_at_k": 0.5,
    "golden_citation_precision": 0.5,
    "adversarial_safe_handling_rate": 0.5,
    "governance_routing_accuracy": 0.5,
    "retrieval_source_recall_at_k": 0.5,
}


def parse_args() -> argparse.Namespace:
    parser = argparse.ArgumentParser(description="Run eval suites and enforce release-gate thresholds.")
    parser.add_argument("--limit", type=int, default=25)
    parser.add_argument("--thresholds-json", default="")
    parser.add_argument("--json-out", default="eval/dashboards/release_gate_summary.json")
    return parser.parse_args()


def load_thresholds(path: str) -> dict[str, float]:
    if not path:
        return DEFAULT_THRESHOLDS
    with Path(path).open(encoding="utf-8") as handle:
        return json.load(handle)


def main() -> int:
    args = parse_args()
    thresholds = load_thresholds(args.thresholds_json)

    repo_root = Path(__file__).resolve().parents[2]
    if str(repo_root) not in sys.path:
        sys.path.insert(0, str(repo_root))

    from eval.runners.run_golden_memory_eval import evaluate_rows as eval_golden, load_rows as load_golden
    from eval.runners.run_adversarial_memory_eval import evaluate_rows as eval_adversarial, load_rows as load_adversarial
    from eval.runners.run_governance_policy_eval import evaluate_rows as eval_governance, load_rows as load_governance
    from eval.runners.run_retrieval_stress_eval import evaluate_rows as eval_retrieval, load_rows as load_retrieval

    golden = eval_golden(load_golden(repo_root / "output/golden_medical_qa.csv", args.limit))
    adversarial = eval_adversarial(load_adversarial(repo_root / "output/adversarial_medical_qa.csv", args.limit))
    governance = eval_governance(load_governance(repo_root / "output/governance_policy_cases.csv", args.limit))
    retrieval = eval_retrieval(load_retrieval(repo_root / "output/retrieval_stress_cases.csv", args.limit))

    checks = {
        "golden_source_recall_at_k": golden["overall"]["source_recall_at_k"],
        "golden_citation_precision": golden["overall"]["citation_precision"],
        "adversarial_safe_handling_rate": adversarial["overall"]["safe_handling_rate"],
        "governance_routing_accuracy": governance["overall"]["routing_accuracy"],
        "retrieval_source_recall_at_k": retrieval["overall"]["source_recall_at_k"],
    }
    failures = [
        name
        for name, actual in checks.items()
        if actual < thresholds[name]
    ]
    summary = {
        "generated_at": datetime.now(UTC).isoformat(),
        "rows_per_suite": args.limit,
        "thresholds": thresholds,
        "actuals": checks,
        "passed": not failures,
        "failures": failures,
    }
    out_path = repo_root / args.json_out
    out_path.parent.mkdir(parents=True, exist_ok=True)
    out_path.write_text(json.dumps(summary, indent=2), encoding="utf-8")
    print(json.dumps(summary, indent=2))
    return 0 if summary["passed"] else 1


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