#!/usr/bin/env python3
"""
optuna_scalping_tpe_space.py — 스캘핑 Optuna 연속(TPE) 탐색 공간
기존 ``_scalp_grids()`` categorical 유지. ``use_macd_cross=False``, ``skip_hts=False``.
"""
from __future__ import annotations

from typing import Any, Dict, List

import optuna

from kis_trader.backtest.optuna_tpe_common import (
    ORDERBOOK_TPE_AXIS_KEYS,
    WHIPSAW_TPE_AXIS_KEYS,
    r1,
    r2,
    r3,
    suggest_orderbook_entry_tpe,
    suggest_whipsaw_tpe,
)

# 매수 시간창(time_*) 은 TPE 제외 — 운영 DB 고정값 사용(과적합·apply 후 진입 잠금 방지).
# 호가·휩쏘: 본 trial 엔진 평가 (사후「필터후」근사 대체).
_SCALP_BASE_KEYS: List[str] = [
    "rsi_period",
    "rsi_oversold",
    "rsi_overbought",
    "sl_pct",
    "tp_pct",
    "tp_max_pct",
    "drop_rate",
    "shoulder_min_high",
    "shoulder_cut_pct",
    "cooldown_min",
    "high_chase_thr",
    "max_daily_chg",
    "min_price",
    "vol_mult",
    "use_defense_filters",
    "require_reversal_candle",
    "max_loss_krw",
    "min_drop_pct_for_loss_cut",
    "min_margin",
    "min_hold_sec",
    "max_daily",
]
SCALP_TPE_AXIS_KEYS: List[str] = (
    list(_SCALP_BASE_KEYS) + list(ORDERBOOK_TPE_AXIS_KEYS) + list(WHIPSAW_TPE_AXIS_KEYS)
)


def scalp_tpe_axis_keys() -> List[str]:
    return list(SCALP_TPE_AXIS_KEYS)


def suggest_scalp_params_tpe(trial: optuna.Trial) -> Dict[str, Any]:
    combo: Dict[str, Any] = {}
    combo["rsi_period"] = trial.suggest_int("rsi_period", 3, 14)
    combo["rsi_oversold"] = r1(trial.suggest_float("rsi_oversold", 15.0, 30.0, step=0.5))
    combo["rsi_overbought"] = r1(trial.suggest_float("rsi_overbought", 65.0, 85.0, step=0.5))
    if combo["rsi_oversold"] >= combo["rsi_overbought"]:
        raise optuna.TrialPruned("rsi oversold >= overbought")

    combo["sl_pct"] = r2(trial.suggest_float("sl_pct", 1.0, 6.0, step=0.1))  # 확장: 1.5→1.0, 4.5→6.0
    combo["tp_pct"] = r2(trial.suggest_float("tp_pct", 0.5, 8.0, step=0.1))  # 확장: 1.5→0.5, 4.0→8.0
    combo["tp_max_pct"] = r2(trial.suggest_float("tp_max_pct", 1.0, 12.0, step=0.5))  # 확장: 2.0→1.0, 8.0→12.0
    if combo["tp_max_pct"] + 1e-9 < combo["tp_pct"]:
        raise optuna.TrialPruned("tp_max < tp")

    combo["drop_rate"] = r2(trial.suggest_float("drop_rate", 0.5, 10.0, step=0.1))  # 확장: 1.0→0.5, 6.0→10.0
    combo["shoulder_min_high"] = r2(
        trial.suggest_float("shoulder_min_high", 0.3, 4.0, step=0.1),
    )
    combo["shoulder_cut_pct"] = r2(
        trial.suggest_float("shoulder_cut_pct", 0.05, 0.6, step=0.05),
    )
    combo["cooldown_min"] = trial.suggest_int("cooldown_min", 0, 15)
    combo["high_chase_thr"] = r3(
        trial.suggest_float("high_chase_thr", 0.95, 1.0, step=0.005),
    )
    combo["max_daily_chg"] = r1(trial.suggest_float("max_daily_chg", 5.0, 60.0, step=1.0))  # 확장: 10→5, 50→60
    combo["min_price"] = trial.suggest_int("min_price", 1000, 8000, step=500)
    combo["vol_mult"] = r2(trial.suggest_float("vol_mult", 0.0, 2.0, step=0.1))
    combo["use_defense_filters"] = trial.suggest_categorical(
        "use_defense_filters", [False, True],
    )
    combo["require_reversal_candle"] = trial.suggest_categorical(
        "require_reversal_candle", [False, True],
    )
    combo["max_loss_krw"] = trial.suggest_int("max_loss_krw", 50000, 300000, step=25000)
    combo["min_drop_pct_for_loss_cut"] = r3(
        trial.suggest_float("min_drop_pct_for_loss_cut", 0.005, 0.03, step=0.001),
    )
    combo["min_margin"] = r2(trial.suggest_float("min_margin", 0.05, 0.5, step=0.05))
    combo["min_hold_sec"] = trial.suggest_int("min_hold_sec", 0, 300, step=10)  # 확장: 120→300초
    combo["max_daily"] = trial.suggest_int("max_daily", 5, 150, step=5)  # 확장: 10→5, 100→150

    combo["use_macd_cross"] = False
    combo["skip_hts_scan_dupes"] = False
    combo.update(suggest_orderbook_entry_tpe(trial))
    combo.update(suggest_whipsaw_tpe(trial))
    return combo