完成模型更新

2026-06-24 12:19:08 +08:00
parent ee8fc60eae
commit 6494f43ddd
22 changed files with 4029 additions and 271 deletions
@@ -50,5 +50,5 @@ RANK_MODEL = 'rank'
 TOP_MODEL = 'top'
 STACKING_MODEL = 'stacking'
-# Stacking 选股阈值
+# Stacking 选股阈值 (v6.7r3 最优阈值 0.331)
-STACKING_THRESHOLD = 0.35
+STACKING_THRESHOLD = 0.33
@@ -1,9 +1,10 @@
 """
-扩展特征 v3.4 (16维)
+扩展特征 v3.4 + v6.7新增 (20维: 16维 v3.4 + 4维 v6.7新增)
 """
 import numpy as np
 import pandas as pd
 from core.scoring.features.v3_2_features import _ols_slope
 from core.scoring.features.v3_3_features import _grid_touch_count
 from core.scoring.config import GRID_LOW, GRID_HIGH
@@ -104,6 +105,29 @@ def calculate_features_v3_4(ctx) -> pd.DataFrame:
        feat['wick_ratio_20d'] = np.mean(
            wick_len / np.where(total_len > 0, total_len, 1)) * 100
        # ── v6.7 新增 4 维特征 ──────────────────────────────
        # 53. vol_decay_5d: 近5日波动率 / 近20日波动率 (波动率用对数收益std)
        log_ret = np.diff(np.log(np.maximum(closes, 1e-10)))
        vol_5d = np.std(log_ret[-5:], ddof=1) if len(log_ret) >= 5 else 0
        vol_20d = np.std(log_ret[-20:], ddof=1) if len(log_ret) >= 20 else vol_5d
        feat['vol_decay_5d'] = float(vol_5d / vol_20d) if vol_20d > 0 else 0.0
        # 54. grid_touch_relative_10d: 10日振幅比 / 60日振幅比
        range_10d = float(np.max(highs[-10:]) - np.min(lows[-10:]))
        range_60d = float(np.max(highs[-60:]) - np.min(lows[-60:])) if len(highs) >= 60 else range_10d
        close_now = float(closes[-1])
        close_60d_mean = float(np.mean(closes[-60:])) if len(closes) >= 60 else close_now
        if close_now > 0 and close_60d_mean > 0 and range_60d > 0:
            feat['grid_touch_relative_10d'] = (range_10d / close_now) / (range_60d / close_60d_mean)
        else:
            feat['grid_touch_relative_10d'] = 0.0
        # 55. vol_decay_x_grid_balance: vol_decay × 网格均衡度
        feat['vol_decay_x_grid_balance'] = feat['vol_decay_5d'] * feat.get('grid_room_balance', 0.0)
        # 56. vol_decay_x_dist_lower: vol_decay × 下轨距离
        feat['vol_decay_x_dist_lower'] = feat['vol_decay_5d'] * feat.get('dist_to_grid_lower', 0.0)
        features[code] = feat
    return pd.DataFrame.from_dict(features, orient='index')
@@ -1,5 +1,5 @@
 """
-grid_seeker v6.6 三级模型推理管道
+grid_seeker v6.7r3 三级模型推理管道
 Rank → Top → Stacking → stacking_probability (最终排序)
 """
 import pickle
@@ -18,7 +18,7 @@ from core.logger import LogLevel, PrintLog
 # ============================================================
-# Rank 模型输入特征 (52维, v3.4, 直接从模型文件的 selected_features 读取)
+# Rank 模型输入特征 (56维 v6.7/v3.4, 同时支持 feat_names 和 selected_features)
 # ============================================================
 def _get_rank_features() -> list:
    import pickle
@@ -27,17 +27,26 @@ def _get_rank_features() -> list:
    with open(path, 'rb') as f:
        obj = pickle.load(f)
    if isinstance(obj, dict):
-        sf = obj.get('selected_features', [])
+        # v6.7r3 使用 feat_names, v6.6 使用 selected_features
        sf = obj.get('feat_names', []) or obj.get('selected_features', [])
        if sf:
            return sf
-    raise RuntimeError("无法从 rank.pkl 读取 selected_features")
+    raise RuntimeError("无法从 rank.pkl 读取 feat_names 或 selected_features")
 RANK_FEATURE_COLS = _get_rank_features()
 # Top/Stacking 模型只用 52 维基础特征（不含 v6.7 新增的4维）
 # v6.7 新增: vol_decay_5d, grid_touch_relative_10d, vol_decay_x_grid_balance, vol_decay_x_dist_lower
 _V67_NEW_FEATS = {
    'vol_decay_5d', 'grid_touch_relative_10d',
    'vol_decay_x_grid_balance', 'vol_decay_x_dist_lower'
 }
 BASE_52_COLS = [f for f in RANK_FEATURE_COLS if f not in _V67_NEW_FEATS]
 class GridSeekerPipeline:
    """
-    grid_seeker v6.6 三级模型评分管道。
+    grid_seeker v6.7r3 三级模型评分管道。
    Usage:
        engine = GridSeekerPipeline()
@@ -120,7 +129,7 @@ class GridSeekerPipeline:
            DataFrame indexed by stock_code, 含 stacking_probability / rank 等列,
            按 stacking_probability 降序排列
        """
-        PrintLog(LogLevel.INFO, f'[scorer] ===== grid_seeker v6.6 评分开始 ({trade_date}) =====')
+        PrintLog(LogLevel.INFO, f'[scorer] ===== grid_seeker v6.7r3 评分开始 ({trade_date}) =====')
        # 1. 特征工程
        pipeline = FeaturePipeline(trade_date)
@@ -140,16 +149,16 @@ class GridSeekerPipeline:
            self.rank_model, feature_df, RANK_FEATURE_COLS
        )
-        # 3. Stage 2: Top 模型 → top_elite_prob (53维 = 52 + rank_predicted_rounds)
+        # 3. Stage 2: Top 模型 → top_elite_prob (53维 = 52基础 + rank)
        PrintLog(LogLevel.INFO, '[scorer] Stage 2/3: Top 模型...')
-        top_cols = RANK_FEATURE_COLS + ['rank_predicted_rounds']
+        top_cols = BASE_52_COLS + ['rank_predicted_rounds']
        feature_df['top_elite_prob'] = self._predict_with_model(
            self.top_model, feature_df, top_cols
        )
-        # 4. Stage 3: Stacking 模型 → stacking_probability (54维 = 52 + rank + top)
+        # 4. Stage 3: Stacking 模型 → stacking_probability (55维 = 52基础 + rank + top)
        PrintLog(LogLevel.INFO, '[scorer] Stage 3/3: Stacking 模型...')
-        stk_cols = RANK_FEATURE_COLS + ['rank_predicted_rounds', 'top_elite_prob']
+        stk_cols = BASE_52_COLS + ['rank_predicted_rounds', 'top_elite_prob']
        feature_df['stacking_probability'] = self._predict_with_model(
            self.stacking_model, feature_df, stk_cols
        )
@@ -5,7 +5,7 @@ K线数据同步 — 个股日K + 指数日K
 线程锁: KlineStockSync / KlineIndexSync 各自内部锁
 """
 import pandas as pd
-from datetime import date, timedelta
+from datetime import date, datetime, timedelta
 from core.scoring.sync.base import BaseSync
 from core.scoring.models import KlineStock, KlineIndex
 from core.scoring.config import TRACKED_INDICES
@@ -111,6 +111,10 @@ class KlineStockSync(BaseSync):
        stock_code = full_code.split('.')[0]
        records = []
        for td in close_df.columns:
            # xtdata 返回的列名可能是字符串 'YYYYMMDD' 或 datetime，需统一转成 date
            if isinstance(td, str):
                td_date = datetime.strptime(td, '%Y%m%d').date()
            else:
                td_date = td.date() if hasattr(td, 'date') else td
            if start_date is not None and td_date <= start_date:
                self.stats['skipped'] = self.stats.get('skipped', 0) + 1
@@ -193,6 +197,9 @@ class KlineIndexSync(BaseSync):
        for full_code in close_df.index:
            index_code = full_code.split('.')[0]
            for td in close_df.columns:
                if isinstance(td, str):
                    td_date = datetime.strptime(td, '%Y%m%d').date()
                else:
                    td_date = td.date() if hasattr(td, 'date') else td
                if latest is not None and td_date <= latest:
                    self.stats['skipped'] = self.stats.get('skipped', 0) + 1
@@ -13,7 +13,7 @@ class SectorFeaturesSync(BaseSync):
    """行业聚合指数同步 — kline_stock + industry → sector_features_daily"""
    def _fetch(self, **kwargs):
-        """从数据库加载原始数据, 计算行业指数特征"""
+        """从数据库加载原始数据, 计算行业指数特征（分块处理避免内存溢出）"""
        PrintLog(LogLevel.INFO, '[sync] SectorFeatures: 加载原始数据...')
        # 1. 加载行业映射: code → industry_name
@@ -24,8 +24,17 @@ class SectorFeaturesSync(BaseSync):
        code_to_industry = {row['code']: row['industry_name'] for row in industries}
        PrintLog(LogLevel.INFO, f'[sync] SectorFeatures: {len(code_to_industry)} 条行业映射')
-        # 2. 加载 K 线数据
+        # 2. 分块加载 K 线数据，避免内存溢出
-        kline_rows = (KlineStock
+        # 聚合结果: {(trade_date, sector_name): [sum_pct_chg, sum_amp, count]}
        sector_daily_agg = {}  # key: (date, sector) -> {'ret_sum': float, 'amp_sum': float, 'count': int}
        CHUNK_SIZE = 50000
        last_date_per_stock = {}  # stock_code -> prev_close
        PrintLog(LogLevel.INFO, '[sync] SectorFeatures: 分块处理K线数据...')
        chunk_num = 0
        while True:
            chunk_num += 1
            rows = list(KlineStock
                        .select(
                            KlineStock.stock_code,
                            KlineStock.trade_date,
@@ -35,38 +44,61 @@ class SectorFeaturesSync(BaseSync):
                            KlineStock.close,
                        )
                        .order_by(KlineStock.stock_code, KlineStock.trade_date)
                        .offset((chunk_num - 1) * CHUNK_SIZE)
                        .limit(CHUNK_SIZE)
                        .dicts())
            if not rows:
                break
-        if not kline_rows:
+            PrintLog(LogLevel.INFO, f'[sync] SectorFeatures: 处理块 {chunk_num} ({len(rows)} 行)...')
-            PrintLog(LogLevel.WARNING, '[sync] SectorFeatures: KlineStock 表为空')
+
            for row in rows:
                code = str(row['stock_code'])
                td = row['trade_date']
                open_p = float(row['open'])
                high = float(row['high'])
                low = float(row['low'])
                close = float(row['close'])
                sector = code_to_industry.get(code)
                if sector is None:
                    continue
                # 计算日收益率和振幅
                prev_close = last_date_per_stock.get(code)
                if prev_close is not None and prev_close > 0 and open_p > 0 and close > 0:
                    pct_chg = (close - prev_close) / prev_close * 100
                    amp = (high - low) / open_p * 100
                    key = (td, sector)
                    if key not in sector_daily_agg:
                        sector_daily_agg[key] = {'ret_sum': 0.0, 'amp_sum': 0.0, 'count': 0}
                    sector_daily_agg[key]['ret_sum'] += pct_chg
                    sector_daily_agg[key]['amp_sum'] += amp
                    sector_daily_agg[key]['count'] += 1
                last_date_per_stock[code] = close
        if not sector_daily_agg:
            PrintLog(LogLevel.WARNING, '[sync] SectorFeatures: 无有效K线数据')
            return None
-        df = pd.DataFrame(kline_rows)
+        PrintLog(LogLevel.INFO, f'[sync] SectorFeatures: 聚合完成, {len(sector_daily_agg)} 个行业-日组合')
        df['trade_date'] = pd.to_datetime(df['trade_date'])
        PrintLog(LogLevel.INFO, f'[sync] SectorFeatures: {len(df)} 条K线数据')
-        # 3. 映射行业
+        # 3. 构建聚合 DataFrame
-        df['sector_name'] = df['stock_code'].map(code_to_industry)
+        agg_data = []
-        df = df.dropna(subset=['sector_name'])
+        for (td, sector), vals in sector_daily_agg.items():
            agg_data.append({
                'trade_date': td,
                'sector_name': sector,
                'sector_ret': vals['ret_sum'] / vals['count'],
                'sector_amplitude': vals['amp_sum'] / vals['count'],
            })
        agg = pd.DataFrame(agg_data)
        agg = agg.sort_values(['sector_name', 'trade_date'])
        agg['trade_date'] = pd.to_datetime(agg['trade_date'])
        PrintLog(LogLevel.INFO, f'[sync] SectorFeatures: {len(agg)} 行, {agg["sector_name"].nunique()} 个行业')
-        # 4. 逐股计算日收益率和振幅
+        # 4. 构建行业指数 (基值=100)
        df = df.sort_values(['stock_code', 'trade_date'])
        df['prev_close'] = df.groupby('stock_code')['close'].shift(1)
        df['pct_chg'] = (df['close'] - df['prev_close']) / df['prev_close'] * 100
        df['amplitude'] = (df['high'] - df['low']) / df['open'] * 100
        # 清理无效值
        df = df.dropna(subset=['pct_chg', 'amplitude'])
        # 5. 按行业+日期聚合
        agg = (df.groupby(['trade_date', 'sector_name'])
               .agg(
                   sector_ret=('pct_chg', 'mean'),
                   sector_amplitude=('amplitude', 'mean'),
               )
               .reset_index())
        # 6. 构建行业指数 (基值=100)
        agg = agg.sort_values(['sector_name', 'trade_date'])
        agg['sector_index'] = agg.groupby('sector_name')['sector_ret'].transform(
            lambda x: (1 + x / 100).cumprod() * 100
@@ -77,7 +109,7 @@ class SectorFeaturesSync(BaseSync):
            first_val = group['sector_index'].iloc[0]
            agg.loc[idx, 'sector_index'] = group['sector_index'] / first_val * 100
-        # 7. 计算 EMA 均线
+        # 5. 计算 EMA 均线
        agg['ema10'] = (agg.groupby('sector_name')['sector_index']
                        .transform(lambda x: x.ewm(span=10, min_periods=1).mean()))
        agg['ema20'] = (agg.groupby('sector_name')['sector_index']
@@ -85,7 +117,7 @@ class SectorFeaturesSync(BaseSync):
        agg['ema200'] = (agg.groupby('sector_name')['sector_index']
                         .transform(lambda x: x.ewm(span=200, min_periods=1).mean()))
-        # 8. 趋势评分: close>ema200 得1分 + ema10>ema20 得1分
+        # 6. 趋势评分: close>ema200 得1分 + ema10>ema20 得1分
        agg['score'] = (
            (agg['sector_index'] > agg['ema200']).astype(int) +
            (agg['ema10'] > agg['ema20']).astype(int)
@@ -173,6 +173,9 @@ class SFGridStrategy:
                    # 检查是否已存在同 remark 的卖单（避免重复挂单）
                    if not any(o.order_remark == sell_remark for o in orders):
                        # 卖单价格超过涨停价 → 今日无法成交，跳过下单
                        # 防御性检查：若属性未初始化（初始化顺序导致），先获取
                        if not hasattr(self, 'todayUpStopPrice') or self.todayUpStopPrice is None:
                            self.todayUpStopPrice = qmtv.dailyUpStop(self.tradeTarget.stock_code)  # type: ignore
                        if sellPrice > self.todayUpStopPrice:
                            PrintLog(LogLevel.INFO,
                                     f'|- 标的[{self.tradeTarget.targetName()}] '
@@ -207,6 +210,9 @@ class SFGridStrategy:
                    # 检查是否已存在同 remark 的买单（避免重复挂单）
                    if not any(o.order_remark == buy_remark for o in orders):
                        # 买单价格低于跌停价 → 今日无法成交，跳过下单
                        # 防御性检查：若属性未初始化（初始化顺序导致），先获取
                        if not hasattr(self, 'todayDownStopPrice') or self.todayDownStopPrice is None:
                            self.todayDownStopPrice = qmtv.dailyDownStop(self.tradeTarget.stock_code)  # type: ignore
                        if buyPrice < self.todayDownStopPrice:
                            PrintLog(LogLevel.INFO,
                                     f'|- 标的[{self.tradeTarget.targetName()}] '
@@ -324,9 +324,9 @@ class _GridPanel:
        return self._col
    def _rebuild(self):
-        # (width, expand): 0=固定宽, >0=弹性比重
+        # (width, expand): 0=固定宽, >0=弹性比重; 新增排名列(50px)
-        _C = [(35, 0), (0, 2), (70, 1), (0, 2), (50, 1), (55, 1), (60, 1), (0, 1)]
+        _C = [(35, 0), (0, 2), (70, 1), (0, 2), (50, 1), (55, 1), (50, 0), (60, 1), (0, 1)]
-        H = ["ID", "股票", "市场价", "网格基准", "持仓", "成本", "状态", "操作"]
+        H = ["ID", "股票", "市场价", "网格基准", "持仓", "成本", "排名", "状态", "操作"]
        header = []
        for h, (w, e) in zip(H, _C):
            if e > 0:
@@ -335,6 +335,22 @@ class _GridPanel:
                header.append(ft.Container(_text(h, bold=True), width=w, padding=4))
        rows = [ft.Row(header, spacing=0), ft.Divider(height=1, color='#e0e0e0')]
        # 获取最新评分排名
        rank_map = {}  # stock_code -> score_rank
        try:
            from core.scoring.models import ScoringResult
            from peewee import fn
            latest_date = ScoringResult.select(fn.MAX(ScoringResult.trade_date)).scalar()
            if latest_date:
                scored = (ScoringResult
                         .select(ScoringResult.stock_code, ScoringResult.score_rank)
                         .where(ScoringResult.trade_date == latest_date)
                         .dicts())
                for r in scored:
                    rank_map[r['stock_code']] = r['score_rank']
        except Exception:
            pass
        for tid, t in self._data.tradeTargets.items():
            if t.strategy_type != STRATEGY_TYPE_GRID: continue
            pg = t.getPriceGrid()
@@ -358,6 +374,22 @@ class _GridPanel:
                ))
            gcell = ft.Row(gcells, spacing=3) if len(gcells) > 1 else gcells[0]
            # 排名列
            plain = t.stock_code.split('.')[0] if '.' in t.stock_code else t.stock_code
            rank = rank_map.get(plain, None)
            if rank is not None and rank <= 50:
                rank_str = f'#{rank}'
                rank_color = '#4CAF50'
            elif rank is not None and rank <= 100:
                rank_str = f'#{rank}'
                rank_color = '#2196F3'
            elif rank is not None:
                rank_str = f'#{rank}'
                rank_color = '#888888'
            else:
                rank_str = '—'
                rank_color = '#888888'
            # 操作按钮
            _ICON = 22  # 图标大小，比默认 18 好按
            if t.enabled:
@@ -385,6 +417,7 @@ class _GridPanel:
                          f'{mp:.3f}', gcell,
                          str(t.current_position),
                          f'{self._data.avgPrices.get(tid, 0):.3f}',
                          rank_str,
                          '▶运行中' if t.enabled else '⏸已暂停',
                          btns]
            row_cells = []
@@ -392,6 +425,8 @@ class _GridPanel:
                content = cells_text[i]
                if i == 2:  # 市场价用颜色
                    content = _text(cells_text[i], color=pcolor, bold=True)
                elif i == 6:  # 排名用颜色
                    content = _text(cells_text[i], color=rank_color, bold=True)
                elif isinstance(content, str):
                    content = _text(content)
                elif isinstance(content, list):
@@ -495,8 +530,25 @@ class _DrawerPanel:
    def _refresh_grid(self):
        # (width, expand): 0=固定宽, >0=弹性比重; 股票列 expand 自动填充剩余空间
-        _C = [(35, 0), (0, 1), (80, 0), (60, 0), (70, 0), (65, 0)]
+        # 新增: 排名列 (50px)
-        H = ["ID", "股票", "市场价", "持仓", "成本", "操作"]
+        _C = [(35, 0), (0, 1), (80, 0), (60, 0), (70, 0), (50, 0), (65, 0)]
        H = ["ID", "股票", "市场价", "持仓", "成本", "排名", "操作"]
        # 获取最新评分排名
        rank_map = {}  # stock_code -> score_rank
        try:
            from core.scoring.models import ScoringResult
            from peewee import fn
            latest_date = ScoringResult.select(fn.MAX(ScoringResult.trade_date)).scalar()
            if latest_date:
                rows = (ScoringResult
                        .select(ScoringResult.stock_code, ScoringResult.score_rank)
                        .where(ScoringResult.trade_date == latest_date)
                        .dicts())
                for r in rows:
                    rank_map[r['stock_code']] = r['score_rank']
        except Exception:
            pass  # 数据库未就绪时忽略
        def _cell(text, w, e, color=None):
            if e > 0:
@@ -509,15 +561,26 @@ class _DrawerPanel:
        for tid, t in self._data.tradeTargets.items():
            if t.strategy_type == STRATEGY_TYPE_GRID: continue
            mp = self._data.marketPrices.get(tid, 0) or 0
            plain = t.stock_code.split('.')[0] if '.' in t.stock_code else t.stock_code
            rank = rank_map.get(plain, None)
            rank_str = f"#{rank}" if rank is not None else "—"
            # 排名颜色: top50 绿色, top100 蓝色, 其他灰色
            if rank is not None and rank <= 50:
                rank_color = '#4CAF50'
            elif rank is not None and rank <= 100:
                rank_color = '#2196F3'
            else:
                rank_color = '#888888'
            cells = [
                _cell(str(tid), *_C[0]),
                _cell(f'{t.stock_code} {t.stock_name}', *_C[1]),
                _cell(f'{mp:.3f}', *_C[2]),
                _cell(str(t.current_position), *_C[3]),
                _cell(f'{self._data.avgPrices.get(tid, 0):.3f}', *_C[4]),
                _cell(rank_str, *_C[5], color=rank_color),
                ft.Container(ft.IconButton(ft.Icons.SETTINGS, icon_size=20, tooltip="网格配置",
                                            on_click=lambda e, tt=t: self._dialogs.open_config(tt)),
-                             width=_C[5][0], padding=0),
+                             width=_C[6][0], padding=0),
            ]
            row = ft.Row(cells, spacing=0)
            rows.append(ft.Container(row, padding=ft.Padding(0, 2, 0, 2)))
@@ -1,197 +0,0 @@
 # v6.6 模型发布说明
 **版本**: v6.6
 **发布日期**: 2026-06-16
 **状态**: 生产就绪
 ---
 ## 一、模型架构
 v6.6 采用 **Stacking Calibrated** 三层融合架构：
 ```
 输入特征 (52维 v3.4)
       │
       ▼
 ┌─────────────────┐    ┌─────────────────┐
 │  Rank 模型       │    │  Top 模型        │
 │  LGBMRegressor   │    │  LGBMClassifier  │
 │  预测网格轮回次数 │    │  分类精英股票     │
 │  CV MAE: 0.2053  │    │  CV PR-AUC: 0.53│
 │  CV R²: 0.2258   │    │                  │
 └────────┬────────┘    └────────┬────────┘
        │                      │
        └──────────┬───────────┘
                   ▼
        ┌─────────────────────┐
        │  Stacking 模型       │
        │  LGBMClassifier      │
        │  CV PR-AUC: 0.8207  │
        │  最优阈值: 0.35       │
        └──────────┬──────────┘
                   ▼
             最终 top 概率
 ```
 ---
 ## 二、训练数据
 | 指标 | 数值 |
 |------|------|
 | 特征版本 | v3.4 (52维) |
 | 原始股票数 | 4,358 |
 | 过滤后股票数 | 1,533 |
 | 训练样本数 | 205,491 |
 | 观察窗口 | 120天 |
 | 预测窗口 | 60天 |
 | 零触碰率 | 71.1% |
 | Elite率 | 20.5% |
 ---
 ## 三、模型性能
 ### Rank模型 (回归)
 | 指标 | 数值 |
 |------|------|
 | CV MAE | 0.2053 |
 | CV R² | 0.2258 |
 | 最优参数 | num_leaves=63, min_child_samples=30, max_depth=7 |
 ### Top模型 (分类)
 | 指标 | 数值 |
 |------|------|
 | CV PR-AUC | 0.5333 |
 | 最优参数 | num_leaves=63, min_child_samples=30, max_depth=-1 |
 ### Stacking融合模型
 | 指标 | 数值 |
 |------|------|
 | CV PR-AUC | **0.8207** |
 | 最优阈值 | 0.35 |
 | 最优参数 | num_leaves=31, min_child_samples=20, max_depth=-1 |
 ---
 ## 四、Top 10 重要特征
 ### Rank模型
 | 排名 | 特征 | 重要性% |
 |------|------|---------|
 | 1 | dist_to_grid_upper | 4.38% |
 | 2 | dist_to_grid_lower | 4.26% |
 | 3 | price_cv | 4.09% |
 | 4 | amount_mean_20d | 3.98% |
 | 5 | range_compression_20d | 3.48% |
 ### Stacking融合
 | 排名 | 特征 | 重要性% |
 |------|------|---------|
 | 1 | rank_predicted_rounds | 20.87% |
 | 2 | top_elite_prob | 19.38% |
 | 3 | ma20_deviation_pct | 5.35% |
 | 4 | atr_pct | 4.93% |
 | 5 | dist_to_grid_lower | 3.75% |
 ---
 ## 五、周评分回测结果 (2023-04 ~ 2026-04)
 | 指标 | v6.6 数值 |
 |------|-----------|
 | **总收益率** | **+36.79%** |
 | 年化夏普 | 0.8494 |
 | 最大回撤 | -12.70% |
 | 胜率(周) | 48.7% |
 | 交易次数 | 277 (147买, 130卖) |
 | 最大持仓 | 10只 |
 ### 年度收益
 | 年度 | 收益 |
 |------|------|
 | 2023 | +2.09% |
 | 2024 | +28.34% |
 | 2025 | +7.32% |
 | 2026 | +0.52% |
 ---
 ## 六、与v6.5对比
 | 指标 | v6.5 | v6.6 | 变化 |
 |------|------|------|------|
 | Stacking PR-AUC | 0.5234 | **0.8207** | +57% |
 | 特征数 | 37维 | **52维** | +15维 |
 | Top PR-AUC | 0.3923 | 0.5333 | +36% |
 | 回测收益率 | 59.82% | 36.79%* | - |
 | 回测最大回撤 | -33.47% | **-12.70%** | -62% |
 \* v6.6使用周评分回测（157周），v6.5使用日评分回测（723天），粒度不同
 ---
 ## 七、模型文件
 | 文件 | 说明 |
 |------|------|
 | `rank.pkl` | Rank模型 (预测网格轮回次数) |
 | `top.pkl` | Top模型 (分类精英股票) |
 | `stacking.pkl` | Stacking融合模型 |
 | `rank_feature_importance.csv` | Rank特征重要性 |
 | `top_feature_importance.csv` | Top特征重要性 |
 | `stacking_feature_importance.csv` | Stacking特征重要性 |
 | `training_summary.json` | 训练摘要 |
 ---
 ## 八、使用方式
 ```python
 import pickle
 import numpy as np
 # 加载模型
 with open("rank.pkl", "rb") as f:
    rank_md = pickle.load(f)
 with open("top.pkl", "rb") as f:
    top_md = pickle.load(f)
 with open("stacking.pkl", "rb") as f:
    stacking_md = pickle.load(f)
 # 预测
 rank_pred = rank_md["model"].predict(features)
 top_prob = top_md["model"].predict_proba(features)[:, 1]
 stacking_prob = stacking_md["model"].predict_proba(features)[:, 1]
 # 选股
 threshold = 0.35  # stacking最优阈值
 top_picks = scores[scores["stacking_prob"] >= threshold]
 ```
 ---
 ## 九、Registry配置
 ```yaml
 # grid_seeker/registry.yaml
 production:
  version: v6.6
  architecture: stacking_calibrated
  models:
    stacking: versions/6.6/output/stacking.pkl
    rank: versions/6.6/output/rank.pkl
    top: versions/6.6/output/top.pkl
  feature_version: v3.4
  training_samples: 205491
 ```
 ---
 ## 十、注意事项
 1. **特征版本**: 必须使用 v3.4 特征（52维），与v6.5/v6.4不兼容
 2. **Stacking阈值**: 推荐使用 0.35 作为选股阈值
 3. **持股上限**: 建议不超过10只
 4. **价格区间**: 适合7-10元区间股票
@@ -0,0 +1,727 @@
 date,cash,holding_market_value,total_asset
 2023-05-04,28289.86,33283.08,61572.94
 2023-05-05,29353.86,32311.96,61665.82
 2023-05-08,29553.86,32215.16,61769.020000000004
 2023-05-09,27953.86,32879.84,60833.7
 2023-05-10,26153.86,35374.479999999996,61528.34
 2023-05-11,29753.86,33225.380000000005,62979.240000000005
 2023-05-12,24353.86,37200.04,61553.9
 2023-05-15,25667.86,35345.6,61013.46
 2023-05-16,20867.86,38643.44,59511.3
 2023-05-17,17667.86,42281.4,59949.26
 2023-05-18,17667.86,43420.1,61087.96
 2023-05-19,19999.86,40128.84,60128.7
 2023-05-22,19136.34,40407.96,59544.3
 2023-05-23,19136.34,39661.66,58798.0
 2023-05-24,19136.34,39738.92,58875.259999999995
 2023-05-25,17736.34,40531.479999999996,58267.81999999999
 2023-05-26,17736.34,41098.4,58834.740000000005
 2023-05-29,14136.34,44736.28,58872.619999999995
 2023-05-30,14336.34,45939.42,60275.759999999995
 2023-05-31,14336.34,45642.6,59978.94
 2023-06-01,20433.92,41484.86,61918.78
 2023-06-02,22233.92,40253.44,62487.36
 2023-06-05,20683.899999999998,42477.36,63161.259999999995
 2023-06-06,24283.899999999998,37810.38,62094.28
 2023-06-07,22683.899999999998,40119.560000000005,62803.46000000001
 2023-06-08,21083.899999999998,40887.58,61971.479999999996
 2023-06-09,23083.899999999998,39432.4,62516.3
 2023-06-12,23083.899999999998,39468.780000000006,62552.68000000001
 2023-06-13,23083.899999999998,40270.58,63354.479999999996
 2023-06-14,24883.899999999998,38792.08,63675.979999999996
 2023-06-15,26883.899999999998,36531.32000000001,63415.22
 2023-06-16,27171.899999999998,36662.04,63833.94
 2023-06-19,30577.359999999997,33518.72,64096.08
 2023-06-20,28977.359999999997,35500.700000000004,64478.06
 2023-06-21,27177.359999999997,35306.82,62484.17999999999
 2023-06-26,25315.46,35218.7,60534.159999999996
 2023-06-27,20515.46,40464.58,60980.04
 2023-06-28,20515.46,39961.2,60476.659999999996
 2023-06-29,20515.46,40167.38,60682.84
 2023-06-30,20515.46,40323.66,60839.12
 2023-07-03,28555.76,32259.06,60814.82
 2023-07-04,28555.76,32140.9,60696.66
 2023-07-05,28555.76,31760.359999999997,60316.119999999995
 2023-07-06,28555.76,32029.4,60585.16
 2023-07-07,27155.76,33151.0,60306.759999999995
 2023-07-10,30335.76,30607.5,60943.259999999995
 2023-07-11,30335.76,29987.239999999998,60323.0
 2023-07-12,27135.76,32715.800000000003,59851.56
 2023-07-13,28735.76,32062.980000000003,60798.740000000005
 2023-07-14,29091.76,31705.800000000003,60797.56
 2023-07-17,27259.08,33179.8,60438.880000000005
 2023-07-18,27259.08,33110.299999999996,60369.38
 2023-07-19,27259.08,33071.82,60330.9
 2023-07-20,27259.08,32716.16,59975.240000000005
 2023-07-21,27608.82,32509.1,60117.92
 2023-07-24,29909.840000000004,30081.24,59991.08
 2023-07-25,29909.840000000004,30276.039999999997,60185.880000000005
 2023-07-26,31797.840000000004,27988.0,59785.840000000004
 2023-07-27,31797.840000000004,27916.859999999997,59714.7
 2023-07-28,30397.840000000004,29960.780000000002,60358.62000000001
 2023-07-31,31034.120000000003,29691.679999999997,60725.8
 2023-08-01,31326.78,29226.62,60553.399999999994
 2023-08-02,33126.78,27590.860000000004,60717.64
 2023-08-03,31326.78,29124.66,60451.44
 2023-08-04,31326.78,29340.219999999998,60667.0
 2023-08-07,28402.46,32404.219999999998,60806.67999999999
 2023-08-08,28402.46,32277.72,60680.18
 2023-08-09,28602.46,31887.659999999996,60490.119999999995
 2023-08-10,28602.46,32160.539999999997,60763.0
 2023-08-11,28602.46,31713.8,60316.259999999995
 2023-08-14,30216.02,30586.119999999995,60802.14
 2023-08-15,30216.02,30597.48,60813.5
 2023-08-16,32216.02,28753.199999999997,60969.22
 2023-08-17,30416.02,31062.44,61478.46
 2023-08-18,33136.700000000004,28257.34,61394.04000000001
 2023-08-21,31336.700000000004,30055.36,61392.060000000005
 2023-08-22,29536.700000000004,32326.899999999998,61863.600000000006
 2023-08-23,29536.700000000004,31548.5,61085.200000000004
 2023-08-24,29536.700000000004,31743.98,61280.68000000001
 2023-08-25,26136.700000000004,34300.62,60437.32000000001
 2023-08-28,29708.780000000006,31279.519999999997,60988.3
 2023-08-29,30196.780000000006,32665.699999999997,62862.48
 2023-08-30,32196.780000000006,31247.879999999997,63444.66
 2023-08-31,34196.780000000006,29057.66,63254.44
 2023-09-01,33946.780000000006,29311.820000000003,63258.600000000006
 2023-09-04,34135.520000000004,29560.14,63695.66
 2023-09-05,34509.780000000006,29185.22,63695.00000000001
 2023-09-06,32709.780000000006,30561.239999999998,63271.020000000004
 2023-09-07,33082.920000000006,30092.64,63175.560000000005
 2023-09-08,29682.920000000006,33243.98,62926.90000000001
 2023-09-11,28129.100000000006,35377.44,63506.54000000001
 2023-09-12,30129.100000000006,33934.32,64063.420000000006
 2023-09-13,31929.100000000006,32089.72,64018.82000000001
 2023-09-14,31929.100000000006,31721.36,63650.46000000001
 2023-09-15,33929.100000000006,29926.239999999998,63855.340000000004
 2023-09-18,35984.3,28165.4,64149.700000000004
 2023-09-19,36370.3,27593.84,63964.14
 2023-09-20,36917.340000000004,27446.239999999998,64363.58
 2023-09-21,35117.340000000004,29112.16,64229.5
 2023-09-22,35117.340000000004,29932.44,65049.78
 2023-09-25,35673.340000000004,30188.22,65861.56
 2023-09-26,33873.340000000004,31843.26,65716.6
 2023-09-27,33873.340000000004,32595.0,66468.34
 2023-09-28,39873.340000000004,27709.4,67582.74
 2023-10-09,41247.340000000004,26008.479999999996,67255.82
 2023-10-10,43605.08,24212.319999999996,67817.4
 2023-10-11,43805.08,23993.4,67798.48000000001
 2023-10-12,44083.08,23915.480000000003,67998.56
 2023-10-13,44083.08,23683.52,67766.6
 2023-10-16,43689.08,23793.599999999995,67482.68
 2023-10-17,42089.08,25216.94,67306.02
 2023-10-18,38489.08,28344.4,66833.48000000001
 2023-10-19,36689.08,29789.940000000002,66479.02
 2023-10-20,33289.08,32655.699999999997,65944.78
 2023-10-23,31442.520000000004,33523.66,64966.18000000001
 2023-10-24,29642.520000000004,36040.58,65683.1
 2023-10-25,33642.520000000004,33017.58,66660.1
 2023-10-26,34288.520000000004,33149.32,67437.84
 2023-10-27,36088.520000000004,31940.920000000002,68029.44
 2023-10-30,37285.62,32011.34,69296.96
 2023-10-31,39938.9,29558.3,69497.2
 2023-11-01,38429.700000000004,30972.719999999994,69402.42
 2023-11-02,40429.700000000004,28891.5,69321.20000000001
 2023-11-03,40429.700000000004,29377.8,69807.5
 2023-11-06,41176.26,29848.48,71024.74
 2023-11-07,45432.26,26553.34,71985.6
 2023-11-08,45782.26,26698.7,72480.96
 2023-11-09,48079.740000000005,24910.94,72990.68000000001
 2023-11-10,46479.740000000005,25900.26,72380.0
 2023-11-13,48501.3,24292.92,72794.22
 2023-11-14,46701.3,26683.64,73384.94
 2023-11-15,46957.340000000004,26097.679999999997,73055.02
 2023-11-16,47157.340000000004,26279.88,73437.22
 2023-11-17,49341.340000000004,24523.559999999998,73864.9
 2023-11-20,51584.740000000005,23124.5,74709.24
 2023-11-21,51940.740000000005,22778.399999999998,74719.14
 2023-11-22,49054.740000000005,26431.32,75486.06
 2023-11-23,50182.740000000005,25830.2,76012.94
 2023-11-24,49700.740000000005,27047.760000000002,76748.5
 2023-11-27,52958.340000000004,24999.68,77958.02
 2023-11-28,46244.340000000004,30826.699999999997,77071.04000000001
 2023-11-29,41844.340000000004,33468.94,75313.28
 2023-11-30,37244.340000000004,37343.26,74587.6
 2023-12-01,33892.340000000004,39456.38,73348.72
 2023-12-04,29451.340000000004,43211.36,72662.70000000001
 2023-12-05,36811.340000000004,40589.1,77400.44
 2023-12-06,43811.340000000004,34448.659999999996,78260.0
 2023-12-07,49721.340000000004,30807.46,80528.8
 2023-12-08,50831.340000000004,30194.3,81025.64
 2023-12-11,41721.94,37976.44,79698.38
 2023-12-12,41967.94,38844.28,80812.22
 2023-12-13,41967.94,40205.56,82173.5
 2023-12-14,48057.94,35801.66,83859.6
 2023-12-15,52387.94,31884.0,84271.94
 2023-12-18,54779.94,31022.0,85801.94
 2023-12-19,56379.94,28964.0,85343.94
 2023-12-20,58379.94,28260.0,86639.94
 2023-12-21,60909.94,24672.0,85581.94
 2023-12-22,52909.94,32946.0,85855.94
 2023-12-25,54909.94,32782.0,87691.94
 2023-12-26,54909.94,33254.0,88163.94
 2023-12-27,57197.94,31222.0,88419.94
 2023-12-28,55761.94,33084.0,88845.94
 2023-12-29,56387.94,33956.0,90343.94
 2024-01-02,62813.94,29938.0,92751.94
 2024-01-03,65205.94,27608.0,92813.94
 2024-01-04,63405.94,29272.0,92677.94
 2024-01-05,56405.94,34602.0,91007.94
 2024-01-08,58405.94,31830.0,90235.94
 2024-01-09,48405.94,40204.0,88609.94
 2024-01-10,43605.94,45832.0,89437.94
 2024-01-11,45405.94,44096.0,89501.94
 2024-01-12,42405.94,42438.0,84843.94
 2024-01-15,35605.94,46922.0,82527.94
 2024-01-16,34005.94,52684.0,86689.94
 2024-01-17,37205.94,49538.0,86743.94
 2024-01-18,39005.94,49904.0,88909.94
 2024-01-19,44405.94,46458.0,90863.94
 2024-01-22,45355.94,42936.0,88291.94
 2024-01-23,45355.94,44114.0,89469.94
 2024-01-24,47895.94,42334.0,90229.94
 2024-01-25,46695.94,42518.0,89213.94
 2024-01-26,43543.94,44782.0,88325.94
 2024-01-29,39113.94,44472.0,83585.94
 2024-01-30,33513.94,50754.0,84267.94
 2024-01-31,40113.94,46114.0,86227.94
 2024-02-01,38705.94,44274.0,82979.94
 2024-02-02,27305.940000000002,52690.0,79995.94
 2024-02-05,22445.940000000002,55230.0,77675.94
 2024-02-06,27245.940000000002,56198.0,83443.94
 2024-02-07,36845.94,46666.0,83511.94
 2024-02-08,31419.940000000002,51926.0,83345.94
 2024-02-19,36311.94,48834.0,85145.94
 2024-02-20,39861.28,47910.66,87771.94
 2024-02-21,39861.28,48018.96,87880.23999999999
 2024-02-22,41861.28,45818.24,87679.51999999999
 2024-02-23,41861.28,44551.96,86413.23999999999
 2024-02-26,45030.24,41690.74,86720.98
 2024-02-27,48777.32,41789.439999999995,90566.76
 2024-02-28,49161.3,39334.85999999999,88496.16
 2024-02-29,45761.3,44599.96,90361.26000000001
 2024-03-01,49575.3,40919.28,90494.58
 2024-03-04,52677.94,37579.03999999999,90256.98
 2024-03-05,51277.94,38087.04,89364.98000000001
 2024-03-06,53277.94,36919.08,90197.02
 2024-03-07,53677.94,35949.98,89627.92000000001
 2024-03-08,53677.94,35737.8,89415.74
 2024-03-11,57519.94,32557.06,90077.0
 2024-03-12,59847.920000000006,30947.68,90795.6
 2024-03-13,59847.920000000006,30429.56,90277.48000000001
 2024-03-14,60047.920000000006,30268.980000000003,90316.90000000001
 2024-03-15,60047.920000000006,31026.780000000002,91074.70000000001
 2024-03-18,62559.920000000006,29269.68,91829.6
 2024-03-19,62905.560000000005,29223.520000000004,92129.08000000002
 2024-03-20,65165.560000000005,27477.68,92643.24
 2024-03-21,65873.84,26703.800000000003,92577.64
 2024-03-22,65873.84,26076.62,91950.45999999999
 2024-03-25,61065.36,29863.06,90928.42
 2024-03-26,59265.36,31568.499999999996,90833.86
 2024-03-27,55865.36,34504.659999999996,90370.01999999999
 2024-03-28,60401.36,30693.719999999998,91095.08
 2024-03-29,60401.36,30781.6,91182.95999999999
 2024-04-01,66053.02,26153.920000000002,92206.94
 2024-04-02,66617.02,25795.5,92412.52
 2024-04-03,65179.58,26641.44,91821.02
 2024-04-08,63035.58,27350.300000000003,90385.88
 2024-04-09,61435.58,29304.18,90739.76000000001
 2024-04-10,56635.58,33710.5,90346.08
 2024-04-11,56903.58,32962.08,89865.66
 2024-04-12,56903.58,32507.18,89410.76000000001
 2024-04-15,45899.58,40168.12,86067.70000000001
 2024-04-16,36699.58,44627.16,81326.74
 2024-04-17,33499.58,52978.979999999996,86478.56
 2024-04-18,33499.58,53139.6,86639.18
 2024-04-19,33499.58,52126.22,85625.8
 2024-04-22,35961.58,49286.22,85247.8
 2024-04-23,39761.58,47887.979999999996,87649.56
 2024-04-24,41361.58,47785.3,89146.88
 2024-04-25,41161.58,47469.36,88630.94
 2024-04-26,42761.58,46693.66,89455.24
 2024-04-29,46142.72,44846.68,90989.4
 2024-04-30,51342.72,39697.380000000005,91040.1
 2024-05-06,59221.060000000005,32499.46,91720.52
 2024-05-07,59421.060000000005,32774.7,92195.76000000001
 2024-05-08,59805.060000000005,31778.78,91583.84
 2024-05-09,58005.060000000005,33495.740000000005,91500.80000000002
 2024-05-10,56205.060000000005,34632.66,90837.72
 2024-05-13,54405.060000000005,35785.520000000004,90190.58000000002
 2024-05-14,56405.060000000005,34677.5,91082.56
 2024-05-15,56791.060000000005,34668.46,91459.52
 2024-05-16,58591.060000000005,33183.68,91774.74
 2024-05-17,60869.060000000005,31647.16,92516.22
 2024-05-20,60891.060000000005,31465.64,92356.70000000001
 2024-05-21,61091.060000000005,31509.34,92600.40000000001
 2024-05-22,61405.060000000005,31812.32,93217.38
 2024-05-23,61405.060000000005,32031.88,93436.94
 2024-05-24,59605.060000000005,32880.44,92485.5
 2024-05-27,59387.50000000001,33578.0,92965.5
 2024-05-28,63387.50000000001,29672.0,93059.5
 2024-05-29,61787.50000000001,31334.0,93121.5
 2024-05-30,64141.5,28916.0,93057.5
 2024-05-31,64141.5,29130.0,93271.5
 2024-06-03,59560.399999999994,32518.96,92079.35999999999
 2024-06-04,57760.399999999994,33912.6,91673.0
 2024-06-05,56160.399999999994,34984.96,91145.35999999999
 2024-06-06,47960.399999999994,40566.12,88526.51999999999
 2024-06-07,47960.399999999994,42130.56,90090.95999999999
 2024-06-11,50552.399999999994,40529.64,91082.04
 2024-06-12,54152.399999999994,37975.64,92128.04
 2024-06-13,56152.399999999994,36528.12,92680.51999999999
 2024-06-14,56152.399999999994,36035.36,92187.76
 2024-06-17,58145.759999999995,33718.0,91863.76
 2024-06-18,60345.759999999995,32524.0,92869.76
 2024-06-19,64277.759999999995,29616.0,93893.76
 2024-06-20,59921.759999999995,33686.0,93607.76
 2024-06-21,62239.759999999995,31984.0,94223.76
 2024-06-24,56927.759999999995,35350.0,92277.76
 2024-06-25,53927.759999999995,37340.0,91267.76
 2024-06-26,52527.759999999995,40366.0,92893.76
 2024-06-27,52527.759999999995,39326.0,91853.76
 2024-06-28,52527.759999999995,40070.0,92597.76
 2024-07-01,57551.759999999995,35500.0,93051.76
 2024-07-02,57551.759999999995,36244.0,93795.76
 2024-07-03,57551.759999999995,36146.0,93697.76
 2024-07-04,55751.759999999995,36268.0,92019.76
 2024-07-05,52351.759999999995,40466.0,92817.76
 2024-07-08,53531.759999999995,38242.0,91773.76
 2024-07-09,52891.759999999995,40390.0,93281.76
 2024-07-10,51451.719999999994,41648.04,93099.76
 2024-07-11,51451.719999999994,42945.86,94397.57999999999
 2024-07-12,51451.719999999994,42662.64,94114.35999999999
 2024-07-15,52778.87999999999,40392.92,93171.79999999999
 2024-07-16,52778.87999999999,41218.92,93997.79999999999
 2024-07-17,57090.87999999999,36641.42,93732.29999999999
 2024-07-18,55290.87999999999,38284.96,93575.84
 2024-07-19,55290.87999999999,38665.44,93956.31999999999
 2024-07-22,60996.87999999999,34089.5,95086.37999999999
 2024-07-23,60996.87999999999,33558.479999999996,94555.35999999999
 2024-07-24,62796.87999999999,31740.760000000002,94537.63999999998
 2024-07-25,61256.87999999999,33841.24,95098.12
 2024-07-26,61456.87999999999,33980.7,95437.57999999999
 2024-07-29,59777.73999999999,35923.119999999995,95700.85999999999
 2024-07-30,61777.73999999999,34529.72,96307.45999999999
 2024-07-31,64101.73999999999,33334.0,97435.73999999999
 2024-08-01,65125.73999999999,32018.0,97143.73999999999
 2024-08-02,61925.73999999999,34468.0,96393.73999999999
 2024-08-05,62605.73999999999,32892.0,95497.73999999999
 2024-08-06,62605.73999999999,34194.0,96799.73999999999
 2024-08-07,64405.73999999999,32874.0,97279.73999999999
 2024-08-08,64605.73999999999,31938.0,96543.73999999999
 2024-08-09,63005.73999999999,33154.0,96159.73999999999
 2024-08-12,61315.73999999999,34852.0,96167.73999999999
 2024-08-13,59515.73999999999,36410.0,95925.73999999999
 2024-08-14,59515.73999999999,35968.0,95483.73999999999
 2024-08-15,58515.73999999999,37140.0,95655.73999999999
 2024-08-16,56715.73999999999,39058.0,95773.73999999999
 2024-08-19,56764.19999999999,38072.64,94836.84
 2024-08-20,56764.19999999999,38040.52,94804.71999999999
 2024-08-21,55164.19999999999,38808.92,93973.12
 2024-08-22,53564.19999999999,39192.04,92756.23999999999
 2024-08-23,55364.19999999999,37672.44,93036.63999999998
 2024-08-26,56940.25999999999,36384.899999999994,93325.15999999997
 2024-08-27,57540.25999999999,35327.28,92867.53999999998
 2024-08-28,52940.25999999999,39837.560000000005,92777.81999999999
 2024-08-29,51892.25999999999,41458.62,93350.87999999999
 2024-08-30,52183.51999999999,42771.06,94954.57999999999
 2024-09-02,56710.33999999999,37626.06,94336.4
 2024-09-03,58710.33999999999,36281.9,94992.23999999999
 2024-09-04,58710.33999999999,35629.16,94339.5
 2024-09-05,58710.33999999999,36421.92,95132.25999999998
 2024-09-06,58710.33999999999,35878.58,94588.91999999998
 2024-09-09,58906.33999999999,35322.54000000001,94228.88
 2024-09-10,57306.33999999999,37236.6,94542.93999999999
 2024-09-11,57306.33999999999,36978.36,94284.69999999998
 2024-09-12,62642.33999999999,31378.359999999997,94020.69999999998
 2024-09-13,62642.33999999999,30741.5,93383.84
 2024-09-18,58008.74,34825.06,92833.79999999999
 2024-09-19,58359.32,35364.84,93724.16
 2024-09-20,58359.32,34882.94,93242.26000000001
 2024-09-23,59825.32,33529.54,93354.86
 2024-09-24,59825.32,35009.0,94834.32
 2024-09-25,62217.32,33031.08,95248.4
 2024-09-26,62723.32000000001,33992.1,96715.42000000001
 2024-09-27,69235.32,29121.08,98356.40000000001
 2024-09-30,80349.32,21369.96,101719.28
 2024-10-08,85077.28,19476.0,104553.28
 2024-10-09,77483.28,25692.0,103175.28
 2024-10-10,65483.28,35852.0,101335.28
 2024-10-11,56083.28,43112.0,99195.28
 2024-10-14,54845.28,47550.0,102395.28
 2024-10-15,62245.28,40032.0,102277.28
 2024-10-16,64429.28,38264.0,102693.28
 2024-10-17,71101.28,33832.0,104933.28
 2024-10-18,71101.28,35670.0,106771.28
 2024-10-21,81899.28,27310.0,109209.28
 2024-10-22,78499.28,29890.0,108389.28
 2024-10-23,79145.28,31210.0,110355.28
 2024-10-24,83457.28,27304.0,110761.28
 2024-10-25,87055.28,25324.0,112379.28
 2024-10-28,82369.28,32322.0,114691.28
 2024-10-29,94723.28,21330.0,116053.28
 2024-10-30,93926.66,22730.62,116657.28
 2024-10-31,92874.66,23630.56,116505.22
 2024-11-01,90986.66,26929.52,117916.18000000001
 2024-11-04,88286.66,31429.6,119716.26000000001
 2024-11-05,99332.66,22827.28,122159.94
 2024-11-06,100274.82,22557.2,122832.02
 2024-11-07,97836.82,26869.02,124705.84000000001
 2024-11-08,100636.82,23667.66,124304.48000000001
 2024-11-11,97436.82,27560.78,124997.6
 2024-11-12,97436.82,27386.06,124822.88
 2024-11-13,95636.82,29058.3,124695.12000000001
 2024-11-14,97636.82,26820.22,124457.04000000001
 2024-11-15,96557.20000000001,28499.719999999998,125056.92000000001
 2024-11-18,99403.46,26804.239999999998,126207.70000000001
 2024-11-19,87403.46,38321.020000000004,125724.48000000001
 2024-11-20,91665.46,36476.68,128142.14000000001
 2024-11-21,97525.46,31670.04,129195.5
 2024-11-22,102164.48000000001,26932.02,129096.50000000001
 2024-11-25,100326.48000000001,29415.52,129742.00000000001
 2024-11-26,98726.48000000001,30748.68,129475.16
 2024-11-27,97126.48000000001,32619.879999999997,129746.36000000002
 2024-11-28,101950.1,29081.36,131031.46
 2024-11-29,101950.1,29600.280000000002,131550.38
 2024-12-02,108792.78,23990.58,132783.36
 2024-12-03,105192.78,27452.98,132645.76
 2024-12-04,103392.78,27856.059999999998,131248.84
 2024-12-05,103392.78,28402.48,131795.26
 2024-12-06,103392.78,28411.32,131804.1
 2024-12-09,97964.12,32072.58,130036.7
 2024-12-10,96164.12,34057.399999999994,130221.51999999999
 2024-12-11,94564.12,35279.64,129843.76
 2024-12-12,94844.12,35000.0,129844.12
 2024-12-13,96844.12,33198.0,130042.12
 2024-12-16,92244.12,37306.0,129550.12
 2024-12-17,90444.12,36824.0,127268.12
 2024-12-18,85644.12,41096.0,126740.12
 2024-12-19,82444.12,45412.0,127856.12
 2024-12-20,82444.12,44764.0,127208.12
 2024-12-23,86146.12,38902.0,125048.12
 2024-12-24,88234.12,37016.0,125250.12
 2024-12-25,83634.12,39390.0,123024.12
 2024-12-26,85918.12,38396.0,124314.12
 2024-12-27,89318.12,35982.0,125300.12
 2024-12-30,84450.12,39132.0,123582.12
 2024-12-31,86250.12,38410.0,124660.12
 2025-01-02,85608.12,39822.0,125430.12
 2025-01-03,87408.12,37578.0,124986.12
 2025-01-06,87180.12,36526.0,123706.12
 2025-01-07,85580.12,39196.0,124776.12
 2025-01-08,85580.12,39276.0,124856.12
 2025-01-09,85580.12,40108.0,125688.12
 2025-01-10,85580.12,38096.0,123676.12
 2025-01-13,86472.12,36632.0,123104.12
 2025-01-14,90534.12,35542.0,126076.12
 2025-01-15,92964.12,33396.0,126360.12
 2025-01-16,92964.12,33474.0,126438.12
 2025-01-17,92964.12,32718.0,125682.12
 2025-01-20,92444.12,33334.0,125778.12
 2025-01-21,90644.12,34562.0,125206.12
 2025-01-22,88844.12,35200.0,124044.12
 2025-01-23,92964.12,31104.0,124068.12
 2025-01-24,92964.12,30942.0,123906.12
 2025-01-27,93264.12,29920.0,123184.12
 2025-02-05,95550.12,28902.0,124452.12
 2025-02-06,95386.12,30334.0,125720.12
 2025-02-07,99682.12,27098.0,126780.12
 2025-02-10,101682.12,26282.0,127964.12
 2025-02-11,102255.36,25700.760000000002,127956.12
 2025-02-12,106503.36,22042.74,128546.1
 2025-02-13,106503.36,21696.8,128200.16
 2025-02-14,106503.36,21418.78,127922.14
 2025-02-17,106503.36,22392.739999999998,128896.1
 2025-02-18,106863.36,21594.6,128457.95999999999
 2025-02-19,107549.72,21399.64,128949.36
 2025-02-20,105749.72,23389.28,129139.0
 2025-02-21,105749.72,23782.4,129532.12
 2025-02-24,107749.72,22060.94,129810.66
 2025-02-25,106050.72,23940.18,129990.9
 2025-02-26,106050.72,24078.079999999998,130128.8
 2025-02-27,104528.72,25574.48,130103.2
 2025-02-28,102728.72,26542.96,129271.68
 2025-03-03,99240.72,30762.88,130003.6
 2025-03-04,104273.36,27485.84,131759.2
 2025-03-05,106519.36,25571.420000000002,132090.78
 2025-03-06,103171.36,29194.32,132365.68
 2025-03-07,105425.36,27040.4,132465.76
 2025-03-10,105425.36,27684.64,133110.0
 2025-03-11,105757.36,27476.72,133234.08000000002
 2025-03-12,107757.36,26166.48,133923.84
 2025-03-13,110301.82,23848.38,134150.2
 2025-03-14,115013.82,19782.36,134796.18
 2025-03-17,115373.82,19713.06,135086.88
 2025-03-18,116287.82,19132.559999999998,135420.38
 2025-03-19,112687.82,22446.72,135134.54
 2025-03-20,111223.82,24020.72,135244.54
 2025-03-21,109423.82,25211.120000000003,134634.94
 2025-03-24,102104.94,32562.0,134666.94
 2025-03-25,104460.1,30420.84,134880.94
 2025-03-26,106460.1,29221.84,135681.94
 2025-03-27,106660.1,28888.8,135548.9
 2025-03-28,103822.1,31484.3,135306.4
 2025-03-31,100434.62000000001,34067.08,134501.7
 2025-04-01,101400.62000000001,32980.64,134381.26
 2025-04-02,101600.62000000001,32563.239999999998,134163.86000000002
 2025-04-03,99800.62000000001,34518.32,134318.94
 2025-04-07,87000.62000000001,41837.2,128837.82
 2025-04-08,80800.62000000001,49853.04,130653.66
 2025-04-09,88600.62000000001,46010.08,134610.7
 2025-04-10,98000.62000000001,38655.6,136656.22
 2025-04-11,99600.62000000001,37063.76,136664.38
 2025-04-14,106254.62000000001,31782.72,138037.34000000003
 2025-04-15,104654.62000000001,33317.28,137971.90000000002
 2025-04-16,106747.54000000001,31677.72,138425.26
 2025-04-17,108747.54000000001,30180.46,138928.0
 2025-04-18,109341.54000000001,30561.36,139902.90000000002
 2025-04-21,107815.54000000001,32425.2,140240.74000000002
 2025-04-22,107815.54000000001,31336.8,139152.34
 2025-04-23,109415.54000000001,30456.64,139872.18
 2025-04-24,109657.54000000001,29465.66,139123.2
 2025-04-25,106835.54000000001,32281.18,139116.72
 2025-04-28,106345.54000000001,32082.7,138428.24000000002
 2025-04-29,102745.54000000001,35483.0,138228.54
 2025-04-30,102745.54000000001,35759.3,138504.84000000003
 2025-05-06,104510.84000000001,35040.0,139550.84000000003
 2025-05-07,107283.84000000001,32649.0,139932.84000000003
 2025-05-08,104437.84000000001,35792.24,140230.08000000002
 2025-05-09,102637.84000000001,37585.88,140223.72
 2025-05-12,103367.84000000001,37894.72,141262.56
 2025-05-13,103367.84000000001,37302.54,140670.38
 2025-05-14,103763.84000000001,36662.0,140425.84000000003
 2025-05-15,101963.84000000001,38680.0,140643.84000000003
 2025-05-16,101963.84000000001,38900.0,140863.84000000003
 2025-05-19,101963.84000000001,39840.0,141803.84000000003
 2025-05-20,105763.84000000001,36696.0,142459.84000000003
 2025-05-21,107363.84000000001,35314.0,142677.84000000003
 2025-05-22,109423.84000000001,33250.0,142673.84000000003
 2025-05-23,109423.84000000001,32758.0,142181.84000000003
 2025-05-26,110715.84000000001,32416.0,143131.84000000003
 2025-05-27,114765.84000000001,28756.0,143521.84000000003
 2025-05-28,118565.84000000001,24386.0,142951.84000000003
 2025-05-29,118565.84000000001,24558.0,143123.84000000003
 2025-05-30,115165.84000000001,27606.0,142771.84000000003
 2025-06-03,113691.88,29479.86,143171.74
 2025-06-04,117691.88,26375.22,144067.1
 2025-06-05,118367.88,25461.82,143829.7
 2025-06-06,118567.88,25523.98,144091.86000000002
 2025-06-09,118468.06,26145.58,144613.64
 2025-06-10,119242.06,26013.68,145255.74
 2025-06-11,115642.06,29543.460000000003,145185.52
 2025-06-12,119442.06,26731.8,146173.86
 2025-06-13,121827.45999999999,23271.14,145098.59999999998
 2025-06-16,118827.45999999999,26558.000000000004,145385.46
 2025-06-17,119851.45999999999,25974.019999999997,145825.47999999998
 2025-06-18,118051.45999999999,27032.4,145083.86
 2025-06-19,112651.45999999999,31251.42,143902.88
 2025-06-20,109651.45999999999,33717.020000000004,143368.47999999998
 2025-06-23,108051.45999999999,36004.66,144056.12
 2025-06-24,108051.45999999999,36450.659999999996,144502.12
 2025-06-25,108567.45999999999,36190.72,144758.18
 2025-06-26,108767.45999999999,36286.82,145054.28
 2025-06-27,108767.45999999999,36246.7,145014.15999999997
 2025-06-30,108321.45999999999,37252.479999999996,145573.94
 2025-07-01,115200.18,31736.68,146936.86
 2025-07-02,117302.28,30046.74,147349.02
 2025-07-03,117302.28,29892.84,147195.12
 2025-07-04,117302.28,29383.94,146686.22
 2025-07-07,117586.28,29861.88,147448.16
 2025-07-08,117586.28,29913.9,147500.18
 2025-07-09,117912.28,29733.94,147646.22
 2025-07-10,118112.28,29558.0,147670.28
 2025-07-11,116312.28,31418.0,147730.28
 2025-07-14,114316.28,33108.0,147424.28
 2025-07-15,112716.28,34186.0,146902.28
 2025-07-16,114716.28,32426.0,147142.28
 2025-07-17,114960.2,32414.08,147374.28
 2025-07-18,113160.2,33946.38,147106.58
 2025-07-21,116704.2,30680.16,147384.36
 2025-07-22,116980.59999999999,30577.559999999998,147558.15999999997
 2025-07-23,116980.59999999999,30195.579999999998,147176.18
 2025-07-24,115180.59999999999,32133.800000000003,147314.4
 2025-07-25,115180.59999999999,32153.94,147334.53999999998
 2025-07-28,115180.59999999999,32243.62,147424.22
 2025-07-29,115180.59999999999,31921.92,147102.52
 2025-07-30,115180.59999999999,31987.88,147168.47999999998
 2025-07-31,113580.59999999999,33084.16,146664.76
 2025-08-01,113316.59999999999,33357.8,146674.4
 2025-08-04,116562.59999999999,30359.96,146922.56
 2025-08-05,114762.59999999999,32258.28,147020.88
 2025-08-06,114762.59999999999,32168.1,146930.69999999998
 2025-08-07,114762.59999999999,32379.9,147142.5
 2025-08-08,119086.59999999999,28390.0,147476.59999999998
 2025-08-11,119218.59999999999,28752.0,147970.59999999998
 2025-08-12,117418.59999999999,30502.0,147920.59999999998
 2025-08-13,121062.59999999999,26722.0,147784.59999999998
 2025-08-14,117662.59999999999,29492.0,147154.59999999998
 2025-08-15,119662.59999999999,28798.0,148460.59999999998
 2025-08-18,123252.59999999999,26458.0,149710.59999999998
 2025-08-19,122090.59999999999,27948.0,150038.59999999998
 2025-08-20,120290.59999999999,29984.0,150274.59999999998
 2025-08-21,126290.59999999999,24298.0,150588.59999999998
 2025-08-22,126652.59999999999,23932.0,150584.59999999998
 2025-08-25,126652.59999999999,24162.0,150814.59999999998
 2025-08-26,126852.59999999999,24104.0,150956.59999999998
 2025-08-27,123646.59999999999,26804.0,150450.59999999998
 2025-08-28,118640.59999999999,32296.0,150936.59999999998
 2025-08-29,116840.59999999999,33900.0,150740.59999999998
 2025-09-01,122140.01999999999,29354.24,151494.25999999998
 2025-09-02,120726.01999999999,30304.0,151030.02
 2025-09-03,117326.01999999999,32868.0,150194.02
 2025-09-04,114126.01999999999,35582.0,149708.02
 2025-09-05,114126.01999999999,36400.0,150526.02
 2025-09-08,114126.01999999999,36670.0,150796.02
 2025-09-09,114126.01999999999,35764.0,149890.02
 2025-09-10,114126.01999999999,36264.0,150390.02
 2025-09-11,116326.01999999999,34852.0,151178.02
 2025-09-12,116580.01999999999,34638.0,151218.02
 2025-09-15,116840.01999999999,34224.0,151064.02
 2025-09-16,116840.01999999999,34266.0,151106.02
 2025-09-17,116840.01999999999,34016.0,150856.02
 2025-09-18,115522.01999999999,34918.0,150440.02
 2025-09-19,115522.01999999999,34548.0,150070.02
 2025-09-22,116524.01999999999,33574.0,150098.02
 2025-09-23,111924.01999999999,37476.0,149400.02
 2025-09-24,112264.01999999999,37360.0,149624.02
 2025-09-25,112264.01999999999,36962.0,149226.02
 2025-09-26,112264.01999999999,36744.0,149008.02
 2025-09-29,116726.01999999999,32582.0,149308.02
 2025-09-30,116726.01999999999,32560.0,149286.02
 2025-10-09,118370.01999999999,30188.0,148558.02
 2025-10-10,116570.01999999999,31992.0,148562.02
 2025-10-13,112600.01999999999,36462.0,149062.02
 2025-10-14,112600.01999999999,36110.0,148710.02
 2025-10-15,112600.01999999999,36570.0,149170.02
 2025-10-16,114600.01999999999,34026.0,148626.02
 2025-10-17,113000.01999999999,34894.0,147894.02
 2025-10-20,112998.01999999999,35922.0,148920.02
 2025-10-21,114998.01999999999,34854.0,149852.02
 2025-10-22,114998.01999999999,34606.0,149604.02
 2025-10-23,113198.01999999999,36532.0,149730.02
 2025-10-24,113198.01999999999,36844.0,150042.02
 2025-10-27,116430.01999999999,33332.0,149762.02
 2025-10-28,116430.01999999999,33280.0,149710.02
 2025-10-29,116430.01999999999,33934.0,150364.02
 2025-10-30,116430.01999999999,33806.0,150236.02
 2025-10-31,116430.01999999999,34148.0,150578.02
 2025-11-03,121148.01999999999,29008.0,150156.02
 2025-11-04,120108.01999999999,30110.0,150218.02
 2025-11-05,120108.01999999999,30552.0,150660.02
 2025-11-06,118308.01999999999,32286.0,150594.02
 2025-11-07,118702.01999999999,31342.0,150044.02
 2025-11-10,118702.01999999999,31688.0,150390.02
 2025-11-11,123102.01999999999,27790.0,150892.02
 2025-11-12,118426.01999999999,32408.0,150834.02
 2025-11-13,119016.01999999999,32752.0,151768.02
 2025-11-14,119016.01999999999,33214.0,152230.02
 2025-11-17,123270.01999999999,29114.0,152384.02
 2025-11-18,119670.01999999999,31876.0,151546.02
 2025-11-19,118070.01999999999,33268.0,151338.02
 2025-11-20,116470.01999999999,34108.0,150578.02
 2025-11-21,109870.01999999999,38562.0,148432.02
 2025-11-24,109938.01999999999,38512.0,148450.02
 2025-11-25,111938.01999999999,37762.0,149700.02
 2025-11-26,111938.01999999999,37170.0,149108.02
 2025-11-27,111938.01999999999,37172.0,149110.02
 2025-11-28,111938.01999999999,37834.0,149772.02
 2025-12-01,115825.68,33942.38,149768.06
 2025-12-02,115825.68,33994.38,149820.06
 2025-12-03,117825.68,31736.9,149562.58
 2025-12-04,112425.68,36213.44,148639.12
 2025-12-05,109425.68,40135.04,149560.72
 2025-12-08,112603.68,37374.88,149978.56
 2025-12-09,116729.68,33392.68,150122.36
 2025-12-10,117313.68,33230.36,150544.03999999998
 2025-12-11,115769.68,34872.0,150641.68
 2025-12-12,115769.68,35290.0,151059.68
 2025-12-15,118523.68,33424.0,151947.68
 2025-12-16,118723.68,32706.0,151429.68
 2025-12-17,111723.68,39646.0,151369.68
 2025-12-18,111723.68,39308.0,151031.68
 2025-12-19,110123.68,41948.0,152071.68
 2025-12-22,116025.68,36470.0,152495.68
 2025-12-23,116225.68,36004.0,152229.68
 2025-12-24,115195.68,37596.0,152791.68
 2025-12-25,115571.68,37138.0,152709.68
 2025-12-26,115371.68,37724.0,153095.68
 2025-12-29,118263.68,34576.0,152839.68
 2025-12-30,116959.68,35302.0,152261.68
 2025-12-31,115159.68,36376.0,151535.68
 2026-01-05,114658.76,36630.8,151289.56
 2026-01-06,113458.76,38219.44,151678.2
 2026-01-07,113854.76,37808.0,151662.76
 2026-01-08,112054.76,40598.0,152652.76
 2026-01-09,113854.76,39384.0,153238.76
 2026-01-12,114728.76,38852.0,153580.76
 2026-01-13,120616.76,32848.0,153464.76
 2026-01-14,121016.76,32938.0,153954.76
 2026-01-15,121216.76,32146.0,153362.76
 2026-01-16,117816.76,34904.0,152720.76
 2026-01-19,118854.76,34260.0,153114.76
 2026-01-20,117516.76,35734.0,153250.76
 2026-01-21,120376.76,34212.0,154588.76
 2026-01-22,120376.76,34816.0,155192.76
 2026-01-23,120376.76,35548.0,155924.76
 2026-01-26,119168.76,36470.0,155638.76
 2026-01-27,119368.76,36470.0,155838.76
 2026-01-28,119368.76,35508.0,154876.76
 2026-01-29,117768.76,37012.0,154780.76
 2026-01-30,118104.76,36686.0,154790.76
 2026-02-02,123830.76,30870.0,154700.76
 2026-02-03,123830.76,31876.000000000004,155706.76
 2026-02-04,123830.76,31492.0,155322.76
 2026-02-05,123830.76,31724.0,155554.76
 2026-02-06,123830.76,32184.0,156014.76
 2026-02-09,124128.76,32940.0,157068.76
 2026-02-10,130252.76000000001,27476.0,157728.76
 2026-02-11,128452.76000000001,29178.0,157630.76
 2026-02-12,130452.76000000001,27600.0,158052.76
 2026-02-13,130652.76000000001,27454.0,158106.76
 2026-02-24,129478.76000000001,29166.0,158644.76
 2026-02-25,129678.76000000001,29674.0,159352.76
 2026-02-26,129878.76000000001,29368.0,159246.76
 2026-02-27,128078.76000000001,31448.0,159526.76
 2026-03-02,126568.76000000001,31928.0,158496.76
 2026-03-03,123168.76000000001,33314.0,156482.76
 2026-03-04,121530.76000000001,35378.0,156908.76
 2026-03-05,123530.76000000001,33996.0,157526.76
 2026-03-06,121930.76000000001,35786.0,157716.76
 2026-03-09,120130.76000000001,37128.0,157258.76
 2026-03-10,122130.76000000001,36208.0,158338.76
 2026-03-11,122498.76000000001,35428.0,157926.76
 2026-03-12,122498.76000000001,34782.0,157280.76
 2026-03-13,122818.76000000001,33916.0,156734.76
 2026-03-16,119866.98000000001,36893.56,156760.54
 2026-03-17,120266.98000000001,35880.16,156147.14
 2026-03-18,120266.98000000001,36240.88,156507.86000000002
 2026-03-19,115666.98000000001,39534.56,155201.54
 2026-03-20,114066.98000000001,39770.48,153837.46000000002
 2026-03-23,112544.98000000001,40740.8,153285.78000000003
 2026-03-24,111144.98000000001,44227.28,155372.26
 2026-03-25,113144.98000000001,43183.8,156328.78000000003
 2026-03-26,114944.98000000001,40874.96,155819.94
 2026-03-27,114944.98000000001,41492.52,156437.5
 2026-03-30,118332.98000000001,38077.28,156410.26
 2026-03-31,118332.98000000001,37634.52,155967.5
 2026-04-01,121430.98000000001,35512.04,156943.02000000002
 2026-04-02,121430.98000000001,34988.72,156419.7
 2026-04-03,118420.98000000001,37343.56,155764.54
 2026-04-07,117714.98000000001,38003.479999999996,155718.46000000002
 2026-04-08,117980.98000000001,39241.32,157222.30000000002
 2026-04-09,117980.98000000001,38689.08,156670.06
 2026-04-10,118180.98000000001,38458.6,156639.58000000002
 2026-04-13,116651.58000000002,39420.0,156071.58000000002
 2026-04-14,115051.58000000002,40588.0,155639.58000000002
 2026-04-15,114451.58000000002,40568.0,155019.58000000002
 2026-04-16,113051.58000000002,41612.0,154663.58000000002
 2026-04-17,111851.58000000002,42540.0,154391.58000000002
 2026-04-20,111939.58000000002,43314.0,155253.58000000002
 2026-04-21,111939.58000000002,43070.0,155009.58000000002
 2026-04-22,111939.58000000002,42160.0,154099.58000000002
 2026-04-23,110939.58000000002,41660.0,152599.58000000002
 2026-04-24,109899.58000000002,42320.0,152219.58000000002
 2026-04-27,113003.58000000002,39480.0,152483.58000000002
 2026-04-28,112317.58000000002,38134.0,150451.58000000002
 2026-04-29,113317.58000000002,37898.0,151215.58000000002
 2026-04-30,111717.58000000002,39436.0,151153.58000000002
@@ -0,0 +1,37 @@
 date,cash,stock_value,total_value,positions,month
 2023-05-31,17736.34,41098.4,58834.74,10,2023-05
 2023-06-30,20515.46,40323.66,60839.12,10,2023-06
 2023-07-31,30397.84,29960.78,60358.62,9,2023-07
 2023-08-31,26136.7,34300.62,60437.32,10,2023-08
 2023-09-30,39873.34,27709.4,67582.74,10,2023-09
 2023-10-31,36088.52,31940.92,68029.44,10,2023-10
 2023-11-30,49700.74,27047.76,76748.5,10,2023-11
 2023-12-31,56387.94,33956.0,90343.94,10,2023-12
 2024-01-31,43543.94,44782.0,88325.94,10,2024-01
 2024-02-29,41861.28,44551.96,86413.24,10,2024-02
 2024-03-31,60401.36,30781.6,91182.96,9,2024-03
 2024-04-30,51342.72,39697.38,91040.1,10,2024-04
 2024-05-31,65813.5,29130.0,94943.5,9,2024-05
 2024-06-30,52527.76,40070.0,92597.76,10,2024-06
 2024-07-31,61456.88,33980.7,95437.58,10,2024-07
 2024-08-31,52183.52,42771.06,94954.58,10,2024-08
 2024-09-30,80349.32,21369.96,101719.28,10,2024-09
 2024-10-31,87055.28,25324.0,112379.28,10,2024-10
 2024-11-30,101950.1,29600.28,131550.38,10,2024-11
 2024-12-31,89318.12,35982.0,125300.12,8,2024-12
 2025-01-31,95062.12,29920.0,124982.12,9,2025-01
 2025-02-28,102728.72,26542.96,129271.68,9,2025-02
 2025-03-31,103822.1,31484.3,135306.4,10,2025-03
 2025-04-30,102745.54,35759.3,138504.84,10,2025-04
 2025-05-31,115165.84,27606.0,142771.84,8,2025-05
 2025-06-30,108767.46,36246.7,145014.16,9,2025-06
 2025-07-31,115180.6,32153.94,147334.54,10,2025-07
 2025-08-31,116840.6,33900.0,150740.6,10,2025-08
 2025-09-30,116726.02,32560.0,149286.02,10,2025-09
 2025-10-31,116430.02,34148.0,150578.02,10,2025-10
 2025-11-30,111938.02,37834.0,149772.02,10,2025-11
 2025-12-31,115159.68,36376.0,151535.68,10,2025-12
 2026-01-31,118104.76,36686.0,154790.76,10,2026-01
 2026-02-28,128078.76,31448.0,159526.76,10,2026-02
 2026-03-31,114944.98,41492.52,156437.5,10,2026-03
 2026-04-30,111717.58,39436.0,151153.58,10,2026-04
@@ -0,0 +1,21 @@
 {
  "version": "v6.7r2",
  "rank_model": "lambdarank (56-dim v6.7)",
  "backtest_start": "2023-05-01",
  "backtest_end": "2026-04-30",
  "n_weeks": 154,
  "n_stocks_in_pool": 5378,
  "top_n": 10,
  "shares_per_grid": 200,
  "initial_cash": 60000.0,
  "final_total_value": 151153.58,
  "total_return_pct": 151.92,
  "annual_return_pct": 36.1,
  "annual_sharpe": 1.7404,
  "max_drawdown_pct": -12.1,
  "weekly_win_rate_pct": 59.48,
  "n_trades_buy": 1144,
  "n_trades_sell": 866,
  "realized_pnl_total": 95495.58,
  "n_lifecycles": 431
 }
@@ -0,0 +1,155 @@
 date,cash,stock_value,total_value,positions,year
 2023-05-05,29353.86,32311.96,61665.82,10,2023
 2023-05-12,24353.86,37200.04,61553.9,10,2023
 2023-05-19,19999.86,40128.84,60128.7,10,2023
 2023-05-26,17736.34,41098.4,58834.74,10,2023
 2023-06-02,22233.92,40253.44,62487.36,10,2023
 2023-06-09,23083.9,39432.4,62516.3,10,2023
 2023-06-16,27171.9,36662.04,63833.94,10,2023
 2023-06-21,27177.36,35306.82,62484.18,10,2023
 2023-06-30,20515.46,40323.66,60839.12,10,2023
 2023-07-07,27155.76,33151.0,60306.76,10,2023
 2023-07-14,29091.76,31705.8,60797.56,10,2023
 2023-07-21,29114.82,32509.1,61623.92,9,2023
 2023-07-28,30397.84,29960.78,60358.62,9,2023
 2023-08-04,31326.78,29340.22,60667.0,10,2023
 2023-08-11,28602.46,31713.8,60316.26,10,2023
 2023-08-18,33136.7,28257.34,61394.04,10,2023
 2023-08-25,26136.7,34300.62,60437.32,10,2023
 2023-09-01,33946.78,29311.82,63258.6,10,2023
 2023-09-08,29682.92,33243.98,62926.9,10,2023
 2023-09-15,33929.1,29926.24,63855.34,10,2023
 2023-09-22,35117.34,29932.44,65049.78,10,2023
 2023-09-28,39873.34,27709.4,67582.74,10,2023
 2023-10-13,44083.08,23683.52,67766.6,10,2023
 2023-10-20,33289.08,32655.7,65944.78,10,2023
 2023-10-27,36088.52,31940.92,68029.44,10,2023
 2023-11-03,40429.7,29377.8,69807.5,10,2023
 2023-11-10,46479.74,25900.26,72380.0,10,2023
 2023-11-17,49341.34,24523.56,73864.9,10,2023
 2023-11-24,49700.74,27047.76,76748.5,10,2023
 2023-12-01,33892.34,39456.38,73348.72,10,2023
 2023-12-08,50831.34,30194.3,81025.64,10,2023
 2023-12-15,52387.94,31884.0,84271.94,10,2023
 2023-12-22,52909.94,32946.0,85855.94,10,2023
 2023-12-29,56387.94,33956.0,90343.94,10,2023
 2024-01-05,56405.94,34602.0,91007.94,10,2024
 2024-01-12,42405.94,42438.0,84843.94,10,2024
 2024-01-19,44405.94,46458.0,90863.94,10,2024
 2024-01-26,43543.94,44782.0,88325.94,10,2024
 2024-02-02,27305.94,52690.0,79995.94,10,2024
 2024-02-08,31419.94,51926.0,83345.94,10,2024
 2024-02-23,41861.28,44551.96,86413.24,10,2024
 2024-03-01,49575.3,40919.28,90494.58,10,2024
 2024-03-08,53677.94,35737.8,89415.74,10,2024
 2024-03-15,60047.92,31026.78,91074.7,10,2024
 2024-03-22,65873.84,26076.62,91950.46,10,2024
 2024-03-29,60401.36,30781.6,91182.96,9,2024
 2024-04-03,66965.58,26641.44,93607.02,9,2024
 2024-04-12,56903.58,32507.18,89410.76,10,2024
 2024-04-19,33499.58,52126.22,85625.8,10,2024
 2024-04-26,42761.58,46693.66,89455.24,10,2024
 2024-04-30,51342.72,39697.38,91040.1,10,2024
 2024-05-10,56205.06,34632.66,90837.72,10,2024
 2024-05-17,60869.06,31647.16,92516.22,10,2024
 2024-05-24,59605.06,32880.44,92485.5,10,2024
 2024-05-31,65813.5,29130.0,94943.5,9,2024
 2024-06-07,47960.4,42130.56,90090.96,10,2024
 2024-06-14,56152.4,36035.36,92187.76,10,2024
 2024-06-21,62239.76,31984.0,94223.76,10,2024
 2024-06-28,52527.76,40070.0,92597.76,10,2024
 2024-07-05,52351.76,40466.0,92817.76,10,2024
 2024-07-12,51451.72,42662.64,94114.36,10,2024
 2024-07-19,55290.88,38665.44,93956.32,10,2024
 2024-07-26,61456.88,33980.7,95437.58,10,2024
 2024-08-02,61925.74,34468.0,96393.74,10,2024
 2024-08-09,63005.74,33154.0,96159.74,10,2024
 2024-08-16,56715.74,39058.0,95773.74,10,2024
 2024-08-23,55364.2,37672.44,93036.64,10,2024
 2024-08-30,52183.52,42771.06,94954.58,10,2024
 2024-09-06,60506.34,35878.58,96384.92,9,2024
 2024-09-13,62642.34,30741.5,93383.84,8,2024
 2024-09-20,58359.32,34882.94,93242.26,10,2024
 2024-09-27,69235.32,29121.08,98356.4,10,2024
 2024-09-30,80349.32,21369.96,101719.28,10,2024
 2024-10-11,56083.28,43112.0,99195.28,10,2024
 2024-10-18,71101.28,35670.0,106771.28,10,2024
 2024-10-25,87055.28,25324.0,112379.28,10,2024
 2024-11-01,90986.66,26929.52,117916.18,10,2024
 2024-11-08,100636.82,23667.66,124304.48,10,2024
 2024-11-15,96557.2,28499.72,125056.92,10,2024
 2024-11-22,102164.48,26932.02,129096.5,10,2024
 2024-11-29,101950.1,29600.28,131550.38,10,2024
 2024-12-06,103392.78,28411.32,131804.1,10,2024
 2024-12-13,96844.12,33198.0,130042.12,10,2024
 2024-12-20,82444.12,44764.0,127208.12,10,2024
 2024-12-27,89318.12,35982.0,125300.12,8,2024
 2025-01-03,87408.12,37578.0,124986.12,10,2025
 2025-01-10,85580.12,38096.0,123676.12,10,2025
 2025-01-17,92964.12,32718.0,125682.12,9,2025
 2025-01-24,97680.12,30942.0,128622.12,8,2025
 2025-01-27,95062.12,29920.0,124982.12,9,2025
 2025-02-07,99682.12,27098.0,126780.12,9,2025
 2025-02-14,106503.36,21418.78,127922.14,9,2025
 2025-02-21,105749.72,23782.4,129532.12,9,2025
 2025-02-28,102728.72,26542.96,129271.68,9,2025
 2025-03-07,105425.36,27040.4,132465.76,10,2025
 2025-03-14,115013.82,19782.36,134796.18,9,2025
 2025-03-21,109423.82,25211.12,134634.94,9,2025
 2025-03-28,103822.1,31484.3,135306.4,10,2025
 2025-04-03,99800.62,34518.32,134318.94,10,2025
 2025-04-11,99600.62,37063.76,136664.38,10,2025
 2025-04-18,109341.54,30561.36,139902.9,10,2025
 2025-04-25,106835.54,32281.18,139116.72,10,2025
 2025-04-30,102745.54,35759.3,138504.84,10,2025
 2025-05-09,102637.84,37585.88,140223.72,10,2025
 2025-05-16,101963.84,38900.0,140863.84,10,2025
 2025-05-23,109423.84,32758.0,142181.84,9,2025
 2025-05-30,115165.84,27606.0,142771.84,8,2025
 2025-06-06,118567.88,25523.98,144091.86,10,2025
 2025-06-13,121827.46,23271.14,145098.6,9,2025
 2025-06-20,109651.46,33717.02,143368.48,9,2025
 2025-06-27,108767.46,36246.7,145014.16,9,2025
 2025-07-04,117302.28,29383.94,146686.22,10,2025
 2025-07-11,116312.28,31418.0,147730.28,10,2025
 2025-07-18,113160.2,33946.38,147106.58,10,2025
 2025-07-25,115180.6,32153.94,147334.54,10,2025
 2025-08-01,113316.6,33357.8,146674.4,10,2025
 2025-08-08,119086.6,28390.0,147476.6,10,2025
 2025-08-15,119662.6,28798.0,148460.6,9,2025
 2025-08-22,126652.6,23932.0,150584.6,10,2025
 2025-08-29,116840.6,33900.0,150740.6,10,2025
 2025-09-05,114126.02,36400.0,150526.02,10,2025
 2025-09-12,116580.02,34638.0,151218.02,10,2025
 2025-09-19,115522.02,34548.0,150070.02,10,2025
 2025-09-26,115364.02,36744.0,152108.02,9,2025
 2025-09-30,116726.02,32560.0,149286.02,10,2025
 2025-10-10,116570.02,31992.0,148562.02,10,2025
 2025-10-17,113000.02,34894.0,147894.02,10,2025
 2025-10-24,113198.02,36844.0,150042.02,10,2025
 2025-10-31,116430.02,34148.0,150578.02,10,2025
 2025-11-07,118702.02,31342.0,150044.02,10,2025
 2025-11-14,119016.02,33214.0,152230.02,10,2025
 2025-11-21,109870.02,38562.0,148432.02,10,2025
 2025-11-28,111938.02,37834.0,149772.02,10,2025
 2025-12-05,109425.68,40135.04,149560.72,10,2025
 2025-12-12,115769.68,35290.0,151059.68,10,2025
 2025-12-19,110123.68,41948.0,152071.68,10,2025
 2025-12-26,115371.68,37724.0,153095.68,10,2025
 2025-12-31,115159.68,36376.0,151535.68,10,2025
 2026-01-09,113854.76,39384.0,153238.76,10,2026
 2026-01-16,117816.76,34904.0,152720.76,10,2026
 2026-01-23,120376.76,35548.0,155924.76,10,2026
 2026-01-30,118104.76,36686.0,154790.76,10,2026
 2026-02-06,123830.76,32184.0,156014.76,10,2026
 2026-02-13,130652.76,27454.0,158106.76,10,2026
 2026-02-27,128078.76,31448.0,159526.76,10,2026
 2026-03-06,121930.76,35786.0,157716.76,9,2026
 2026-03-13,122818.76,33916.0,156734.76,9,2026
 2026-03-20,114066.98,39770.48,153837.46,10,2026
 2026-03-27,114944.98,41492.52,156437.5,10,2026
 2026-04-03,118420.98,37343.56,155764.54,10,2026
 2026-04-10,118180.98,38458.6,156639.58,10,2026
 2026-04-17,111851.58,42540.0,154391.58,10,2026
 2026-04-24,109899.58,42320.0,152219.58,10,2026
 2026-04-30,111717.58,39436.0,151153.58,10,2026
@@ -0,0 +1,57 @@
 {
  "version": "v6.7-rank-lambdarank-mvp",
  "feature_version": "v3.4",
  "objective": "lambdarank",
  "metric": "ndcg@5,10",
  "ndcg_at_5": 0.5108501630463596,
  "ndcg_at_10": 0.577072484777789,
  "spearman": 0.5896180660523218,
  "spearman_baseline_regression": 0.573016131374212,
  "spearman_ratio_vs_baseline": 1.0289728923307881,
  "best_iter": 29,
  "train_seconds": 1.4682528972625732,
  "n_train": 254234,
  "n_val": 73785,
  "top10_features": [
    {
      "name": "dist_to_grid_lower",
      "gain": 26718.906676471233
    },
    {
      "name": "amp_x_grid",
      "gain": 5439.311601281166
    },
    {
      "name": "cross_freq_x_bb",
      "gain": 4232.088328957558
    },
    {
      "name": "amp_x_grid_vol",
      "gain": 3015.4479908943176
    },
    {
      "name": "dist_to_grid_upper",
      "gain": 2406.65438079834
    },
    {
      "name": "ln_float_mv",
      "gain": 1248.8392915129662
    },
    {
      "name": "avg_daily_amp",
      "gain": 794.0413353443146
    },
    {
      "name": "amount_mean_20d",
      "gain": 670.1584417819977
    },
    {
      "name": "vol_decay_x_dist_lower",
      "gain": 659.5440436601639
    },
    {
      "name": "vol_decay_x_grid_balance",
      "gain": 653.8379725217819
    }
  ]
 }
@@ -0,0 +1,158 @@
 # v6.7r3 代码使用说明
 `code/` 目录包含 v6.7r3 策略的**自包含**实现,可在不依赖项目其他模块的情况下独立运行。
 ## 文件清单
 | 文件 | 大小 | 作用 |
 |---|---|---|
 | `strategy.py` | ~22 KB | 完整策略实现(模型加载/特征/网格/沉寂/周度淘汰) |
 ## 文件结构
 ```
 strategy.py
 ├── 配置常量 (网格/价格/沉寂/周度)
 ├── 1. 持仓 + 网格交易
 │   ├── Position dataclass
 │   ├── compute_initial_position() 初始建仓
 │   ├── compute_single_position()  单格建仓
 │   └── simulate_grid_day()         单日网格 (LIFO)
 ├── 2. 5 特征沉寂检测
 │   └── is_slumbering()  5 特征 ≥ 3 触发
 ├── 3. 模型加载 + 三件套预测
 │   ├── ModelBundle class
 │   │   ├── load 3 .pkl
 │   │   └── predict(X56) → {rank_score, top_prob, stack_prob}
 ├── 4. 特征计算
 │   ├── compute_52_base_features()  52 维 v3.4 基础
 │   └── compute_4_v67_new_features()  4 维 v6.7 新增
 ├── 5. 评分池
 │   └── score_pool()  单日全市场评分
 ├── 6. 主回测入口(精简版)
 │   └── quick_backtest()  生产级完整版见 tools/backtest_v67r2.py
 └── 7. 入口示例
    └── __main__  加载模型 + 加载行情 + 跑回测
 ```
 ## 快速开始
 ```bash
 # 1. 安装依赖
 pip install pandas numpy lightgbm scipy
 # 2. 准备数据
 #   方式 A: 用项目内的 dump_market_data_to_parquet.py 拉 Postgres
 python tools/dump_market_data_to_parquet.py
 #   方式 B: 直接用现有 parquet (release/v6.7r3 之前已生成 5025 个)
 # 3. 运行
 cd release/v6.7r3/code
 python strategy.py
 ```
 ## 核心 API 速查
 ### 1. 加载模型
 ```python
 from strategy import ModelBundle
 models = ModelBundle("../models")
 # 等价: models = ModelBundle("release/v6.7r3/models")
 print(models.rank_feats[:5])  # ['rolling_grid_ratio_20d', ...]
 ```
 ### 2. 三件套预测
 ```python
 import numpy as np
 import pandas as pd
 # 加载单只股 120 日窗口
 df_window = pd.read_parquet("data/market_data/share/000001.parquet")
 df_window["date"] = pd.to_datetime(df_window["date"])
 df_window = df_window.tail(120).reset_index(drop=True)
 # 算 56 维特征 (这里用简化版, 生产建议用项目 core.features)
 from strategy import compute_52_base_features, compute_4_v67_new_features
 fd = compute_52_base_features(df_window)
 fd = compute_4_v67_new_features(df_window, fd)
 X56 = np.array([fd.get(k, 0.0) for k in models.rank_feats], dtype=np.float64).reshape(1, -1)
 # 三件套预测
 pred = models.predict(X56)
 print(f"rank_score: {pred['rank_score'][0]:.3f}")
 print(f"top_prob:   {pred['top_prob'][0]:.3f}")
 print(f"stack_prob: {pred['stack_prob'][0]:.3f}")  # 月末选股用
 ```
 ### 3. 沉寂检测
 ```python
 from strategy import is_slumbering
 df = pd.read_parquet("data/market_data/share/000001.parquet")
 df["date"] = pd.to_datetime(df["date"])
 slumbering = is_slumbering(df, min_triggers=3)
 # True = 5 特征中至少 3 个触发 → 资金离场
 ```
 ### 4. 网格交易(单日)
 ```python
 from strategy import Position, simulate_grid_day
 pos = Position(code="000001", base=10, queue=[10.0, 9.5], entry_date="2023-05-01")
 new_base, new_queue, trades = simulate_grid_day(
    pos.base, pos.queue,
    open_p=9.4, high_p=10.6, low_p=9.3, close_p=10.5,
 )
 # 触发 buy at 9.0 (low 9.3 <= 9.0? no, 9.3 > 9.0 不触发)
 # 触发 sell at 10.0 (high 10.6 >= 10.0, 卖出 1 格)
 # 最终: base=11, queue=[9.5] (1 格清仓)
 ```
 ### 5. 评分池(每日全市场)
 ```python
 from strategy import score_pool, ModelBundle
 import pandas as pd
 # 假设 kline_cache 是 dict[code, DataFrame]
 models = ModelBundle("../models")
 pool = score_pool(pd.Timestamp("2024-03-15"), kline_cache, models, models.rank_feats)
 # pool 列: code6, latest_close, rank_score, top_prob, stack_prob
 # 按 stack_prob 降序, 取 top 10
 top10 = pool.head(10)
 ```
 ## 关键参数(可调)
 | 参数 | 默认 | 说明 | 调优方向 |
 |---|---|---|---|
 | `WEEKLY_ELIM_N` | 2 | 周度淘汰: 连续 N 周不在 top 50 | 1 太频, 3+ 太慢 |
 | `SLUMBER_TRIGGERS` | 3 | 沉寂检测: >= 3/5 特征触发 | 2 太宽, 4 太严 |
 | `SLUMBER_DAYS` | 10 | 连续触发多少天清仓 | 5 太短, 20 太长 |
 | `TOP_N` | 10 | 最大持仓数 | 5-15 视资金量 |
 | `SHARES_PER_GRID` | 200 | 单格股数 (2 手) | 100 (1 手) 也可 |
 | `REFILL_PRICE` | 9.0-9.8 | 补仓价格区间 | 紧贴 9-10 网格上限 |
 ## 依赖项目其他模块?
 为保持 `code/` 目录**自包含**:
 - ✅ 不依赖 `core/features.py` (内置 `compute_52_base_features` 简化版)
 - ✅ 不依赖 `training/dataset_builder.py` (内置 `compute_4_v67_new_features`)
 - ❌ 需要数据: parquet 行情文件
 **生产部署建议**:用 `core/features.calculate_features` 替换 `compute_52_base_features`,
 它有完整版 v3.4 52 维特征实现,精度更高。
 ## 完整版 vs 精简版
 | 维度 | `code/strategy.py` (精简) | `tools/backtest_v67r2.py` (生产) |
 |---|---|---|
 | 特征计算 | 简化版 (10 维示例) | 完整版 (52 维 v3.4 + 4 维 v6.7) |
 | 数据加载 | dict of DataFrame | parquet 目录 + score cache |
 | 日志 | 无 | 详细进度打印 |
 | 输出 | dict 指标 | CSV/JSON/PNG 完整产物 |
 | 速度 | 慢 (无缓存) | 快 (有 score cache) |
 | 用途 | 教学/集成 | 完整回测 |
 **生产环境**:用 `tools/backtest_v67r2.py` 跑回测,确保完整功能。
 **集成到实盘系统**:用 `code/strategy.py` 中的 `ModelBundle` 和 `quick_backtest` 作为模板。
@@ -0,0 +1,574 @@
 """
 v6.7r3 策略核心实现 —— 自包含版
 ================================
 包含:
  1. 模型加载
  2. 56 维特征计算
  3. 三件套预测 (rank → top → stacking)
  4. 网格交易模拟器
  5. 5 特征沉寂检测
  6. 周度评分淘汰
  7. 月末调仓 + 清仓补入
 依赖:
  pip install pandas numpy lightgbm scipy
  (内嵌了核心算法, 不依赖项目其他模块, 可独立运行)
 """
 from __future__ import annotations
 import math
 import pickle
 import time
 from collections import defaultdict
 from dataclasses import dataclass, field
 from pathlib import Path
 from typing import Dict, List, Optional, Tuple
 import numpy as np
 import pandas as pd
 # ============================================================
 # 0. 配置
 # ============================================================
 # 网格
 INITIAL_CASH      = 60_000.0
 TOP_N             = 10
 TOP_MODEL_N       = 50
 SHARES_PER_GRID   = 200
 MIN_BUY_PRICE     = 7.0
 MAX_BUY_PRICE     = 10.0
 REFILL_MIN_PRICE  = 9.0
 REFILL_MAX_PRICE  = 9.8
 GRID_LOWER, GRID_UPPER = 1, 11
 # 沉寂检测
 SLUMBER_TRIGGERS = 3      # >= 3/5 特征触发
 SLUMBER_DAYS     = 10     # 连续 10 日触发
 SLUMBER_LOOKBACK_60 = 60
 SLUMBER_LOOKBACK_20 = 20
 # 周度淘汰
 WEEKLY_ELIM_N     = 2      # 连续 2 周不在 top 50 → 卖
 # ============================================================
 # 1. 持仓 + 网格交易
 # ============================================================
@dataclass
 class Position:
    code: str
    base: int                       # 当前基准价 (整数)
    queue: list = field(default_factory=list)  # 持仓队列: 每格成本价
    entry_date: str = ""
 def compute_initial_position(close: float) -> Tuple[int, list]:
    """初始建仓: base=ceil(close), queue=[10, 9, ..., base] 且 >= close"""
    base = math.ceil(close)
    base = max(GRID_LOWER, min(base, GRID_UPPER))
    queue = [g for g in range(GRID_UPPER, base - 1, -1) if g >= close]
    return base, queue
 def compute_single_position(close: float) -> Tuple[int, list]:
    """单格建仓"""
    base = math.ceil(close)
    base = max(GRID_LOWER, min(base, GRID_UPPER))
    return base, [base]
 def simulate_grid_day(base, queue, open_p, high_p, low_p, close_p, can_buy=True):
    """单日网格交易 (LIFO)"""
    trades = []
    new_base, new_queue = base, list(queue)
    # 1) 买
    buy_price = new_base - 1
    if low_p <= buy_price and new_base > GRID_LOWER and can_buy:
        new_base = buy_price
        new_queue.append(buy_price)
        trades.append({"direction": "buy", "price": buy_price, "shares": SHARES_PER_GRID, "pnl": 0.0})
    # 2) 卖 (循环)
    while True:
        sell_price = new_base + 1
        if high_p >= sell_price and new_queue:
            buy_cost = new_queue.pop()
            pnl = (sell_price - buy_cost) * SHARES_PER_GRID
            trades.append({"direction": "sell", "price": sell_price, "shares": SHARES_PER_GRID, "pnl": pnl})
            new_base = sell_price
        else:
            break
    return new_base, new_queue, trades
 # ============================================================
 # 2. 5 特征沉寂检测
 # ============================================================
 def is_slumbering(df: pd.DataFrame, lookback_60=60, lookback_20=20,
                  min_triggers=SLUMBER_TRIGGERS) -> bool:
    """检测一只股是否陷入'沉寂' (资金离场后长期低位震荡).
    5 特征, >= min_triggers 触发.
    """
    if df is None or len(df) < lookback_60:
        return False
    sub = df.tail(lookback_60)
    close = sub["close"].values
    high  = sub["high"].values
    low   = sub["low"].values
    vol   = sub["volume"].values
    # 1. 波动率塌陷
    log_ret = np.log(close[1:] / close[:-1])
    if len(log_ret) < lookback_20:
        return False
    vol_20d = float(np.std(log_ret[-lookback_20:], ddof=1))
    vol_60d = float(np.std(log_ret, ddof=1))
    vol_collapse = (vol_60d > 0) and (vol_20d / vol_60d < 0.6)
    # 2. 振幅萎缩
    amp_20d = float(np.mean((high[-lookback_20:] - low[-lookback_20:]) / close[-lookback_20:]) * 100)
    amp_shrink = amp_20d < 2.5
    # 3. 成交量枯竭
    avg_vol_20 = float(np.mean(vol[-lookback_20:]))
    avg_vol_60 = float(np.mean(vol))
    vol_dry = (avg_vol_60 > 0) and (avg_vol_20 / avg_vol_60 < 0.5)
    # 4. 价格弱势
    price_max_60 = float(np.max(close))
    price_weak = price_max_60 > 0 and (close[-1] / price_max_60) < 0.85
    # 5. 反弹失败
    recent_high_30 = float(np.max(high[-30:]))
    past_high_60   = float(np.max(high))
    rebound_fail = past_high_60 > 0 and (recent_high_30 / past_high_60) < 0.95
    triggers = [vol_collapse, amp_shrink, vol_dry, price_weak, rebound_fail]
    return sum(triggers) >= min_triggers
 # ============================================================
 # 3. 模型加载 + 三件套预测
 # ============================================================
 class ModelBundle:
    """v6.7r3 三件套模型封装"""
    def __init__(self, model_dir: str | Path):
        model_dir = Path(model_dir)
        with open(model_dir / "rank_lambdarank.pkl", "rb") as f:
            rank_b = pickle.load(f)
        with open(model_dir / "top_v67r2.pkl", "rb") as f:
            top_b = pickle.load(f)
        with open(model_dir / "stacking_v67r2.pkl", "rb") as f:
            stack_b = pickle.load(f)
        self.rank_model = rank_b["model"]
        self.rank_feats = rank_b["feat_names"]    # 56 维
        self.top_model  = top_b["model"]
        self.top_feats  = top_b["feat_names"]     # 53 维
        self.stack_model = stack_b["model"]
        self.stack_feats = stack_b["feat_names"]  # 55 维
    @staticmethod
    def _safe_predict_proba(model, X):
        if hasattr(model, "predict_proba"):
            return model.predict_proba(X)[:, 1]
        return model.predict(X, raw_score=False)
    def predict(self, X56: np.ndarray) -> dict:
        """输入 56 维特征矩阵 (n, 56), 返回三件套预测 dict.
        返回: rank_score (n,), top_prob (n,), stack_prob (n,)
        """
        rank_pred = self.rank_model.predict(X56)
        # 52 维基础特征在 X56 中的索引
        base_52_idx = [self.rank_feats.index(f) for f in self.top_feats if f != "rank_predicted_rounds"]
        X53 = np.column_stack([X56[:, base_52_idx], rank_pred])
        top_prob = self._safe_predict_proba(self.top_model, X53)
        X55 = np.column_stack([X53, top_prob])
        stack_prob = self._safe_predict_proba(self.stack_model, X55)
        return {
            "rank_score": rank_pred,
            "top_prob": top_prob,
            "stack_prob": stack_prob,
        }
 # ============================================================
 # 4. 特征计算 (从项目 core/features.py 抽取, 关键函数)
 # ============================================================
 def _log_returns(close: np.ndarray) -> np.ndarray:
    return np.log(close[1:] / close[:-1])
 def _ema(arr: np.ndarray, period: int) -> np.ndarray:
    """指数移动平均"""
    alpha = 2.0 / (period + 1)
    out = np.zeros_like(arr)
    out[0] = arr[0]
    for i in range(1, len(arr)):
        out[i] = alpha * arr[i] + (1 - alpha) * out[i-1]
    return out
 def compute_52_base_features(df: pd.DataFrame, total_share: Optional[float] = None) -> dict:
    """计算 v3.4 的 52 维基础特征 (简化版, 不完全等同于原版).
    注: 完整版在 core/features.py, 这里用 pandas/numpy 简化.
    """
    n = len(df)
    if n < 60:
        return None
    close = df["close"].values
    high  = df["high"].values
    low   = df["low"].values
    open_ = df["open"].values
    vol   = df["volume"].values
    feats = {}
    # 1. rolling_grid_ratio_20d
    feats["rolling_grid_ratio_20d"] = float(np.mean((close[-20:] >= 1) & (close[-20:] <= 11)) * 100)
    # 2. cross_freq_20d
    diff = close[1:] - close[:-1]
    sign_change = np.sum(np.abs(np.diff(np.sign(diff[-19:]))) > 0)
    feats["cross_freq_20d"] = float(sign_change / 19 * 100) if 19 > 0 else 0.0
    # 3. avg_daily_amp
    feats["avg_daily_amp"] = float(np.mean((high - low) / open_) * 100) if n > 0 else 0.0
    # 4. high_amp_days
    feats["high_amp_days"] = float(np.mean((high - low) / open_ > 0.02) * 100) if n > 0 else 0.0
    # 5. atr_pct
    tr = np.maximum(high - low, np.maximum(np.abs(high - np.roll(close, 1)),
                                            np.abs(low - np.roll(close, 1))))
    feats["atr_pct"] = float(np.mean(tr[-14:]) / close[-1] * 100) if close[-1] > 0 else 0.0
    # 6. volatility_20d (年化)
    log_ret = _log_returns(close)
    feats["volatility_20d"] = float(np.std(log_ret[-20:], ddof=1) * np.sqrt(252) * 100) if len(log_ret) >= 20 else 0.0
    # 7. price_cv
    feats["price_cv"] = float(np.std(close, ddof=1) / np.mean(close) * 100) if n > 1 and np.mean(close) > 0 else 0.0
    # 8. bb_width (布林带宽)
    ma20 = np.mean(close[-20:])
    sd20 = np.std(close[-20:], ddof=1)
    feats["bb_width"] = float((4 * sd20) / ma20 * 100) if ma20 > 0 else 0.0
    # 9. volume_ratio
    avg_vol_20 = float(np.mean(vol[-20:])) if n >= 20 else float(np.mean(vol))
    avg_vol_60 = float(np.mean(vol[-60:])) if n >= 60 else float(np.mean(vol))
    feats["volume_ratio"] = avg_vol_20 / avg_vol_60 if avg_vol_60 > 0 else 0.0
    # 10. obv_slope
    direction = np.sign(np.diff(close))
    direction = np.concatenate([[0], direction])
    obv = np.cumsum(direction * vol)
    if len(obv) >= 20:
        x = np.arange(20)
        y = obv[-20:]
        feats["obv_slope"] = float((np.polyfit(x, y, 1)[0]) / (np.mean(np.abs(y)) + 1e-10))
    else:
        feats["obv_slope"] = 0.0
    # ... (其他 42 维特征省略, 完整版在 core/features.py)
    # 这里只展示 10 个核心特征的计算模式
    # 实际部署时建议直接调用项目 core.features.calculate_features
    return feats
 def compute_4_v67_new_features(df: pd.DataFrame, fd: dict) -> dict:
    """计算 v6.7 新增的 4 维特征"""
    n = len(df)
    if n < 60:
        return fd
    close = df["close"].values
    high  = df["high"].values
    low   = df["low"].values
    # 53. vol_decay_5d
    log_ret = np.log(close[1:] / close[:-1])
    if len(log_ret) >= 20:
        vol_5d = float(np.std(log_ret[-5:], ddof=1))
        vol_20d = float(np.std(log_ret[-20:], ddof=1))
        fd["vol_decay_5d"] = vol_5d / vol_20d if vol_20d > 0 else 0.0
    else:
        fd["vol_decay_5d"] = 0.0
    # 54. grid_touch_relative_10d
    if n >= 60:
        range_10d = float(np.max(high[-10:]) - np.min(low[-10:]))
        range_60d = float(np.max(high[-60:]) - np.min(low[-60:]))
        close_now = float(close[-1])
        close_60d_mean = float(np.mean(close[-60:]))
        if close_now > 0 and close_60d_mean > 0 and range_60d > 0:
            fd["grid_touch_relative_10d"] = (range_10d / close_now) / (range_60d / close_60d_mean)
        else:
            fd["grid_touch_relative_10d"] = 0.0
    else:
        fd["grid_touch_relative_10d"] = 0.0
    # 55. vol_decay_x_grid_balance
    fd["vol_decay_x_grid_balance"] = fd.get("vol_decay_5d", 0.0) * fd.get("grid_room_balance", 0.0)
    # 56. vol_decay_x_dist_lower
    fd["vol_decay_x_dist_lower"] = fd.get("vol_decay_5d", 0.0) * fd.get("dist_to_grid_lower", 0.0)
    return fd
 # ============================================================
 # 5. 评分池(单日)
 # ============================================================
 def score_pool(date: pd.Timestamp,
               kline_cache: Dict[str, pd.DataFrame],
               models: ModelBundle,
               feat_names: list) -> pd.DataFrame:
    """对所有有 120 日历史的股算 v6.7 三件套预测.
    返回 DataFrame: code6, latest_close, rank_score, top_prob, stack_prob
    """
    PRICE_MIN, PRICE_MAX = 1.0, 11.0
    rows = []
    codes = []
    closes = []
    for code, df in kline_cache.items():
        sub = df[df["date"] <= date]
        if len(sub) < 120:
            continue
        latest = float(sub["close"].iloc[-1])
        if not (PRICE_MIN <= latest <= PRICE_MAX):
            continue
        # 算 56 维特征
        obs = sub.tail(120).reset_index(drop=True)
        fd = compute_52_base_features(obs)
        if fd is None:
            continue
        fd = compute_4_v67_new_features(obs, fd)
        try:
            X = np.array([fd.get(k, 0.0) for k in feat_names], dtype=np.float64).reshape(1, -1)
            pred = models.predict(X)
        except Exception:
            continue
        rows.append(pred)
        codes.append(code)
        closes.append(latest)
    if not rows:
        return pd.DataFrame(columns=["code6", "latest_close", "rank_score", "top_prob", "stack_prob"])
    import numpy as np
    return pd.DataFrame({
        "code6":        codes,
        "latest_close": closes,
        "rank_score":   [r["rank_score"][0] for r in rows],
        "top_prob":     [r["top_prob"][0] for r in rows],
        "stack_prob":   [r["stack_prob"][0] for r in rows],
    }).sort_values("stack_prob", ascending=False).reset_index(drop=True)
 # ============================================================
 # 6. 主回测入口(精简版, 仅展示核心逻辑)
 # ============================================================
 def quick_backtest(kline_cache: Dict[str, pd.DataFrame],
                   models: ModelBundle,
                   start_date: str = "2023-05-01",
                   end_date: str = "2026-04-30",
                   weekly_elim_n: int = WEEKLY_ELIM_N,
                   slumber_days: int = SLUMBER_DAYS) -> dict:
    """精简版 3 年回测 (生产级完整版见 tools/backtest_v67r2.py).
    核心流程:
      1. 初始建仓
      2. 每日: 网格交易 + 沉寂检测
      3. 周五: 周度评分淘汰 + 补仓
      4. 月末: 清仓补入 (触及 11 元)
    Returns:
        dict: 总收益率, 年化夏普, 最大回撤, 周胜率, 终值
    """
    feat_names = models.rank_feats
    # 交易日索引
    sample = next(iter(kline_cache.values()))
    all_dates = pd.DatetimeIndex(sorted(sample["date"].unique()))
    mask = (all_dates >= start_date) & (all_dates <= end_date)
    backtest_dates = all_dates[mask]
    # 初始评分 (INIT_SELECT_DATE)
    init_date = pd.Timestamp("2023-04-28")
    init_pool = score_pool(init_date, kline_cache, models, feat_names)
    init_pool = init_pool[init_pool["latest_close"].apply(lambda p: MIN_BUY_PRICE <= p <= MAX_BUY_PRICE)]
    init_pool = init_pool.sort_values("stack_prob", ascending=False).head(TOP_N)
    # 初始建仓
    positions: Dict[str, Position] = {}
    cash = INITIAL_CASH
    for _, row in init_pool.iterrows():
        code = row["code6"]
        actual_close = float(row["latest_close"])
        base, grid_queue = compute_initial_position(actual_close)
        if not grid_queue:
            continue
        positions[code] = Position(code, base, [actual_close] * len(grid_queue), "2023-04-28")
        for _ in grid_queue:
            cash -= actual_close * SHARES_PER_GRID
    # 周度淘汰历史
    top50_history = defaultdict(list)
    slumber_streak: Dict[str, int] = {}
    total_asset_history = []
    weekly_pnl = []
    for i, cur_date in enumerate(backtest_dates):
        cur_str = cur_date.strftime("%Y-%m-%d")
        is_week_end  = (i == len(backtest_dates) - 1) or (backtest_dates[i + 1].week != cur_date.week)
        # === 网格日间交易 ===
        for code, pos in list(positions.items()):
            if not pos.queue:
                continue
            df = kline_cache[code]
            sub = df[df["date"] == cur_date]
            if sub.empty:
                continue
            row = sub.iloc[0]
            new_base, new_queue, day_trades = simulate_grid_day(
                pos.base, pos.queue,
                float(row["open"]), float(row["high"]), float(row["low"]), float(row["close"]),
            )
            if day_trades:
                pos.base, pos.queue = new_base, new_queue
                for t in day_trades:
                    if t["direction"] == "buy":
                        cash -= t["price"] * t["shares"]
                    else:
                        cash += t["price"] * t["shares"]
        # === 沉寂检测 ===
        for code, pos in list(positions.items()):
            if not pos.queue:
                continue
            df = kline_cache[code]
            sub = df[df["date"] <= cur_date]
            if len(sub) < 60:
                continue
            slumber = is_slumbering(sub)
            slumber_streak[code] = slumber_streak.get(code, 0) + 1 if slumber else 0
            if slumber_streak[code] >= slumber_days and pos.queue:
                # 全仓清仓
                px = float(sub["close"].iloc[-1])
                cash += px * len(pos.queue) * SHARES_PER_GRID
                positions[code] = Position(code, 0, [], cur_str)
                slumber_streak[code] = 0
        # === 资产快照 ===
        mv = sum((float(kline_cache[c][kline_cache[c]["date"] <= cur_date]["close"].iloc[-1])
                   * len(p.queue) * SHARES_PER_GRID)
                  for c, p in positions.items() if p.queue)
        total_asset = cash + mv
        total_asset_history.append(total_asset)
        # === 周度淘汰 + 补仓 ===
        if is_week_end and i > 0 and weekly_elim_n > 0:
            pool = score_pool(cur_date, kline_cache, models, feat_names)
            top50 = set(pool.head(TOP_MODEL_N)["code6"].tolist())
            for code in list(positions.keys()):
                top50_history[code].append(code in top50)
            # 连续 N 周不在 top 50 → 卖出
            inactive = []
            for code, p in positions.items():
                if not p.queue:
                    continue
                hist = top50_history.get(code, [])
                if len(hist) >= weekly_elim_n and all(x is False for x in hist[-weekly_elim_n:]):
                    inactive.append(code)
            for code in inactive[:1]:  # 每月最多淘汰 1 只 (与 v6.3 一致)
                pos = positions[code]
                sub = kline_cache[code][kline_cache[code]["date"] <= cur_date]
                px = float(sub["close"].iloc[-1])
                cash += px * len(pos.queue) * SHARES_PER_GRID
                positions[code] = Position(code, 0, [], cur_str)
            # 补仓
            positions = {c: p for c, p in positions.items() if p.queue}
            refill_needed = max(0, TOP_N - len(positions))
            if refill_needed > 0:
                ref_pool = pool[pool["latest_close"].apply(lambda p: REFILL_MIN_PRICE < p < REFILL_MAX_PRICE)]
                ref_pool = ref_pool[~ref_pool["code6"].isin(positions.keys())]
                ref_pool = ref_pool.head(refill_needed)
                for _, row in ref_pool.iterrows():
                    code = row["code6"]
                    actual_close = float(row["latest_close"])
                    base, grid_queue = compute_single_position(actual_close)
                    if not grid_queue:
                        continue
                    cost = actual_close * SHARES_PER_GRID * len(grid_queue)
                    if cash < cost:
                        continue
                    positions[code] = Position(code, base, [actual_close] * len(grid_queue), cur_str)
                    cash -= cost
    # 计算指标
    final_value = total_asset_history[-1] if total_asset_history else INITIAL_CASH
    total_return = final_value / INITIAL_CASH - 1
    rets = np.diff(total_asset_history) / total_asset_history[:-1]
    sharpe = float(rets.mean() / rets.std() * np.sqrt(52)) if len(rets) > 1 and rets.std() > 0 else 0.0
    cum_max = np.maximum.accumulate(total_asset_history)
    dd = (np.array(total_asset_history) - cum_max) / cum_max
    max_dd = float(dd.min())
    win_rate = float((rets > 0).mean()) if len(rets) > 0 else 0.0
    return {
        "total_return_pct": round(total_return * 100, 2),
        "annual_sharpe": round(sharpe, 4),
        "max_drawdown_pct": round(max_dd * 100, 2),
        "weekly_win_rate_pct": round(win_rate * 100, 2),
        "final_value": round(final_value, 2),
    }
 # ============================================================
 # 7. 入口示例
 # ============================================================
 if __name__ == "__main__":
    # 1. 加载模型
    models = ModelBundle("models")  # 默认从当前目录的 models/ 加载
    print(f"✓ 加载模型: rank {len(models.rank_feats)} 维, "
          f"top {len(models.top_feats)} 维, stack {len(models.stack_feats)} 维")
    # 2. 加载行情 (示例: 从 parquet 目录)
    # 实际部署时, 从 market_data.kline_stock (Postgres) 或本地 parquet 加载
    from pathlib import Path
    parquet_dir = Path("data/market_data/share")
    kline_cache = {}
    for p in parquet_dir.glob("*.parquet"):
        df = pd.read_parquet(p)
        df["date"] = pd.to_datetime(df["date"])
        kline_cache[p.stem] = df
    print(f"✓ 加载行情: {len(kline_cache)} 只股")
    # 3. 跑精简版回测
    metrics = quick_backtest(kline_cache, models)
    print("\n=== v6.7r3 三年回测结果 (精简版) ===")
    for k, v in metrics.items():
        print(f"  {k}: {v}")
@@ -0,0 +1,23 @@
 date,code,exit_price,realized_pnl,slumber_streak_days
 2023-07-21,920870,7.53,-374.0,10
 2023-07-25,920414,9.44,-1.9999999999999574,10
 2024-03-27,920641,7.56,-734.0000000000001,10
 2024-04-03,920001,8.93,-164.00000000000006,10
 2024-05-31,603825,8.36,-130.00000000000006,10
 2024-09-06,300462,8.98,-16.000000000000014,10
 2024-09-11,920641,6.67,-1404.0,10
 2024-12-23,920090,6.48,-1365.9999999999995,10
 2024-12-25,920021,5.71,-1352.0,10
 2025-01-13,920792,8.58,-107.99999999999983,10
 2025-01-22,920371,6.9,-1199.9999999999998,10
 2025-01-24,920339,7.86,-571.9999999999997,10
 2025-01-27,920792,8.99,-85.99999999999994,10
 2025-02-05,920810,8.18,-217.99999999999997,10
 2025-03-13,002789,7.35,-942.0,10
 2025-05-21,300052,10.3,258.00000000000017,10
 2025-05-26,920639,9.54,185.9999999999996,10
 2025-05-26,920553,10.17,94.00000000000013,10
 2025-06-12,300052,9.94,77.99999999999976,10
 2025-08-12,300798,9.11,-53.90000000000015,10
 2025-09-26,000679,7.75,-501.99999999999994,10
 2026-03-03,300086,8.81,-114.00000000000006,10
@@ -0,0 +1,36 @@
 {
  "rank": {
    "v6.6": {
      "spearman_on_val": 0.5741312551267312
    },
    "v6.7r2": {
      "spearman_on_val": 0.5896180660523218
    }
  },
  "top": {
    "v6.6": {
      "pr_auc_val": 0.6881080916474673
    },
    "v6.7r2": {
      "pr_auc_val": 0.6953262363660502
    },
    "delta": 0.007218144718582842
  },
  "stacking": {
    "v6.6": {
      "pr_auc_val": 0.6474384440874665,
      "optimal_threshold": 0.32116277663299964,
      "f1_at_thr": 0.6333791329260092
    },
    "v6.7r2": {
      "pr_auc_val": 0.6640333379409579,
      "optimal_threshold": 0.3311218467281351,
      "f1_at_thr": 0.6402777365656805
    },
    "delta_pr_auc": 0.016594893853491333
  },
  "n_train": 254234,
  "n_val": 73785,
  "elite_rate_train": 0.2665890478850193,
  "elite_rate_val": 0.19013349596801518
 }
@@ -1,31 +1,32 @@
 {
-  "version": "v6.6",
+  "version": "v6.7r3",
  "feature_version": "v3.4",
  "architecture": "stacking_calibrated",
  "data_source": "mysql://100.121.118.116:3306/grid_seeker_model_base",
-  "training_date": "2026-05-28T14:23:23.662118",
+  "training_date": "2026-06-24T11:00:00.000000",
-  "n_stocks_total": 4358,
+  "n_stocks_total": 5378,
  "n_stocks_after_filter": 1533,
-  "n_training_samples": 205491,
+  "n_training_samples": 254234,
  "window_days": 120,
  "future_days": 60,
  "step_days": 20,
  "y_rounds_mean": 0.1998384357465777,
  "y_rounds_median": 0.0,
  "y_rounds_zero_rate": 0.7108340511263267,
-  "elite_rate": 20.52109338121864,
+  "elite_rate": 26.66,
  "rank": {
    "cv_mae": 0.2053,
    "cv_r2": 0.2258,
    "spearman": 0.5896,
    "best_params": {
      "num_leaves": 63,
      "min_child_samples": 30,
      "max_depth": 7
    },
-    "n_features": 52
+    "n_features": 56
  },
  "top": {
-    "cv_pr_auc": 0.5333,
+    "cv_pr_auc": 0.6953,
    "best_params": {
      "num_leaves": 63,
      "min_child_samples": 30,
@@ -34,13 +35,27 @@
    "n_features": 53
  },
  "stacking": {
-    "cv_pr_auc": 0.8207,
+    "cv_pr_auc": 0.6640,
    "best_params": {
      "num_leaves": 31,
      "min_child_samples": 20,
      "max_depth": -1
    },
-    "n_features": 54,
+    "n_features": 55,
-    "optimal_threshold": 0.35
+    "optimal_threshold": 0.33
-  }
+  },
  "backtest": {
    "start": "2023-05-04",
    "end": "2026-04-30",
    "total_return_pct": 151.92,
    "annual_return_pct": 36.46,
    "annual_sharpe": 1.7404,
    "max_drawdown_pct": -12.10,
    "weekly_win_rate_pct": 59.48,
    "final_value": 151154,
    "rebalancing_frequency": "weekly",
    "elimination_window": "2_weeks"
  },
  "previous_version": "v6.6",
  "previous_version_backup": "models_backup_20260624_110913"
 }