lm_code/test2.py

import os
import time
import uuid
import datetime
from dataclasses import dataclass

from tqdm import tqdm
from loguru import logger

from bitmart.api_contract import APIContract
from bitmart.lib.cloud_exceptions import APIException

from 交易.tools import send_dingtalk_message


@dataclass
class StrategyConfig:
    # =============================
    # 1m | ETH 永续 | 控止损≤5/日
    # =============================

    # ===== 合约 =====
    contract_symbol: str = "ETHUSDT"
    open_type: str = "cross"
    leverage: str = "30"

    # ===== K线与指标 =====
    step_min: int = 1
    lookback_min: int = 240
    ema_len: int = 36
    atr_len: int = 14

    # ===== ADX 趋势过滤（新增）=====
    # 目的：单边趋势（ADX高）时，抑制/禁止逆势均值回归单，避免反复反向开仓止损
    enable_adx_filter: bool = True
    adx_len: int = 14
    adx_threshold: float = 25.0  # 常用：20~30区间，你可按回测调整
    # 过滤模式：
    # - "block_countertrend": 只禁止逆着 DI 的方向开仓（推荐，既防反手又不完全停机）
    # - "block_all": ADX 高时直接不允许任何新开仓（更保守）
    adx_mode: str = "block_countertrend"
    # 趋势保护冷却：当 ADX 高且刚止损，延长冷却，减少“止损->立刻反手”的连环
    cooldown_sec_after_sl_extra: int = 40

    # =========================================================
    # ✅ 自动阈值：ATR/Price 分位数基准（更稳，不被短时噪声带跑）
    # =========================================================
    vol_baseline_window: int = 60
    vol_baseline_quantile: float = 0.65
    vol_scale_min: float = 0.80
    vol_scale_max: float = 1.60

    # ✅ baseline 每 60 秒刷新一次（体感更明显、也省CPU）
    base_ratio_refresh_sec: int = 180

    # =========================================================
    # ✅ 动态 floor（方案一）
    #   floor = clamp(min, base_k * base_ratio, max)
    #   目的：跟着典型波动变，过滤小噪声；tp/sl 也随环境自适应
    # =========================================================
    # entry_dev_floor 动态
    entry_dev_floor_min: float = 0.0012  # 0.12%
    entry_dev_floor_max: float = 0.0030  # 0.30%
    entry_dev_floor_base_k: float = 1.10  # entry_floor = 1.10 * base_ratio

    # tp_floor 动态
    tp_floor_min: float = 0.0006  # 0.06%
    tp_floor_max: float = 0.0020  # 0.20%
    tp_floor_base_k: float = 0.55  # tp_floor = 0.55 * base_ratio

    # sl_floor 动态
    sl_floor_min: float = 0.0018  # 0.18%
    sl_floor_max: float = 0.0060  # 0.60%
    sl_floor_base_k: float = 1.35  # sl_floor = 1.35 * base_ratio

    # =========================================================
    # ✅ 动态阈值倍率
    # =========================================================
    entry_k: float = 1.45
    tp_k: float = 0.65
    sl_k: float = 1.05

    # ===== 时间/冷却 =====
    max_hold_sec: int = 75
    cooldown_sec_after_exit: int = 20

    # ===== 下单/仓位 =====
    risk_percent: float = 0.004
    min_size: int = 1
    max_size: int = 5000

    # ===== 日内风控 =====
    daily_loss_limit: float = 0.02
    daily_profit_cap: float = 0.01

    # ===== 危险模式过滤 =====
    atr_ratio_kill: float = 0.0038
    big_body_kill: float = 0.010

    # ===== 轮询节奏 =====
    klines_refresh_sec: int = 10
    tick_refresh_sec: int = 1
    status_notify_sec: int = 60

    # =========================================================
    # ✅ 止损后同向入场加门槛（你原来的逻辑保留）
    # =========================================================
    reentry_penalty_mult: float = 1.55
    reentry_penalty_max_sec: int = 180
    reset_band_k: float = 0.45
    reset_band_floor: float = 0.0006

    # =========================================================
    # ✅ 止损后同方向 SL 放宽幅度与"止损时 vol_scale"联动
    # =========================================================
    post_sl_sl_max_sec: int = 90
    post_sl_mult_min: float = 1.02
    post_sl_mult_max: float = 1.16
    post_sl_vol_alpha: float = 0.20


class BitmartFuturesMeanReversionBot:
    def __init__(self, cfg: StrategyConfig):
        self.cfg = cfg

        self.api_key = os.getenv("BITMART_API_KEY", "").strip()
        self.secret_key = os.getenv("BITMART_SECRET_KEY", "").strip()
        self.memo = os.getenv("BITMART_MEMO", "合约交易").strip()

        if not self.api_key or not self.secret_key:
            raise RuntimeError("请先设置环境变量 BITMART_API_KEY / BITMART_SECRET_KEY / BITMART_MEMO(可选)")

        self.contractAPI = APIContract(self.api_key, self.secret_key, self.memo, timeout=(5, 15))

        # 持仓状态: -1 空, 0 无, 1 多
        self.pos = 0
        self.entry_price = None
        self.entry_ts = None
        self.last_exit_ts = 0

        # 日内权益基准
        self.day_start_equity = None
        self.trading_enabled = True
        self.day_tag = datetime.date.today()

        # 缓存
        self._klines_cache = None
        self._klines_cache_ts = 0
        self._last_status_notify_ts = 0

        # ✅ base_ratio 缓存
        self._base_ratio_cached = 0.0015  # 初始化默认值 0.15%
        self._base_ratio_ts = 0.0

        # ✅ 止损后"同向入场加门槛"状态
        self.last_sl_dir = 0  # 1=多止损，-1=空止损，0=无
        self.last_sl_ts = 0.0

        # ✅ 止损后"同方向 SL 联动放宽"状态
        self.post_sl_dir = 0
        self.post_sl_ts = 0.0
        self.post_sl_vol_scale = 1.0  # 记录止损时的 vol_scale

        self.pbar = tqdm(total=60, desc="运行中(秒)", ncols=90)
        logger.info(f"初始化完成，基准波动率默认值: {self._base_ratio_cached * 100:.4f}%")

    # ----------------- 通用工具 -----------------
    def ding(self, msg, error=False):
        prefix = "❌bitmart：" if error else "🔔bitmart："
        if error:
            for _ in range(3):
                send_dingtalk_message(f"{prefix}{msg}")
        else:
            send_dingtalk_message(f"{prefix}{msg}")

    def set_leverage(self) -> bool:
        try:
            resp = self.contractAPI.post_submit_leverage(
                contract_symbol=self.cfg.contract_symbol,
                leverage=self.cfg.leverage,
                open_type=self.cfg.open_type
            )[0]
            if resp.get("code") == 1000:
                logger.success(f"设置杠杆成功：{self.cfg.open_type} + {self.cfg.leverage}x")
                return True
            logger.error(f"设置杠杆失败: {resp}")
            self.ding(f"设置杠杆失败: {resp}", error=True)
            return False
        except Exception as e:
            logger.error(f"设置杠杆异常: {e}")
            self.ding(f"设置杠杆异常: {e}", error=True)
            return False

    # ----------------- 行情/指标 -----------------
    def get_klines_cached(self):
        now = time.time()
        if self._klines_cache is not None and (now - self._klines_cache_ts) < self.cfg.klines_refresh_sec:
            return self._klines_cache

        kl = self.get_klines()
        if kl:
            self._klines_cache = kl
            self._klines_cache_ts = now
        return self._klines_cache

    def get_klines(self):
        try:
            end_time = int(time.time())
            start_time = end_time - 60 * self.cfg.lookback_min

            resp = self.contractAPI.get_kline(
                contract_symbol=self.cfg.contract_symbol,
                step=self.cfg.step_min,
                start_time=start_time,
                end_time=end_time
            )[0]

            if resp.get("code") != 1000:
                logger.error(f"获取K线失败: {resp}")
                return None

            data = resp.get("data", [])
            formatted = []
            for k in data:
                formatted.append({
                    "id": int(k["timestamp"]),
                    "open": float(k["open_price"]),
                    "high": float(k["high_price"]),
                    "low": float(k["low_price"]),
                    "close": float(k["close_price"]),
                })
            formatted.sort(key=lambda x: x["id"])
            return formatted
        except Exception as e:
            logger.error(f"获取K线异常: {e}")
            self.ding(f"获取K线异常: {e}", error=True)
            return None

    def get_last_price(self, fallback_close: float) -> float:
        try:
            if hasattr(self.contractAPI, "get_contract_details"):
                r = self.contractAPI.get_contract_details(contract_symbol=self.cfg.contract_symbol)[0]
                d = r.get("data") if isinstance(r, dict) else None
                if isinstance(d, dict):
                    for key in ("last_price", "mark_price", "index_price"):
                        if key in d and d[key] is not None:
                            return float(d[key])

            if hasattr(self.contractAPI, "get_ticker"):
                r = self.contractAPI.get_ticker(contract_symbol=self.cfg.contract_symbol)[0]
                d = r.get("data") if isinstance(r, dict) else None
                if isinstance(d, dict):
                    for key in ("last_price", "price", "last", "close"):
                        if key in d and d[key] is not None:
                            return float(d[key])
        except Exception:
            pass

        return float(fallback_close)

    @staticmethod
    def ema(values, n: int) -> float:
        k = 2 / (n + 1)
        e = values[0]
        for v in values[1:]:
            e = v * k + e * (1 - k)
        return e

    @staticmethod
    def atr(klines, n: int) -> float:
        if len(klines) < n + 1:
            return 0.0
        trs = []
        for i in range(-n, 0):
            cur = klines[i]
            prev = klines[i - 1]
            tr = max(
                cur["high"] - cur["low"],
                abs(cur["high"] - prev["close"]),
                abs(cur["low"] - prev["close"]),
            )
            trs.append(tr)
        return sum(trs) / len(trs)

    @staticmethod
    def _wilder_smooth(prev: float, val: float, n: int) -> float:
        # Wilder smoothing: prev - prev/n + val
        return prev - (prev / n) + val

    def adx(self, klines, n: int):
        """
        返回 (adx, plus_di, minus_di)
        采用经典 Wilder ADX/DI 计算（足够稳，避免趋势期逆势反复开仓）
        """
        if len(klines) < (2 * n + 2):
            return 0.0, 0.0, 0.0

        highs = [k["high"] for k in klines]
        lows = [k["low"] for k in klines]
        closes = [k["close"] for k in klines]

        # 计算 TR, +DM, -DM 序列（从1开始）
        tr_list = []
        pdm_list = []
        mdm_list = []

        for i in range(1, len(klines)):
            high = highs[i]
            low = lows[i]
            prev_close = closes[i - 1]
            prev_high = highs[i - 1]
            prev_low = lows[i - 1]

            tr = max(high - low, abs(high - prev_close), abs(low - prev_close))
            up_move = high - prev_high
            down_move = prev_low - low

            pdm = up_move if (up_move > down_move and up_move > 0) else 0.0
            mdm = down_move if (down_move > up_move and down_move > 0) else 0.0

            tr_list.append(tr)
            pdm_list.append(pdm)
            mdm_list.append(mdm)

        # 初始平滑值（第一个n段的和）
        tr14 = sum(tr_list[:n])
        pdm14 = sum(pdm_list[:n])
        mdm14 = sum(mdm_list[:n])

        def safe_div(a, b):
            return (a / b) if b != 0 else 0.0

        plus_di = 100.0 * safe_div(pdm14, tr14)
        minus_di = 100.0 * safe_div(mdm14, tr14)
        dx = 100.0 * safe_div(abs(plus_di - minus_di), (plus_di + minus_di))
        dx_list = [dx]

        # 继续平滑并计算后续 DX
        for i in range(n, len(tr_list)):
            tr14 = self._wilder_smooth(tr14, tr_list[i], n)
            pdm14 = self._wilder_smooth(pdm14, pdm_list[i], n)
            mdm14 = self._wilder_smooth(mdm14, mdm_list[i], n)

            plus_di = 100.0 * safe_div(pdm14, tr14)
            minus_di = 100.0 * safe_div(mdm14, tr14)
            dx = 100.0 * safe_div(abs(plus_di - minus_di), (plus_di + minus_di))
            dx_list.append(dx)

        # ADX 是 DX 的 Wilder 平滑，常见做法：先取前 n 个 DX 的均值作为初值
        if len(dx_list) < (n + 1):
            return 0.0, plus_di, minus_di

        adx0 = sum(dx_list[:n]) / n
        adx_val = adx0
        for j in range(n, len(dx_list)):
            adx_val = (adx_val * (n - 1) + dx_list[j]) / n

        return float(adx_val), float(plus_di), float(minus_di)

    def is_danger_market(self, klines, price: float) -> bool:
        last = klines[-1]
        body = abs(last["close"] - last["open"]) / last["open"] if last["open"] else 0.0
        if body >= self.cfg.big_body_kill:
            return True

        a = self.atr(klines, self.cfg.atr_len)
        atr_ratio = (a / price) if price > 0 else 0.0
        if atr_ratio >= self.cfg.atr_ratio_kill:
            return True

        return False

    def atr_ratio_baseline(self, klines) -> float:
        """简化版ATR基准计算"""
        window = min(self.cfg.vol_baseline_window, len(klines) - self.cfg.atr_len - 1)
        if window <= 10:
            logger.warning(f"数据不足计算基准: {len(klines)}根K线")
            return 0.0

        ratios = []
        step = 3
        for i in range(-window, 0, step):
            if len(klines) + i < self.cfg.atr_len + 1:
                continue

            start_idx = len(klines) + i - self.cfg.atr_len
            end_idx = len(klines) + i
            if start_idx < 0 or end_idx <= start_idx:
                continue

            sub_klines = klines[start_idx:end_idx]
            if len(sub_klines) >= self.cfg.atr_len + 1:
                a = self.atr(sub_klines, self.cfg.atr_len)
                price = klines[end_idx - 1]["close"]
                if a > 0 and price > 0:
                    ratio = a / price
                    if 0.0001 < ratio < 0.01:
                        ratios.append(ratio)

        if len(ratios) < 5:
            a = self.atr(klines[-60:], self.cfg.atr_len)
            price = klines[-1]["close"]
            if a > 0 and price > 0:
                baseline = a / price
                logger.debug(f"使用全量数据计算基准: {baseline * 100:.4f}%")
                return baseline
            return 0.0

        ratios.sort()
        idx = min(len(ratios) - 1, max(0, int(self.cfg.vol_baseline_quantile * (len(ratios) - 1))))
        baseline = ratios[idx]
        logger.debug(
            f"基准计算: 样本数={len(ratios)}, 基准={baseline * 100:.4f}%, "
            f"范围=[{ratios[0] * 100:.4f}%, {ratios[-1] * 100:.4f}%]"
        )
        return baseline

    def get_base_ratio_cached(self, klines) -> float:
        """获取缓存的基准波动率，定期刷新"""
        now = time.time()
        refresh_sec = self.cfg.base_ratio_refresh_sec

        if (self._base_ratio_cached is None or (now - self._base_ratio_ts) >= refresh_sec):
            baseline = self.atr_ratio_baseline(klines)
            if baseline > 0.0001:
                self._base_ratio_cached = baseline
                self._base_ratio_ts = now
                logger.info(f"基准波动率更新: {baseline * 100:.4f}%")
            else:
                current_price = klines[-1]["close"] if klines else 3000
                if current_price > 4000:
                    default_baseline = 0.0010
                elif current_price > 3500:
                    default_baseline = 0.0012
                elif current_price > 3000:
                    default_baseline = 0.0015
                elif current_price > 2500:
                    default_baseline = 0.0018
                else:
                    default_baseline = 0.0020

                self._base_ratio_cached = default_baseline
                self._base_ratio_ts = now
                logger.warning(
                    f"使用价格动态默认基准: {default_baseline * 100:.4f}% (价格=${current_price:.0f})"
                )

        return self._base_ratio_cached

    @staticmethod
    def _clamp(x: float, lo: float, hi: float) -> float:
        return max(lo, min(hi, x))

    def dynamic_thresholds(self, atr_ratio: float, base_ratio: float):
        """
        ✅ entry/tp/sl 全部动态（修复版）：
        - vol_scale：atr_ratio/base_ratio 限幅
        - floor：方案一 (floor = clamp(min, k*base_ratio, max))
        - 最终阈值：max(floor, k * vol_scale * atr_ratio)
        """
        if atr_ratio <= 0:
            logger.warning(f"ATR比率异常: {atr_ratio}")
            atr_ratio = 0.001

        if base_ratio < 0.0005:
            base_ratio = max(0.001, atr_ratio * 1.2)
            logger.debug(f"基准太小，使用调整后的atr_ratio: {base_ratio * 100:.4f}%")

        if base_ratio > 0:
            raw_scale = atr_ratio / base_ratio
            vol_scale = self._clamp(raw_scale, self.cfg.vol_scale_min, self.cfg.vol_scale_max)
            logger.debug(
                f"vol_scale: {raw_scale:.2f} → {vol_scale:.2f} (atr={atr_ratio * 100:.3f}%, base={base_ratio * 100:.3f}%)"
            )
        else:
            vol_scale = 1.0
            logger.warning("基准无效，使用默认vol_scale=1.0")

        entry_floor_raw = self.cfg.entry_dev_floor_base_k * base_ratio
        entry_floor = self._clamp(entry_floor_raw, self.cfg.entry_dev_floor_min, self.cfg.entry_dev_floor_max)

        tp_floor_raw = self.cfg.tp_floor_base_k * base_ratio
        tp_floor = self._clamp(tp_floor_raw, self.cfg.tp_floor_min, self.cfg.tp_floor_max)

        sl_floor_raw = self.cfg.sl_floor_base_k * base_ratio
        sl_floor = self._clamp(sl_floor_raw, self.cfg.sl_floor_min, self.cfg.sl_floor_max)

        entry_dev_atr_part = self.cfg.entry_k * vol_scale * atr_ratio
        entry_dev = max(entry_floor, entry_dev_atr_part)

        tp_atr_part = self.cfg.tp_k * vol_scale * atr_ratio
        tp = max(tp_floor, tp_atr_part)

        sl_atr_part = self.cfg.sl_k * vol_scale * atr_ratio
        sl = max(sl_floor, sl_atr_part)

        entry_dev = max(entry_dev, self.cfg.entry_dev_floor_min)

        logger.info(
            f"动态阈值: atr={atr_ratio * 100:.4f}%, base={base_ratio * 100:.4f}%, "
            f"vol_scale={vol_scale:.2f}, floor={entry_floor * 100:.4f}%, "
            f"atr_part={entry_dev_atr_part * 100:.4f}%, 最终entry_dev={entry_dev * 100:.4f}%"
        )

        return entry_dev, tp, sl, vol_scale, entry_floor, tp_floor, sl_floor

    # ----------------- 账户/仓位 -----------------
    def get_assets_available(self) -> float:
        try:
            resp = self.contractAPI.get_assets_detail()[0]
            if resp.get("code") != 1000:
                return 0.0
            data = resp.get("data")
            if isinstance(data, dict):
                return float(data.get("available_balance", 0))
            if isinstance(data, list):
                for asset in data:
                    if asset.get("currency") == "USDT":
                        return float(asset.get("available_balance", 0))
            return 0.0
        except Exception as e:
            logger.error(f"余额查询异常: {e}")
            return 0.0

    def get_position_status(self) -> bool:
        try:
            resp = self.contractAPI.get_position(contract_symbol=self.cfg.contract_symbol)[0]
            if resp.get("code") != 1000:
                return False

            positions = resp.get("data", [])
            if not positions:
                self.pos = 0
                return True

            p = positions[0]
            self.pos = 1 if p["position_type"] == 1 else -1
            return True
        except Exception as e:
            logger.error(f"持仓查询异常: {e}")
            self.ding(f"持仓查询异常: {e}", error=True)
            return False

    def get_equity_proxy(self) -> float:
        return self.get_assets_available()

    def refresh_daily_baseline(self):
        today = datetime.date.today()
        if today != self.day_tag:
            self.day_tag = today
            self.day_start_equity = None
            self.trading_enabled = True
            self.ding(f"新的一天({today})：重置日内风控基准")

    def risk_kill_switch(self):
        self.refresh_daily_baseline()
        equity = self.get_equity_proxy()
        if equity <= 0:
            return

        if self.day_start_equity is None:
            self.day_start_equity = equity
            logger.info(f"日内权益基准设定：{equity:.2f} USDT")
            return

        pnl = (equity - self.day_start_equity) / self.day_start_equity
        if pnl <= -self.cfg.daily_loss_limit:
            self.trading_enabled = False
            self.ding(f"触发日止损：{pnl * 100:.2f}% -> 停机", error=True)

        if pnl >= self.cfg.daily_profit_cap:
            self.trading_enabled = False
            self.ding(f"达到日盈利封顶：{pnl * 100:.2f}% -> 停机")

    # ----------------- 下单 -----------------
    def calculate_size(self, price: float) -> int:
        bal = self.get_assets_available()
        if bal < 10:
            return 0

        margin = bal * self.cfg.risk_percent
        lev = int(self.cfg.leverage)

        # ⚠️ 沿用你的原假设：1张≈0.001ETH
        size = int((margin * lev) / (price * 0.001))
        size = max(self.cfg.min_size, size)
        size = min(self.cfg.max_size, size)
        return size

    def place_market_order(self, side: int, size: int) -> bool:
        if size <= 0:
            return False

        client_order_id = f"mr_{int(time.time())}_{uuid.uuid4().hex[:8]}"
        try:
            resp = self.contractAPI.post_submit_order(
                contract_symbol=self.cfg.contract_symbol,
                client_order_id=client_order_id,
                side=side,
                mode=1,
                type="market",
                leverage=self.cfg.leverage,
                open_type=self.cfg.open_type,
                size=size
            )[0]

            logger.info(f"order_resp: {resp}")

            if resp.get("code") == 1000:
                return True

            self.ding(f"下单失败: {resp}", error=True)
            return False

        except APIException as e:
            logger.error(f"API下单异常: {e}")
            self.ding(f"API下单异常: {e}", error=True)
            return False

        except Exception as e:
            logger.error(f"下单未知异常: {e}")
            self.ding(f"下单未知异常: {e}", error=True)
            return False

    def close_position_all(self):
        if self.pos == 1:
            ok = self.place_market_order(3, 999999)
            if ok:
                self.pos = 0
        elif self.pos == -1:
            ok = self.place_market_order(2, 999999)
            if ok:
                self.pos = 0

    # ----------------- 止损后机制 -----------------
    def _reentry_penalty_active(self, dev: float, entry_dev: float) -> bool:
        """检查是否需要应用重新入场惩罚（你原逻辑保留）"""
        if self.last_sl_dir == 0:
            return False

        if (time.time() - self.last_sl_ts) > self.cfg.reentry_penalty_max_sec:
            self.last_sl_dir = 0
            return False

        reset_band = max(self.cfg.reset_band_floor, self.cfg.reset_band_k * entry_dev)
        if abs(dev) <= reset_band:
            self.last_sl_dir = 0
            return False

        return True

    def _post_sl_dynamic_mult(self) -> float:
        """计算止损后SL放宽倍数"""
        if self.post_sl_dir == 0:
            return 1.0

        if (time.time() - self.post_sl_ts) > self.cfg.post_sl_sl_max_sec:
            self.post_sl_dir = 0
            self.post_sl_vol_scale = 1.0
            return 1.0

        raw = 1.0 + self.cfg.post_sl_vol_alpha * (self.post_sl_vol_scale - 1.0)
        raw = max(1.0, raw)  # 不缩小止损，只放宽
        return max(self.cfg.post_sl_mult_min, min(self.cfg.post_sl_mult_max, raw))

    # ----------------- 交易逻辑 -----------------
    def in_cooldown(self) -> bool:
        """检查是否在冷却期内（新增：止损后可额外延长冷却，用于抑制反手连环）"""
        cd = self.cfg.cooldown_sec_after_exit
        if self.last_sl_ts and (time.time() - self.last_sl_ts) < self.cfg.reentry_penalty_max_sec:
            cd += max(0, self.cfg.cooldown_sec_after_sl_extra)
        return (time.time() - self.last_exit_ts) < cd

    def _adx_blocks_entry(self, adx_val: float, plus_di: float, minus_di: float, want_dir: int) -> bool:
        """
        ADX 趋势过滤：
        - want_dir: 1=想开多, -1=想开空
        """
        if not self.cfg.enable_adx_filter:
            return False
        if adx_val < self.cfg.adx_threshold:
            return False

        if self.cfg.adx_mode == "block_all":
            return True

        # block_countertrend：只禁止逆 DI 方向
        # 上升趋势：+DI > -DI => 禁止开空
        # 下降趋势：-DI > +DI => 禁止开多
        if plus_di > minus_di and want_dir == -1:
            return True
        if minus_di > plus_di and want_dir == 1:
            return True
        return False

    def maybe_enter(self, price: float, ema_value: float, entry_dev: float,
                    adx_val: float, plus_di: float, minus_di: float):
        """检查并执行入场（新增：ADX趋势过滤，防止趋势期逆势反复开仓）"""
        if self.pos != 0:
            return
        if self.in_cooldown():
            return

        dev = (price - ema_value) / ema_value if ema_value else 0.0
        size = self.calculate_size(price)
        if size <= 0:
            return

        penalty_active = self._reentry_penalty_active(dev, entry_dev)

        long_th = -entry_dev
        short_th = entry_dev

        if penalty_active:
            if self.last_sl_dir == 1:
                long_th = -entry_dev * self.cfg.reentry_penalty_mult
                logger.info(f"多头止损后惩罚生效: long_th={long_th * 100:.3f}%")
            elif self.last_sl_dir == -1:
                short_th = entry_dev * self.cfg.reentry_penalty_mult
                logger.info(f"空头止损后惩罚生效: short_th={short_th * 100:.3f}%")

        logger.info(
            f"入场检查: price={price:.2f}, ema={ema_value:.2f}, dev={dev * 100:.3f}% "
            f"(entry_dev={entry_dev * 100:.3f}%, long_th={long_th * 100:.3f}%, short_th={short_th * 100:.3f}%) "
            f"ADX={adx_val:.2f} +DI={plus_di:.2f} -DI={minus_di:.2f} "
            f"size={size}, penalty={penalty_active}, last_sl_dir={self.last_sl_dir}"
        )

        # 先判断信号，再用 ADX 过滤（这样日志更直观）
        if dev <= long_th:
            if self._adx_blocks_entry(adx_val, plus_di, minus_di, want_dir=1):
                logger.warning(
                    f"ADX过滤：趋势期禁止逆势开多（ADX={adx_val:.2f}, +DI={plus_di:.2f}, -DI={minus_di:.2f}）"
                )
                return

            if self.place_market_order(1, size):
                self.pos = 1
                self.entry_price = price
                self.entry_ts = time.time()
                self.ding(f"✅开多：dev={dev * 100:.3f}% size={size} entry={price:.2f}")

        elif dev >= short_th:
            if self._adx_blocks_entry(adx_val, plus_di, minus_di, want_dir=-1):
                logger.warning(
                    f"ADX过滤：趋势期禁止逆势开空（ADX={adx_val:.2f}, +DI={plus_di:.2f}, -DI={minus_di:.2f}）"
                )
                return

            if self.place_market_order(4, size):
                self.pos = -1
                self.entry_price = price
                self.entry_ts = time.time()
                self.ding(f"✅开空：dev={dev * 100:.3f}% size={size} entry={price:.2f}")

    def maybe_exit(self, price: float, tp: float, sl: float, vol_scale: float):
        """检查并执行出场"""
        if self.pos == 0 or self.entry_price is None or self.entry_ts is None:
            return

        hold = time.time() - self.entry_ts

        if self.pos == 1:
            pnl = (price - self.entry_price) / self.entry_price
        else:
            pnl = (self.entry_price - price) / self.entry_price

        sl_mult = 1.0
        if self.post_sl_dir == self.pos and self.post_sl_dir != 0:
            sl_mult = self._post_sl_dynamic_mult()
        effective_sl = sl * sl_mult

        if pnl >= tp:
            self.close_position_all()
            self.ding(f"🎯止盈：pnl={pnl * 100:.3f}% price={price:.2f} tp={tp * 100:.3f}%")
            self.entry_price, self.entry_ts = None, None
            self.last_exit_ts = time.time()

        elif pnl <= -effective_sl:
            sl_dir = self.pos
            self.close_position_all()
            self.ding(
                f"🛑止损：pnl={pnl * 100:.3f}% price={price:.2f} "
                f"sl={sl * 100:.3f}% effective_sl={effective_sl * 100:.3f}%(×{sl_mult:.2f})",
                error=True
            )

            # 记录止损方向与时间
            self.last_sl_dir = sl_dir
            self.last_sl_ts = time.time()

            self.post_sl_dir = sl_dir
            self.post_sl_ts = time.time()
            self.post_sl_vol_scale = float(vol_scale)

            self.entry_price, self.entry_ts = None, None
            self.last_exit_ts = time.time()

        elif hold >= self.cfg.max_hold_sec:
            self.close_position_all()
            self.ding(f"⏱超时：hold={int(hold)}s pnl={pnl * 100:.3f}% price={price:.2f}")
            self.entry_price, self.entry_ts = None, None
            self.last_exit_ts = time.time()

    def notify_status_throttled(self, price: float, ema_value: float, dev: float, bal: float,
                                atr_ratio: float, base_ratio: float, vol_scale: float,
                                entry_dev: float, tp: float, sl: float,
                                entry_floor: float, tp_floor: float, sl_floor: float,
                                adx_val: float, plus_di: float, minus_di: float):
        """限频状态通知"""
        now = time.time()
        if (now - self._last_status_notify_ts) < self.cfg.status_notify_sec:
            return
        self._last_status_notify_ts = now

        direction_str = "多" if self.pos == 1 else ("空" if self.pos == -1 else "无")
        penalty_active = self._reentry_penalty_active(dev, entry_dev)

        sl_mult = 1.0
        if self.pos != 0 and self.post_sl_dir == self.pos:
            sl_mult = self._post_sl_dynamic_mult()

        base_age = int(now - self._base_ratio_ts) if self._base_ratio_ts else -1

        msg = (
            f"【BitMart {self.cfg.contract_symbol}｜1m均值回归(动态阈值+ADX过滤)】\n"
            f"📊 状态：{direction_str}\n"
            f"💰 现价：{price:.2f} | EMA{self.cfg.ema_len}：{ema_value:.2f}\n"
            f"📈 偏离：{dev * 100:.3f}% (入场阈值：±{entry_dev * 100:.3f}%)\n"
            f"🌊 波动率：ATR比={atr_ratio * 100:.3f}% | 基准={base_ratio * 100:.3f}% | 缩放={vol_scale:.2f}\n"
            f"🧭 趋势：ADX={adx_val:.2f} | +DI={plus_di:.2f} | -DI={minus_di:.2f} "
            f"(阈值={self.cfg.adx_threshold:.1f}, 模式={self.cfg.adx_mode})\n"
            f"🎯 动态Floor：入场={entry_floor * 100:.3f}% | 止盈={tp_floor * 100:.3f}% | 止损={sl_floor * 100:.3f}%\n"
            f"💰 止盈/止损：{tp * 100:.3f}% / {sl * 100:.3f}% (盈亏比：{tp / sl:.2f})\n"
            f"🔄 基准刷新：{self.cfg.base_ratio_refresh_sec}s (已过={base_age}s)\n"
            f"⚠️ 止损同向加门槛：{'开启' if penalty_active else '关闭'} (方向={self.last_sl_dir})\n"
            f"💳 可用余额：{bal:.2f} USDT | 杠杆：{self.cfg.leverage}x\n"
            f"⏱️ 持仓限制：{self.cfg.max_hold_sec}s | 冷却：{self.cfg.cooldown_sec_after_exit}s"
        )
        self.ding(msg)

    def action(self):
        """主循环"""
        if not self.set_leverage():
            self.ding("杠杆设置失败，停止运行", error=True)
            return

        while True:
            now_dt = datetime.datetime.now()
            self.pbar.n = now_dt.second
            self.pbar.refresh()

            # 1. 获取K线数据
            klines = self.get_klines_cached()
            if not klines or len(klines) < (max(self.cfg.ema_len + 5, 2 * self.cfg.adx_len + 5)):
                logger.warning("K线数据不足，等待...")
                time.sleep(1)
                continue

            # 2. 计算技术指标
            last_k = klines[-1]
            closes = [k["close"] for k in klines[-(self.cfg.ema_len + 1):]]
            ema_value = self.ema(closes, self.cfg.ema_len)

            price = self.get_last_price(fallback_close=float(last_k["close"]))
            dev = (price - ema_value) / ema_value if ema_value else 0.0

            # 3. 计算波动率相关指标
            a = self.atr(klines, self.cfg.atr_len)
            atr_ratio = (a / price) if price > 0 else 0.0
            base_ratio = self.get_base_ratio_cached(klines)

            # 4. 计算动态阈值
            entry_dev, tp, sl, vol_scale, entry_floor, tp_floor, sl_floor = self.dynamic_thresholds(
                atr_ratio, base_ratio
            )

            # 5. 计算 ADX（新增）
            adx_val, plus_di, minus_di = self.adx(klines, self.cfg.adx_len)
            logger.info(f"ADX: {adx_val:.2f} (+DI={plus_di:.2f}, -DI={minus_di:.2f})")

            # 6. 风控检查
            self.risk_kill_switch()

            # 7. 获取持仓状态
            if not self.get_position_status():
                time.sleep(1)
                continue

            # 8. 检查交易是否启用
            if not self.trading_enabled:
                if self.pos != 0:
                    self.close_position_all()
                logger.warning("交易被禁用（风控触发），等待...")
                time.sleep(5)
                continue

            # 9. 检查危险市场
            if self.is_danger_market(klines, price):
                logger.warning("危险模式：高波动/大实体K，暂停开仓")
                self.maybe_exit(price, tp, sl, vol_scale)
                time.sleep(self.cfg.tick_refresh_sec)
                continue

            # 10. 执行交易逻辑
            self.maybe_exit(price, tp, sl, vol_scale)
            self.maybe_enter(price, ema_value, entry_dev, adx_val, plus_di, minus_di)

            # 11. 状态通知
            bal = self.get_assets_available()
            self.notify_status_throttled(
                price, ema_value, dev, bal,
                atr_ratio, base_ratio, vol_scale,
                entry_dev, tp, sl,
                entry_floor, tp_floor, sl_floor,
                adx_val, plus_di, minus_di
            )

            time.sleep(self.cfg.tick_refresh_sec)


if __name__ == "__main__":
    """
    Windows PowerShell:
      setx BITMART_API_KEY "你的key"
      setx BITMART_SECRET_KEY "你的secret"
      setx BITMART_MEMO "合约交易"
    重新打开终端再运行。

    Linux/macOS:
      export BITMART_API_KEY="你的key"
      export BITMART_SECRET_KEY="你的secret"
      export BITMART_MEMO "合约交易"
    """
    cfg = StrategyConfig()
    bot = BitmartFuturesMeanReversionBot(cfg)

    logger.remove()
    logger.add(lambda msg: tqdm.write(msg, end=""), level="INFO")

    try:
        bot.action()
    except KeyboardInterrupt:
        logger.info("程序被用户中断")
        bot.ding("🤖 策略已手动停止")
    except Exception as e:
        logger.error(f"程序异常退出: {e}")
        bot.ding(f"❌ 策略异常退出: {e}", error=True)
        raise