"""Run the full daily post-stop re-entry study from one candidate replay.

The expensive point-in-time setup replay and cross-sectional ranking happen
once.  Every policy, lookback, transaction-cost, capacity, and holdout arm then
uses that same qualified daily candidate set, so differences come only from the
portfolio/re-entry rules being compared.
"""

from __future__ import annotations

import argparse
import asyncio
import copy
import json
import multiprocessing
import os
import pickle
import sys
from collections import Counter
from concurrent.futures import ProcessPoolExecutor, as_completed
from datetime import date, datetime
from pathlib import Path
from typing import Any

from sqlalchemy import select
from sqlalchemy.ext.asyncio import AsyncSession, async_sessionmaker, create_async_engine

ROOT = Path(__file__).resolve().parents[1]
if str(ROOT) not in sys.path:
    sys.path.insert(0, str(ROOT))

POLICY_NAMES = (
    "immediate",
    "next_session",
    "cooldown_5",
    "gate_reset",
    "gate_reset_improved",
    "two_session_confirmation",
)
CACHE_VERSION = "daily-reentry-matrix-v2-full-ranking-universe"


def _sqlite_url(path: Path) -> str:
    return f"sqlite+aiosqlite:///{path.resolve().as_posix()}"


def _parse_args() -> argparse.Namespace:
    parser = argparse.ArgumentParser(description=__doc__)
    parser.add_argument("snapshot", help="SQLite backtest snapshot.")
    parser.add_argument(
        "--start-date",
        default=None,
        help="Optional earliest replay/simulation date (YYYY-MM-DD).",
    )
    parser.add_argument("--workers", type=int, default=6)
    parser.add_argument("--out", default=None)
    parser.add_argument(
        "--candidate-cache",
        default=None,
        help=(
            "Optional pickle cache. It stores only the ranked, production-qualified "
            "daily candidates, not the much larger raw replay."
        ),
    )
    parser.add_argument(
        "--policies",
        nargs="+",
        choices=POLICY_NAMES,
        default=list(POLICY_NAMES),
    )
    parser.add_argument("--base-cost-per-side-pct", type=float, default=0.1)
    parser.add_argument("--base-capacity", type=int, default=10)
    parser.add_argument(
        "--costs-per-side-pct",
        type=float,
        nargs="+",
        default=[0.1, 0.2, 0.3],
    )
    parser.add_argument(
        "--capacities", type=int, nargs="+", default=[5, 10, 15]
    )
    parser.add_argument(
        "--holdout-split",
        default="2025-01-01",
        help="Train/test split date (YYYY-MM-DD), or 'none' to disable.",
    )
    parser.add_argument("--quiet", action="store_true")
    return parser.parse_args()


def _default_output_path() -> Path:
    stamp = datetime.now().strftime("%Y%m%d-%H%M%S")
    return Path("reports") / f"daily-reentry-matrix-{stamp}.json"


class PrecomputedDailyEngine:
    """Exact date/symbol lookup over the already-ranked production gate."""

    def __init__(self, qualified_candidates: list[dict]) -> None:
        self.by_key = {
            (row["symbol"], date.fromisoformat(row["date"]).toordinal()): row
            for row in qualified_candidates
        }

    def candidate(self, symbol: str, asof_ord: int) -> dict | None:
        row = self.by_key.get((symbol, asof_ord))
        return dict(row) if row is not None else None


class ReentryPolicy:
    """Stateful policy evaluated after every initial-stop exit."""

    def __init__(
        self,
        name: str,
        engine: PrecomputedDailyEngine,
        ranking_key: str,
    ) -> None:
        if name not in POLICY_NAMES:
            raise ValueError(f"Unknown re-entry policy: {name}")
        self.name = name
        self.engine = engine
        self.ranking_key = ranking_key
        self.checks = 0
        self.gate_passes = 0
        self.emitted = Counter()

    def __call__(
        self,
        symbol: str,
        asof_ord: int,
        state: dict,
        _bar: Any,
    ) -> dict | None:
        self.checks += 1
        sessions = int(state["sessions_since_stop"])
        candidate = self.engine.candidate(symbol, asof_ord)
        if candidate is None:
            state["gate_went_unqualified"] = True
            state["qualified_streak"] = 0
            return None
        self.gate_passes += 1

        # Two-session confirmation means two complete post-stop closes. The
        # stop day's close (sessions=0) deliberately does not count.
        if self.name == "two_session_confirmation" and sessions == 0:
            state["qualified_streak"] = 0
            return None
        state["qualified_streak"] = int(state.get("qualified_streak", 0)) + 1

        reason: str | None = None
        if self.name == "immediate":
            reason = "gate_still_or_again_qualified"
        elif self.name == "next_session":
            if sessions >= 1:
                reason = "stop_day_block_complete"
        elif self.name == "cooldown_5":
            if sessions >= 5:
                reason = "5_session_cooldown_complete"
        elif self.name == "gate_reset":
            if state["gate_went_unqualified"]:
                reason = "gate_failed_then_requalified"
        elif self.name == "gate_reset_improved":
            previous_rank = state.get("previous_rank")
            current_rank = candidate.get(self.ranking_key)
            rank_not_weaker = (
                current_rank is not None
                and (
                    previous_rank is None
                    or float(current_rank) >= float(previous_rank)
                )
            )
            if (
                state["gate_went_unqualified"]
                and float(candidate["stop"]) > float(state["previous_stop"])
                and rank_not_weaker
            ):
                reason = "gate_reset_with_improved_stop_and_rank"
        elif self.name == "two_session_confirmation":
            if state["qualified_streak"] >= 2:
                reason = "two_qualified_post_stop_closes"

        if reason is None:
            return None
        emitted = dict(candidate)
        emitted["_reentry_reason"] = reason
        self.emitted[reason] += 1
        return emitted

    def summary(self) -> dict:
        return {
            "daily_checks": self.checks,
            "qualified_checks": self.gate_passes,
            "emitted_candidates_by_reason": dict(self.emitted),
        }


def _trade_summary(trades: list[dict]) -> dict:
    reentries = [trade for trade in trades if trade.get("is_reentry")]
    waits = [
        int(trade["reentry_wait_sessions"])
        for trade in reentries
        if trade.get("reentry_wait_sessions") is not None
    ]
    return {
        "transaction_cost": round(
            sum(float(trade["transaction_cost"]) for trade in trades), 2
        ),
        "reentry_trades": len(reentries),
        "same_day_reentries": sum(wait == 0 for wait in waits),
        "next_session_reentries": sum(wait == 1 for wait in waits),
        "reentries_within_5_sessions": sum(wait <= 5 for wait in waits),
        "avg_reentry_wait_sessions": (
            round(sum(waits) / len(waits), 1) if waits else None
        ),
        "reentry_win_rate": (
            round(
                sum(float(trade["pnl"]) > 0 for trade in reentries)
                / len(reentries)
                * 100.0,
                1,
            )
            if reentries
            else None
        ),
        "reentry_total_pnl": round(
            sum(float(trade["pnl"]) for trade in reentries), 2
        ),
    }


def _parse_optional_date(value: str | None, option: str) -> date | None:
    if value is None or value.strip().lower() == "none":
        return None
    try:
        return date.fromisoformat(value)
    except ValueError as exc:
        raise SystemExit(f"{option} must use YYYY-MM-DD or 'none'") from exc


def _max_date(left: date | None, right: date | None) -> date | None:
    if left is None:
        return right
    if right is None:
        return left
    return max(left, right)


async def _main() -> None:
    args = _parse_args()
    snapshot = Path(args.snapshot)
    if not snapshot.exists():
        raise SystemExit(f"Snapshot not found: {snapshot}")
    requested_start = _parse_optional_date(args.start_date, "--start-date")
    holdout_split = _parse_optional_date(args.holdout_split, "--holdout-split")
    if args.workers < 1:
        raise SystemExit("--workers must be positive")
    if args.base_capacity < 1 or any(value < 1 for value in args.capacities):
        raise SystemExit("capacities must be positive")
    all_costs = sorted(
        set([args.base_cost_per_side_pct, *args.costs_per_side_pct])
    )
    if any(value < 0 or value >= 100 for value in all_costs):
        raise SystemExit("cost percentages must be in [0, 100)")
    all_capacities = sorted(set([args.base_capacity, *args.capacities]))
    policies = tuple(dict.fromkeys(args.policies))

    os.environ["BACKTEST_SNAPSHOT_OFFLINE"] = "1"
    os.environ["BACKTEST_ALLOW_SPAWN"] = "1"

    from app.models.ticker import Ticker
    from app.services import backtest_service as bt
    from app.services.admin_service import get_activation_config
    from app.services.paper_trade_service import get_exit_policy
    from app.services.recommendation_service import get_recommendation_config

    db_engine = create_async_engine(_sqlite_url(snapshot), pool_pre_ping=True)
    Session = async_sessionmaker(
        db_engine, class_=AsyncSession, expire_on_commit=False
    )
    try:
        async with Session() as db:
            recommendation_config = await get_recommendation_config(db)
            activation = await get_activation_config(db)
            exit_config = await get_exit_policy(db)
            benchmark_closes = await bt._load_benchmark_closes_for_backtest(
                db, days=None, refresh=False
            )
            ticker_result = await db.execute(select(Ticker).order_by(Ticker.symbol))
            symbols = [ticker.symbol for ticker in ticker_result.scalars().all()]
            prices: dict[str, tuple] = {}
            for index, symbol in enumerate(symbols, 1):
                columns = await bt._fetch_columns(db, symbol)
                if columns is not None:
                    prices[symbol] = columns
                if not args.quiet and index % 50 == 0:
                    print(f"loaded prices: {index}/{len(symbols)}", flush=True)
    finally:
        await db_engine.dispose()

    replay_start = requested_start or date(1900, 1, 1)
    snapshot_stat = snapshot.stat()
    cache_key = {
        "version": CACHE_VERSION,
        "snapshot": str(snapshot.resolve()),
        "snapshot_size": snapshot_stat.st_size,
        "snapshot_mtime_ns": snapshot_stat.st_mtime_ns,
        "start_date": replay_start.isoformat(),
        "cadence": "daily",
        "target_model": "production_gtl",
    }
    cache_path = Path(args.candidate_cache) if args.candidate_cache else None
    qualified_candidates: list[dict] | None = None
    entry_candidate_count = 0
    entry_candidates_by_direction: dict[str, int] = {}
    if cache_path is not None and cache_path.exists():
        with cache_path.open("rb") as handle:
            cached = pickle.load(handle)  # noqa: S301 - trusted local cache
        if cached.get("key") == cache_key:
            qualified_candidates = list(cached["qualified_candidates"])
            entry_candidate_count = int(cached["entry_candidate_count"])
            entry_candidates_by_direction = dict(
                cached["entry_candidates_by_direction"]
            )
            if not args.quiet:
                print(f"loaded qualified candidate cache: {cache_path}", flush=True)
        elif not args.quiet:
            print(f"candidate cache mismatch; rebuilding: {cache_path}", flush=True)

    if qualified_candidates is None:
        candidates: list[dict] = []
        workers = max(1, min(int(args.workers), multiprocessing.cpu_count() - 1))
        context = bt._mp_context() or multiprocessing.get_context("spawn")
        with ProcessPoolExecutor(max_workers=workers, mp_context=context) as pool:
            futures = {
                pool.submit(
                    bt._replay_candidates_for_period,
                    symbol,
                    columns,
                    recommendation_config,
                    activation,
                    benchmark_closes,
                    replay_start,
                    "daily",
                    True,
                ): symbol
                for symbol, columns in prices.items()
            }
            for index, future in enumerate(as_completed(futures), 1):
                candidates.extend(future.result())
                if not args.quiet and index % 25 == 0:
                    print(f"daily replay: {index}/{len(futures)} tickers", flush=True)

        entry_candidate_count = len(candidates)
        entry_candidates_by_direction = dict(
            Counter(row["direction"] for row in candidates)
        )
        bt._assign_momentum_percentiles(candidates)
        bt._assign_residual_momentum_percentiles(candidates)
        bt._assign_low_volatility_percentiles(candidates)
        bt._assign_activation_momentum_percentiles(candidates)
        bt._assign_residual_high_vol_blend(candidates)
        threshold = float(activation.get("min_momentum_percentile", 80.0))
        for candidate in candidates:
            candidate["qualified"] = bt._momentum_qualifies(candidate, threshold)
        qualified_candidates = [
            candidate
            for candidate in candidates
            if candidate["qualified"] and candidate.get("direction") == "long"
        ]
        del candidates
        if cache_path is not None:
            cache_path.parent.mkdir(parents=True, exist_ok=True)
            with cache_path.open("wb") as handle:
                pickle.dump(
                    {
                        "key": cache_key,
                        "entry_candidate_count": entry_candidate_count,
                        "entry_candidates_by_direction": (
                            entry_candidates_by_direction
                        ),
                        "qualified_candidates": qualified_candidates,
                    },
                    handle,
                    protocol=pickle.HIGHEST_PROTOCOL,
                )
            if not args.quiet:
                print(f"wrote qualified candidate cache: {cache_path}", flush=True)

    if not qualified_candidates:
        raise RuntimeError("Daily replay produced no production-qualified candidates")

    strategy = next(
        row for row in bt.PORTFOLIO_MONITOR_STRATEGIES if row.get("is_production")
    )
    entry_config = bt._entry_variant_config(str(strategy["entry_variant"]))
    if entry_config is None:
        raise RuntimeError("Production entry configuration missing")
    ranking_key = str(
        entry_config.get("ranking_key") or entry_config["percentile_key"]
    )
    threshold = float(activation.get("min_momentum_percentile", 80.0))
    exit_policy = bt.LIVE_EXIT_MODE_TO_SIM.get(
        str(exit_config.get("mode", "atr_trailing")), "atr_trail3"
    )
    hold_days = int(exit_config.get("hold_days", max(bt.TIME_EXIT_DAYS)))
    trail_multiplier = float(
        exit_config.get("atr_multiplier", bt.ATR_TRAIL_MULTIPLIER)
    )
    qualified_symbols = {row["symbol"] for row in qualified_candidates}
    simulation_prices = {
        symbol: columns
        for symbol, columns in prices.items()
        if symbol in qualified_symbols
    }
    daily_engine = PrecomputedDailyEngine(qualified_candidates)
    latest_ord = max(
        date.fromisoformat(row["date"]).toordinal()
        for row in qualified_candidates
    )
    latest_date = date.fromordinal(latest_ord)
    if holdout_split is not None and not (
        (requested_start or date.min) < holdout_split <= latest_date
    ):
        raise SystemExit(
            f"--holdout-split must be after the start and no later than {latest_date}"
        )

    run_cache: dict[tuple, dict] = {}
    completed_sims = 0

    def run_policy(
        policy_name: str,
        *,
        start_date: date | None,
        end_date: date | None,
        cost_pct: float,
        capacity: int,
    ) -> dict:
        nonlocal completed_sims
        key = (policy_name, start_date, end_date, float(cost_pct), int(capacity))
        if key in run_cache:
            return copy.deepcopy(run_cache[key])
        policy = ReentryPolicy(policy_name, daily_engine, ranking_key)
        sim = bt._simulate_portfolio(
            qualified_candidates,
            simulation_prices,
            benchmark_closes,
            exit_policy,
            hold_days,
            ranking_key=ranking_key,
            max_positions=capacity,
            risk_per_trade=float(entry_config["risk_per_trade"]),
            atr_trail_multiplier=trail_multiplier,
            cost_per_side=cost_pct / 100.0,
            start_date=start_date,
            end_date=end_date,
            post_stop_reentry_fn=policy,
            include_trades=True,
        )
        if sim is None:
            raise RuntimeError(f"Policy {policy_name} produced no trades")
        trades = list(sim.pop("trade_details"))
        events = list(sim.pop("reentry_events", []))
        row = {
            **sim,
            "turnover": _trade_summary(trades),
            "policy": policy.summary(),
            "reentry_events": events,
        }
        run_cache[key] = row
        completed_sims += 1
        if not args.quiet:
            print(
                f"portfolio simulations: {completed_sims} "
                f"({policy_name}, cost={cost_pct}%, capacity={capacity})",
                flush=True,
            )
        return copy.deepcopy(row)

    primary: list[dict] = []
    for lookback in bt.PORTFOLIO_MONITOR_LOOKBACKS:
        lookback_start = bt._lookback_start(latest_ord, lookback["days"])
        sim_start = _max_date(requested_start, lookback_start)
        for policy_name in policies:
            row = run_policy(
                policy_name,
                start_date=sim_start,
                end_date=None,
                cost_pct=args.base_cost_per_side_pct,
                capacity=args.base_capacity,
            )
            if lookback["lookback"] != "all":
                row.pop("reentry_events", None)
            primary.append({
                "arm": policy_name,
                "lookback": lookback["lookback"],
                "lookback_label": lookback["label"],
                "capacity": args.base_capacity,
                **row,
            })

    immediate_baseline_parity: dict
    if "immediate" in policies:
        direct_daily_baseline = bt._simulate_portfolio(
            qualified_candidates,
            simulation_prices,
            benchmark_closes,
            exit_policy,
            hold_days,
            ranking_key=ranking_key,
            max_positions=args.base_capacity,
            risk_per_trade=float(entry_config["risk_per_trade"]),
            atr_trail_multiplier=trail_multiplier,
            cost_per_side=args.base_cost_per_side_pct / 100.0,
            start_date=requested_start,
        )
        if direct_daily_baseline is None:
            raise RuntimeError("Direct daily no-lockdown baseline produced no trades")
        immediate_all = next(
            row
            for row in primary
            if row["lookback"] == "all" and row["arm"] == "immediate"
        )
        parity_fields = tuple(sorted(direct_daily_baseline))
        parity_differences = {
            field: {
                "direct_daily_baseline": direct_daily_baseline.get(field),
                "immediate_callback": immediate_all.get(field),
            }
            for field in parity_fields
            if direct_daily_baseline.get(field) != immediate_all.get(field)
        }
        if parity_differences:
            raise RuntimeError(
                "Immediate callback diverges from direct daily baseline: "
                f"{parity_differences}"
            )
        immediate_baseline_parity = {
            "passed": True,
            "compared_fields": list(parity_fields),
            "direct_daily_baseline": direct_daily_baseline,
        }
    else:
        immediate_baseline_parity = {
            "passed": None,
            "skipped": "immediate policy was not selected",
        }

    robustness: list[dict] = []
    for cost_pct in all_costs:
        for capacity in all_capacities:
            for policy_name in policies:
                row = run_policy(
                    policy_name,
                    start_date=requested_start,
                    end_date=None,
                    cost_pct=cost_pct,
                    capacity=capacity,
                )
                row.pop("reentry_events", None)
                robustness.append({
                    "arm": policy_name,
                    "lookback": "all",
                    "cost_per_side_pct_requested": cost_pct,
                    "capacity": capacity,
                    **row,
                })

    holdout: list[dict] = []
    if holdout_split is not None:
        for segment, segment_start, segment_end in (
            ("train", requested_start, holdout_split),
            ("test", _max_date(requested_start, holdout_split), None),
        ):
            for policy_name in policies:
                row = run_policy(
                    policy_name,
                    start_date=segment_start,
                    end_date=segment_end,
                    cost_pct=args.base_cost_per_side_pct,
                    capacity=args.base_capacity,
                )
                row.pop("reentry_events", None)
                holdout.append({
                    "arm": policy_name,
                    "segment": segment,
                    "split_date": holdout_split.isoformat(),
                    "capacity": args.base_capacity,
                    **row,
                })

    output = Path(args.out) if args.out else _default_output_path()
    report = {
        "generated_at": datetime.now().astimezone().isoformat(),
        "snapshot": str(snapshot.resolve()),
        "period_start_requested": (
            requested_start.isoformat() if requested_start else None
        ),
        "last_eligible_candidate_date": latest_date.isoformat(),
        "tickers_loaded": len(prices),
        "tickers_qualified": len(qualified_symbols),
        "entry_candidates": entry_candidate_count,
        "entry_candidates_by_direction": entry_candidates_by_direction,
        "qualified_candidates": len(qualified_candidates),
        "params": {
            "entry_cadence": "daily",
            "target_model": "production_gtl",
            "ranking_universe": "all_long_and_short_setups",
            "policies": list(policies),
            "base_cost_per_side_pct": args.base_cost_per_side_pct,
            "base_capacity": args.base_capacity,
            "robustness_costs_per_side_pct": all_costs,
            "robustness_capacities": all_capacities,
            "holdout_split": (
                holdout_split.isoformat() if holdout_split else None
            ),
            "setup_stop_atr_multiplier": bt.ATR_MULTIPLIER,
            "exit_policy": exit_policy,
            "exit_atr_multiplier": trail_multiplier,
            "hold_days": hold_days,
            "risk_per_trade": float(entry_config["risk_per_trade"]),
            "momentum_percentile_floor": threshold,
            "ranking_key": ranking_key,
        },
        "primary_lookback_matrix": primary,
        "immediate_baseline_parity": immediate_baseline_parity,
        "cost_capacity_robustness": robustness,
        "holdout": holdout,
        "portfolio_simulations_executed": completed_sims,
        "note": (
            "The point-in-time daily setup replay and universe ranking are executed "
            "once. Long and short setups both contribute to the production-faithful "
            "cross-sectional percentiles; only qualified longs are tradable. All arms "
            "reuse that identical production-qualified candidate set. The immediate "
            "callback, when selected, must match a direct daily no-lockdown "
            "simulation exactly. "
            "Immediate is the daily no-lockdown baseline; next_session blocks only "
            "the stop day; cooldown_5 permits re-entry at wait_sessions=5; gate_reset "
            "requires an unqualified close before requalification; "
            "gate_reset_improved additionally requires a higher stop and a non-weaker "
            "production rank; two_session_confirmation requires two consecutive "
            "qualified post-stop closes and excludes the stop day's close. Transaction "
            "costs alter cash and position sizing, not just reported P&L."
        ),
    }
    output.parent.mkdir(parents=True, exist_ok=True)
    output.write_text(json.dumps(report, indent=2) + "\n", encoding="utf-8")

    print(f"Report written: {output}")
    for row in primary:
        if row["lookback"] == "all":
            print(
                f"{row['arm']}: Sharpe {row['sharpe']}, CAGR {row['cagr_pct']}%, "
                f"DD {row['max_drawdown_pct']}%, trades {row['trades']}, "
                f"reentries {row['turnover']['reentry_trades']}, "
                f"fees ${row['turnover']['transaction_cost']}"
            )


if __name__ == "__main__":
    asyncio.run(_main())