diff --git a/app/__init__.py b/app/__init__.py
index 96dbc9c94..c05708298 100644
--- a/app/__init__.py
+++ b/app/__init__.py
@@ -37,6 +37,9 @@ def create_app(config_name='default'):
     from app.api.data_jobs_api import data_jobs_bp
     from app.routes.ml_factor_routes import ml_factor_routes
     from app.routes.realtime_analysis_routes import realtime_analysis_routes
+    from app.routes.heatmap import heatmap_routes
+    from app.routes.pattern_screen import pattern_screen_bp
+    from app.api.pattern_screen_api import pattern_screen_api
     app.register_blueprint(api_bp, url_prefix='/api')
     app.register_blueprint(ml_factor_bp)
     app.register_blueprint(text2sql_bp)
@@ -50,7 +53,10 @@ def create_app(config_name='default'):
     app.register_blueprint(data_jobs_bp)
     app.register_blueprint(ml_factor_routes)
     app.register_blueprint(realtime_analysis_routes)
-    
+    app.register_blueprint(heatmap_routes)
+    app.register_blueprint(pattern_screen_bp)
+    app.register_blueprint(pattern_screen_api)
+
     from app.main import main_bp
     app.register_blueprint(main_bp)
     
diff --git a/app/api/pattern_screen_api.py b/app/api/pattern_screen_api.py
new file mode 100644
index 000000000..4de13a87f
--- /dev/null
+++ b/app/api/pattern_screen_api.py
@@ -0,0 +1,47 @@
+"""形态选股 API 端点。"""
+from flask import Blueprint, request, jsonify
+from loguru import logger
+
+pattern_screen_api = Blueprint('pattern_screen_api', __name__, url_prefix='/api/pattern-screen')
+
+
+@pattern_screen_api.route('/groups', methods=['GET'])
+def get_groups():
+    """返回形态分组元数据（含命中数）。"""
+    try:
+        from app.services.pattern_screen_service import get_pattern_screen_service
+        svc = get_pattern_screen_service()
+        groups = svc.get_groups()
+        return jsonify({'code': 200, 'message': '成功', 'data': groups})
+    except Exception as e:
+        logger.error(f"获取形态分组失败: {e}")
+        return jsonify({'code': 500, 'message': f'服务器错误: {str(e)}', 'data': None}), 500
+
+
+@pattern_screen_api.route('/screen', methods=['POST'])
+def screen():
+    """执行形态筛选，返回结果表格。"""
+    try:
+        data = request.get_json() or {}
+        patterns = data.get('patterns', [])
+        sort_by = data.get('sort_by', 'pct_chg')
+        order = data.get('order', 'desc')
+        limit = data.get('limit', 50)
+        offset = data.get('offset', 0)
+
+        from app.services.pattern_screen_service import get_pattern_screen_service
+        svc = get_pattern_screen_service()
+        result = svc.screen(
+            patterns=patterns,
+            sort_by=sort_by,
+            order=order,
+            limit=limit,
+            offset=offset,
+        )
+        return jsonify({'code': 200, 'message': '成功', 'data': result})
+    except ValueError as e:
+        logger.warning(f"形态筛选参数错误: {e}")
+        return jsonify({'code': 400, 'message': str(e), 'data': None}), 400
+    except Exception as e:
+        logger.error(f"形态筛选失败: {e}")
+        return jsonify({'code': 500, 'message': f'服务器错误: {str(e)}', 'data': None}), 500
diff --git a/app/routes/heatmap.py b/app/routes/heatmap.py
new file mode 100644
index 000000000..9a089effb
--- /dev/null
+++ b/app/routes/heatmap.py
@@ -0,0 +1,30 @@
+"""板块热力图页面路由。"""
+from flask import Blueprint, render_template
+from loguru import logger
+from app.services.heatmap_service import HeatmapService
+
+heatmap_routes = Blueprint('heatmap_routes', __name__, url_prefix='/heatmap')
+
+
+@heatmap_routes.route('/')
+def index():
+    """板块热力图页面。"""
+    try:
+        service = HeatmapService()
+        sectors, stocks = service.get_heatmap_data()
+        trade_date = sectors[0]['trade_date'] if sectors else ''
+        return render_template(
+            'heatmap.html',
+            sectors_json=sectors,
+            stocks_json=stocks,
+            trade_date=trade_date,
+        )
+    except Exception as e:
+        logger.error(f"热力图加载失败: {e}")
+        return render_template(
+            'heatmap.html',
+            sectors_json=[],
+            stocks_json=[],
+            trade_date='',
+            error='数据加载失败，请确认 data/data.parquet 是否存在',
+        )
diff --git a/app/routes/pattern_screen.py b/app/routes/pattern_screen.py
new file mode 100644
index 000000000..5b56eee08
--- /dev/null
+++ b/app/routes/pattern_screen.py
@@ -0,0 +1,10 @@
+"""形态选股页面路由。"""
+from flask import Blueprint, render_template
+
+pattern_screen_bp = Blueprint('pattern_screen', __name__)
+
+
+@pattern_screen_bp.route('/pattern-screen/')
+def index():
+    """形态选股页面。"""
+    return render_template('pattern_screen.html')
diff --git a/app/services/heatmap_service.py b/app/services/heatmap_service.py
new file mode 100644
index 000000000..889129cfd
--- /dev/null
+++ b/app/services/heatmap_service.py
@@ -0,0 +1,57 @@
+"""板块热力图数据服务 — 读取 data/data.parquet 并按行业聚合。"""
+import os
+import pandas as pd
+import numpy as np
+from loguru import logger
+
+
+class HeatmapService:
+    """板块热力图数据聚合服务。"""
+
+    def __init__(self):
+        self._data_path = os.path.join(
+            os.path.dirname(os.path.dirname(os.path.dirname(__file__))),
+            'data', 'data.parquet'
+        )
+
+    def get_heatmap_data(self):
+        """读取 parquet，返回 (sectors_json, stocks_json)。
+
+        Returns:
+            tuple: (list[dict], list[dict])
+                - sectors: 板块聚合数据，按 total_mv 降序
+                - stocks: 全量个股数据（已过滤停牌）
+        """
+        df = pd.read_parquet(self._data_path)
+        trade_date = str(df['trade_date'].iloc[0])
+
+        # 过滤停牌/退市
+        df = df[df['close'] > 0].copy()
+
+        # 板块聚合
+        sectors = []
+        for industry, group in df.groupby('industry'):
+            total_mv_sum = group['total_mv'].sum()
+            if total_mv_sum == 0:
+                continue
+            avg_pct_chg = np.average(group['pct_chg'], weights=group['total_mv'])
+            sectors.append({
+                'name': industry,
+                'avg_pct_chg': round(avg_pct_chg, 2),
+                'total_mv': round(total_mv_sum, 2),
+                'stock_count': len(group),
+                'up_count': int((group['pct_chg'] > 0).sum()),
+                'down_count': int((group['pct_chg'] < 0).sum()),
+                'net_mf_amount': round(group['net_mf_amount'].sum(), 2),
+                'trade_date': trade_date,
+            })
+
+        sectors.sort(key=lambda x: x['total_mv'], reverse=True)
+
+        # 个股明细
+        stock_cols = ['ts_code', 'name', 'industry', 'pct_chg', 'close',
+                      'total_mv', 'net_mf_amount', 'turnover_rate']
+        stocks_df = df[stock_cols].sort_values('pct_chg', ascending=False)
+        stocks = stocks_df.where(stocks_df.notna(), None).to_dict(orient='records')
+
+        return sectors, stocks
diff --git a/app/services/pattern_screen_service.py b/app/services/pattern_screen_service.py
new file mode 100644
index 000000000..a50e205ac
--- /dev/null
+++ b/app/services/pattern_screen_service.py
@@ -0,0 +1,345 @@
+"""Pattern screen service for stock pattern filtering and metadata.
+
+Provides pattern group metadata and AND-filtered screening across pattern columns
+in the market data parquet file.
+"""
+
+from flask import current_app
+import pandas as pd
+import numpy as np
+
+
+# Pattern group metadata with full field list
+PATTERN_GROUPS = [
+    {
+        'id': 'single_candle',
+        'label': '单根K线',
+        'fields': [
+            {'key': 'pattern_bull_candle', 'label': '阳线'},
+            {'key': 'pattern_bear_candle', 'label': '阴线'},
+            {'key': 'pattern_hammer', 'label': '锤头线'},
+            {'key': 'pattern_doji', 'label': '十字星'},
+            {'key': 'pattern_spinning_top', 'label': '纺锤线'},
+            {'key': 'pattern_shooting_star', 'label': '流星线'},
+            {'key': 'pattern_long_upper_shadow', 'label': '长上影线'},
+            {'key': 'pattern_long_lower_shadow', 'label': '长下影线'},
+            {'key': 'pattern_gravestone_doji', 'label': '墓碑十字'},
+            {'key': 'pattern_dragonfly_doji', 'label': '蜻蜓十字'},
+            {'key': 'pattern_hanging_man', 'label': '上吊线'},
+            {'key': 'pattern_inverted_hammer', 'label': '倒锤头'},
+            {'key': 'pattern_big_bull', 'label': '大阳线'},
+            {'key': 'pattern_big_bear', 'label': '大阴线'},
+            {'key': 'pattern_medium_bull', 'label': '中阳线'},
+            {'key': 'pattern_medium_bear', 'label': '中阴线'},
+            {'key': 'pattern_small_bull', 'label': '小阳线'},
+            {'key': 'pattern_small_bear', 'label': '小阴线'},
+            {'key': 'pattern_no_body', 'label': '无实体线'},
+            {'key': 'pattern_no_upper_bull', 'label': '光头阳线'},
+            {'key': 'pattern_no_upper_bear', 'label': '光头阴线'},
+            {'key': 'pattern_no_lower_bull', 'label': '光脚阳线'},
+            {'key': 'pattern_no_lower_bear', 'label': '光脚阴线'},
+            {'key': 'pattern_t_shape', 'label': 'T字线'},
+            {'key': 'pattern_inverted_t_shape', 'label': '倒T字线'},
+            {'key': 'pattern_low_open_high_close', 'label': '低开高走'},
+            {'key': 'pattern_high_open_low_close', 'label': '高开低走'},
+            {'key': 'pattern_gap_up', 'label': '跳空高开'},
+            {'key': 'pattern_gap_down', 'label': '跳空低开'},
+            {'key': 'pattern_close_above_prev_close', 'label': '收盘站上前收'},
+            {'key': 'pattern_close_below_prev_close', 'label': '收盘跌破前收'},
+            {'key': 'pattern_gap_reclaim_prev_close', 'label': '低开收回前收'},
+            {'key': 'pattern_gap_fade_below_prev_close', 'label': '高开回落失守前收'},
+            {'key': 'pattern_close_high', 'label': '收盘近最高'},
+            {'key': 'pattern_flat_open_high_close', 'label': '平开高走'},
+            {'key': 'pattern_flat_open_low_close', 'label': '平开低走'},
+            {'key': 'pattern_gap_up_close_bull', 'label': '高开收阳'},
+            {'key': 'pattern_gap_down_close_bear', 'label': '低开收阴'},
+            {'key': 'pattern_open_near_high_close_high', 'label': '开盘即最高附近收盘'},
+            {'key': 'pattern_open_near_low_close_low', 'label': '开盘即最低附近收盘'},
+            {'key': 'pattern_flat_open', 'label': '平开'},
+            {'key': 'pattern_gap_up_fill', 'label': '高开补缺'},
+            {'key': 'pattern_gap_down_fill', 'label': '低开补缺'},
+            {'key': 'pattern_pin_bar', 'label': 'Pin Bar'},
+            {'key': 'pattern_reversal_prelude', 'label': '反包前兆'},
+            {'key': 'pattern_high_resistance', 'label': '高位受阻'},
+            {'key': 'pattern_low_stabilization', 'label': '低位止跌'},
+            {'key': 'pattern_break_prev_high', 'label': '向上突破前高'},
+            {'key': 'pattern_break_prev_low', 'label': '向下跌破前低'},
+            {'key': 'pattern_false_break', 'label': '假突破回落'},
+            {'key': 'pattern_false_breakdown_recovery', 'label': '假跌破回升'},
+        ]
+    },
+    {
+        'id': 'double_candle',
+        'label': '双根K线',
+        'fields': [
+            {'key': 'pattern_bullish_engulfing', 'label': '阳包阴'},
+            {'key': 'pattern_bearish_engulfing', 'label': '阴包阳'},
+            {'key': 'pattern_inside_bar', 'label': '孕线'},
+            {'key': 'pattern_dark_cloud', 'label': '乌云盖顶'},
+            {'key': 'pattern_piercing', 'label': '刺透形态'},
+            {'key': 'pattern_tweezer_top', 'label': '镊子顶'},
+            {'key': 'pattern_tweezer_bottom', 'label': '镊子底'},
+            {'key': 'pattern_gap_break', 'label': '跳空上攻'},
+            {'key': 'pattern_gap_down_break', 'label': '跳空下跌'},
+            {'key': 'pattern_gap_up_no_fill', 'label': '跳空不补上行'},
+            {'key': 'pattern_gap_down_no_fill', 'label': '跳空不补下行'},
+            {'key': 'pattern_reversal_bar', 'label': '反转包线'},
+            {'key': 'pattern_flat_top', 'label': '平头顶'},
+            {'key': 'pattern_flat_bottom', 'label': '平头底'},
+            {'key': 'pattern_island_reversal', 'label': '岛形反转'},
+            {'key': 'pattern_t_limit', 'label': 'T字板'},
+            {'key': 'pattern_limit_reversal_wrap', 'label': '涨停反包'},
+            {'key': 'pattern_vol_up', 'label': '放量上涨'},
+            {'key': 'pattern_vol_down', 'label': '缩量下跌'},
+            {'key': 'pattern_double_volume_bar', 'label': '倍量柱'},
+            {'key': 'pattern_breakout_volume_confirm', 'label': '突破放量确认'},
+        ]
+    },
+    {
+        'id': 'triple_candle',
+        'label': '三根K线',
+        'fields': [
+            {'key': 'pattern_morning_star', 'label': '早晨之星'},
+            {'key': 'pattern_evening_star', 'label': '黄昏之星'},
+            {'key': 'pattern_morning_doji_star', 'label': '启明星'},
+            {'key': 'pattern_evening_doji_star', 'label': '黄昏十字星'},
+            {'key': 'pattern_three_black_crows', 'label': '三只乌鸦'},
+            {'key': 'pattern_red_three', 'label': '红三兵'},
+            {'key': 'pattern_three_outside_up', 'label': '三外升'},
+            {'key': 'pattern_three_outside_down', 'label': '三外降'},
+            {'key': 'pattern_rising_three_methods', 'label': '上升三法'},
+            {'key': 'pattern_falling_three_methods', 'label': '下降三法'},
+            {'key': 'pattern_three_up', 'label': '三连阳'},
+            {'key': 'pattern_three_down', 'label': '三连阴'},
+            {'key': 'pattern_three_yang_kaitai', 'label': '三阳开泰'},
+            {'key': 'pattern_three_yin_breakdown', 'label': '三阴破位'},
+        ]
+    },
+    {
+        'id': 'trend_structure',
+        'label': '趋势结构',
+        'fields': [
+            {'key': 'pattern_up_trend', 'label': '上升趋势'},
+            {'key': 'pattern_down_trend', 'label': '下降趋势'},
+            {'key': 'pattern_sideways', 'label': '横盘整理'},
+            {'key': 'pattern_golden_cross', 'label': '均线金叉'},
+            {'key': 'pattern_duck_head', 'label': '老鸭头'},
+            {'key': 'pattern_double_bottom', 'label': '双底'},
+            {'key': 'pattern_arc_bottom', 'label': '圆弧底'},
+            {'key': 'pattern_ma_bull', 'label': '均线多头排列'},
+            {'key': 'pattern_high_tight', 'label': '高位强势整理'},
+            {'key': 'pattern_pullback_hold', 'label': '回踩不破'},
+            {'key': 'pattern_trend_continue', 'label': '趋势中继'},
+            {'key': 'pattern_ma_spread_bull', 'label': '均线发散多头'},
+        ]
+    },
+    {
+        'id': 'volume_price',
+        'label': '量价关系',
+        'fields': [
+            {'key': 'pattern_box_breakout', 'label': '箱体放量突破'},
+            {'key': 'pattern_vol_price_up', 'label': '量价齐升'},
+            {'key': 'pattern_platform_break', 'label': '平台突破'},
+            {'key': 'pattern_triangle_squeeze', 'label': '三角收敛突破'},
+            {'key': 'pattern_limit_turnover_strong', 'label': '涨停换手强'},
+            {'key': 'pattern_price_volume_bear_divergence', 'label': '价量顶背离'},
+            {'key': 'pattern_price_volume_bull_divergence', 'label': '价量底背离'},
+            {'key': 'pattern_price_down_volume_up', 'label': '价跌量增'},
+            {'key': 'pattern_volume_staircase', 'label': '量能阶梯'},
+            {'key': 'pattern_pullback_volume_shrink', 'label': '回调缩量'},
+            {'key': 'pattern_high_turnover', 'label': '高换手'},
+            {'key': 'pattern_limit_up_volume_shrink', 'label': '一字板缩量'},
+            {'key': 'pattern_false_breakout_volume_weak', 'label': '假突破量弱'},
+            {'key': 'pattern_floor_volume_price', 'label': '地量价稳'},
+            {'key': 'pattern_blowoff_volume_price', 'label': '天量滞涨'},
+            {'key': 'pattern_v_reversal', 'label': 'V型反转'},
+        ]
+    },
+    {
+        'id': 'compound',
+        'label': '复合形态',
+        'fields': [
+            {'key': 'pattern_first_limit', 'label': '首板'},
+            {'key': 'pattern_multi_limit', 'label': '连板'},
+            {'key': 'pattern_one_word_limit', 'label': '一字板'},
+            {'key': 'pattern_limit_down_to_up', 'label': '地天板'},
+            {'key': 'pattern_lotus_breakout', 'label': '莲花突破'},
+            {'key': 'pattern_midway_refuel', 'label': '中继加油'},
+            {'key': 'pattern_consolidation_platform', 'label': '整理平台'},
+            {'key': 'pattern_n_breakout', 'label': 'N字突破'},
+            {'key': 'pattern_gap_breakaway', 'label': '跳空突破'},
+            {'key': 'pattern_channel_breakout', 'label': '通道突破'},
+            {'key': 'pattern_flag_breakout', 'label': '旗形突破'},
+        ]
+    },
+    {
+        'id': 'momentum',
+        'label': '动量突破',
+        'fields': [
+            {'key': 'break_high_20', 'label': '突破20日新高'},
+            {'key': 'break_high_60', 'label': '突破60日新高'},
+            {'key': 'break_high_120', 'label': '突破120日新高'},
+            {'key': 'break_high_250', 'label': '突破250日新高'},
+            {'key': 'consec_up_3', 'label': '连续上涨3日'},
+            {'key': 'consec_up_5', 'label': '连续上涨5日'},
+        ]
+    },
+]
+
+# Allowed sort columns
+SORT_WHITELIST = [
+    "ts_code", "pct_chg", "close", "amount", "total_mv",
+    "turnover_rate", "vol_ratio_5", "consec_up_days"
+]
+
+# Columns to return in screen results
+DISPLAY_COLUMNS = [
+    "ts_code", "name", "industry", "pct_chg", "close", "amount",
+    "total_mv", "turnover_rate", "vol_ratio_5"
+]
+
+
+class PatternScreenService:
+    """Service for pattern metadata and AND-filtered screening."""
+
+    def __init__(self):
+        self._df = None
+
+    def _load_df(self) -> pd.DataFrame:
+        """Load market data from parquet file."""
+        data_dir = current_app.config.get('DATA_DIR', 'data')
+        path = f"{data_dir}/data.parquet"
+        return pd.read_parquet(path)
+
+    def _ensure_df(self):
+        """Lazy-load DataFrame if not cached."""
+        if self._df is None:
+            self._df = self._load_df()
+
+    def get_groups(self) -> list:
+        """Return pattern groups with hit counts filtered by DataFrame columns.
+
+        Returns:
+            List of group dicts with id, label, and fields (each field has
+            key, label, count). Fields whose columns are not present in the
+            DataFrame are excluded.
+        """
+        self._ensure_df()
+        df_cols = set(self._df.columns)
+
+        groups = []
+        for group in PATTERN_GROUPS:
+            fields = []
+            for field in group['fields']:
+                if field['key'] in df_cols:
+                    count = int(self._df[field['key']].fillna(0).sum())
+                    fields.append({
+                        'key': field['key'],
+                        'label': field['label'],
+                        'count': count
+                    })
+            if fields:
+                groups.append({
+                    'id': group['id'],
+                    'label': group['label'],
+                    'fields': fields
+                })
+        return groups
+
+    def screen(
+        self,
+        patterns: list = None,
+        sort_by: str = "pct_chg",
+        order: str = "desc",
+        limit: int = 20,
+        offset: int = 0
+    ) -> dict:
+        """Screen stocks by pattern filters with AND logic.
+
+        Args:
+            patterns: List of pattern column keys to filter (AND logic).
+            sort_by: Column to sort by (must be in SORT_WHITELIST).
+            order: 'asc' or 'desc'.
+            limit: Max rows to return (capped at 500).
+            offset: Rows to skip for pagination.
+
+        Returns:
+            Dict with keys: total, offset, limit, trade_date, rows (list of dicts).
+
+        Raises:
+            ValueError: If sort_by, order, or pattern keys are invalid.
+        """
+        if patterns is None:
+            patterns = []
+
+        self._ensure_df()
+        df = self._df.copy()
+
+        # Validate sort_by
+        if sort_by not in SORT_WHITELIST:
+            raise ValueError(f"sort_by must be one of {SORT_WHITELIST}, got: {sort_by}")
+
+        # Validate order
+        if order not in ('asc', 'desc'):
+            raise ValueError(f"order must be 'asc' or 'desc', got: {order}")
+
+        # Validate pattern keys
+        df_cols = set(df.columns)
+        for p in patterns:
+            if p not in df_cols:
+                raise ValueError(f"Pattern '{p}' not found in data")
+
+        # Apply AND filter: all selected patterns must be 1
+        if patterns:
+            mask = df[patterns].fillna(0).eq(1).all(axis=1)
+            df = df[mask]
+
+        # Get trade_date from first row (if any)
+        trade_date = None
+        if not df.empty and 'trade_date' in df.columns:
+            trade_date = str(df['trade_date'].iloc[0])
+
+        # Sort
+        ascending = (order == 'asc')
+        df = df.sort_values(by=sort_by, ascending=ascending)
+
+        # Total before pagination
+        total = len(df)
+
+        # Cap limit
+        limit = min(limit, 500)
+
+        # Select display columns (only those present)
+        cols_to_show = [c for c in DISPLAY_COLUMNS if c in df.columns]
+        df = df[cols_to_show]
+
+        # Paginate
+        df = df.iloc[offset:offset + limit]
+
+        # Convert to dict of dicts, NaN -> None
+        df = df.replace({np.nan: None})
+        rows = df.to_dict(orient='records')
+
+        return {
+            'total': total,
+            'offset': offset,
+            'limit': limit,
+            'trade_date': trade_date,
+            'rows': rows
+        }
+
+    def invalidate_cache(self):
+        """Clear cached DataFrame to force reload on next access."""
+        self._df = None
+
+
+# Module-level singleton
+_service = None
+
+
+def get_pattern_screen_service() -> PatternScreenService:
+    """Get or create the singleton PatternScreenService instance."""
+    global _service
+    if _service is None:
+        _service = PatternScreenService()
+    return _service
diff --git a/app/templates/base.html b/app/templates/base.html
index c31193e08..4c8d2cf80 100644
--- a/app/templates/base.html
+++ b/app/templates/base.html
@@ -331,6 +331,16 @@
                             <i class="fas fa-filter me-1"></i>选股筛选
                         </a>
                     </li>
+                    <li class="nav-item">
+                        <a class="nav-link" href="{{ url_for('heatmap_routes.index') }}">
+                            <i class="fas fa-th-large me-1"></i>板块热力图
+                        </a>
+                    </li>
+                    <li class="nav-item">
+                        <a class="nav-link" href="{{ url_for('pattern_screen.index') }}">
+                            <i class="fas fa-crosshairs me-1"></i>形态选股
+                        </a>
+                    </li>
                     <li class="nav-item dropdown">
                         <a class="nav-link dropdown-toggle" href="#" role="button" data-bs-toggle="dropdown">
                             <i class="fas fa-brain me-1"></i>多因子模型
diff --git a/app/templates/heatmap.html b/app/templates/heatmap.html
new file mode 100644
index 000000000..588a63dbf
--- /dev/null
+++ b/app/templates/heatmap.html
@@ -0,0 +1,264 @@
+{% extends "base.html" %}
+
+{% block title %}板块热力图{% endblock %}
+
+{% block extra_css %}
+<style>
+.heatmap-page { padding: 20px 0; }
+.heatmap-header {
+    display: flex; align-items: center; justify-content: space-between;
+    margin-bottom: 16px; flex-wrap: wrap; gap: 12px;
+}
+.heatmap-title { font-size: 20px; font-weight: 600; color: #e2e8f0; }
+.heatmap-date { color: #94a3b8; font-size: 14px; }
+.sort-btns { display: flex; gap: 6px; }
+.sort-btn {
+    padding: 4px 12px; border-radius: 4px; border: 1px solid #334155;
+    background: transparent; color: #94a3b8; font-size: 12px; cursor: pointer;
+    transition: all 0.2s;
+}
+.sort-btn:hover { border-color: #6366f1; color: #e2e8f0; }
+.sort-btn.active { background: #6366f1; color: #fff; border-color: #6366f1; }
+.heatmap-legend {
+    display: flex; align-items: center; gap: 8px; font-size: 12px; color: #94a3b8;
+}
+.legend-bar {
+    width: 120px; height: 12px; border-radius: 3px;
+    background: linear-gradient(to right, #27ae60, #a9dfbf, #bdc3c7, #f5b7b1, #c0392b);
+}
+#treemap-chart { width: 100%; height: 520px; }
+.stock-detail-panel {
+    margin-top: 16px; display: none;
+    animation: slideDown 0.3s ease;
+}
+@keyframes slideDown {
+    from { opacity: 0; transform: translateY(-10px); }
+    to { opacity: 1; transform: translateY(0); }
+}
+.panel-header {
+    display: flex; align-items: center; justify-content: space-between;
+    padding: 12px 16px; background: #1e293b; border-radius: 8px 8px 0 0;
+    border-bottom: 1px solid #334155;
+}
+.panel-title { font-size: 16px; font-weight: 600; color: #e2e8f0; }
+.panel-summary { font-size: 13px; color: #94a3b8; }
+.panel-close {
+    background: none; border: none; color: #94a3b8; cursor: pointer;
+    font-size: 18px; padding: 0 4px;
+}
+.panel-close:hover { color: #e2e8f0; }
+.stock-table { width: 100%; }
+.stock-table thead th {
+    background: #0f172a; color: #94a3b8; font-size: 12px; font-weight: 500;
+    padding: 8px 12px; border-bottom: 1px solid #1e293b; position: sticky; top: 0;
+    text-transform: uppercase; letter-spacing: 0.5px;
+}
+.stock-table tbody tr { border-bottom: 1px solid #1e293b; }
+.stock-table tbody tr:hover { background: rgba(99, 102, 241, 0.08); }
+.stock-table td { padding: 8px 12px; color: #e2e8f0; font-size: 13px; }
+.stock-table td:first-child { font-weight: 500; }
+.stock-table td.pct-up { color: #ef4444; }
+.stock-table td.pct-down { color: #22c55e; }
+.stock-table td.pct-flat { color: #94a3b8; }
+.error-banner {
+    padding: 20px; text-align: center; color: #ef4444;
+    background: #1e293b; border-radius: 8px; margin: 20px 0;
+}
+</style>
+{% endblock %}
+
+{% block content %}
+<div class="container-fluid heatmap-page">
+    {% if error is defined and error %}
+    <div class="error-banner">{{ error }}</div>
+    {% else %}
+    <div class="heatmap-header">
+        <div>
+            <span class="heatmap-title">板块热力图</span>
+            <span class="heatmap-date" style="margin-left:12px;">{{ trade_date }}</span>
+        </div>
+        <div style="display:flex;align-items:center;gap:20px;">
+            <div class="sort-btns">
+                <button class="sort-btn active" data-sort="pct" onclick="switchSort('pct')">涨跌幅排序</button>
+                <button class="sort-btn" data-sort="mv" onclick="switchSort('mv')">市值排序</button>
+            </div>
+            <div class="heatmap-legend">
+                <span>跌</span>
+                <div class="legend-bar"></div>
+                <span>涨</span>
+            </div>
+        </div>
+    </div>
+    <div id="treemap-chart"></div>
+    <div id="stock-detail" class="stock-detail-panel"></div>
+    {% endif %}
+</div>
+{% endblock %}
+
+{% block extra_js %}
+<script>
+(function() {
+    var sectors = {{ sectors_json | tojson }};
+    var stocks = {{ stocks_json | tojson }};
+
+    // ---- Treemap ----
+    var chart = echarts.init(document.getElementById('treemap-chart'), 'financial');
+
+    function pctColor(val) {
+        if (val > 0) {
+            var t = Math.min(val / 6, 1);
+            return 'rgb(' + Math.round(192 + (245-192)*(1-t)) + ',' +
+                            Math.round(57 + (183-57)*(1-t)) + ',' +
+                            Math.round(43 + (177-43)*(1-t)) + ')';
+        } else if (val < 0) {
+            var t = Math.min(Math.abs(val) / 6, 1);
+            return 'rgb(' + Math.round(39 + (169-39)*(1-t)) + ',' +
+                            Math.round(174 + (223-174)*(1-t)) + ',' +
+                            Math.round(96 + (191-96)*(1-t)) + ')';
+        }
+        return '#bdc3c7';
+    }
+
+    var treemapData = sectors.map(function(s) {
+        return {
+            name: s.name,
+            value: s.total_mv,
+            total_mv: s.total_mv,
+            avg_pct_chg: s.avg_pct_chg,
+            stock_count: s.stock_count,
+            up_count: s.up_count,
+            down_count: s.down_count,
+            net_mf_amount: s.net_mf_amount,
+            itemStyle: { color: pctColor(s.avg_pct_chg) }
+        };
+    });
+
+    function applySort(data, mode) {
+        // 深拷贝避免修改原数组
+        var sorted = data.map(function(d) {
+            var item = {};
+            for (var k in d) item[k] = d[k];
+            // 涨跌幅排序：矩形面积 = 涨跌幅绝对值 × 系数，让波动大的板块更突出
+            if (mode === 'pct') {
+                item.value = Math.max(Math.abs(d.avg_pct_chg), 0.01);
+            } else {
+                item.value = d.total_mv;
+            }
+            return item;
+        });
+        sorted.sort(function(a, b) { return b.value - a.value; });
+        return sorted;
+    }
+
+    function buildOption(data) {
+        return {
+            tooltip: {
+                formatter: function(info) {
+                    var d = info.data;
+                    return '<b>' + d.name + '</b><br/>' +
+                        '加权涨跌: <b>' + d.avg_pct_chg.toFixed(2) + '%</b><br/>' +
+                        '涨/跌: ' + d.up_count + '/' + d.down_count +
+                        ' (共' + d.stock_count + '只)<br/>' +
+                        '净流入: ' + (d.net_mf_amount / 10000).toFixed(2) + '亿';
+                }
+            },
+            series: [{
+                type: 'treemap',
+                data: data,
+                roam: false,
+                nodeClick: false,
+                breadcrumb: { show: false },
+                levels: [{
+                    itemStyle: { borderColor: '#0f172a', borderWidth: 2, gapWidth: 2 },
+                    upperLabel: { show: false }
+                }],
+                label: {
+                    show: true,
+                    formatter: function(params) {
+                        var d = params.data;
+                        var sign = d.avg_pct_chg >= 0 ? '+' : '';
+                        return d.name + '\n' + sign + d.avg_pct_chg.toFixed(2) + '%';
+                    },
+                    fontSize: 13,
+                    color: '#fff',
+                    textShadowColor: 'rgba(0,0,0,0.5)',
+                    textShadowBlur: 4
+                }
+            }]
+        };
+    }
+
+    var currentSort = 'pct';
+    chart.setOption(buildOption(applySort(treemapData, 'pct')));
+
+    // ---- 排序切换 ----
+    window.switchSort = function(mode) {
+        if (mode === currentSort) return;
+        currentSort = mode;
+
+        document.querySelectorAll('.sort-btn').forEach(function(btn) {
+            btn.classList.toggle('active', btn.dataset.sort === mode);
+        });
+
+        chart.setOption(buildOption(applySort(treemapData, mode)));
+    };
+
+    // ---- Sector Click → Stock Table ----
+    var activeIndustry = null;
+
+    window.closeSectorPanel = function() {
+        document.getElementById('stock-detail').style.display = 'none';
+        activeIndustry = null;
+    };
+
+    chart.on('click', function(params) {
+        var industry = params.data.name;
+        var panel = document.getElementById('stock-detail');
+
+        if (activeIndustry === industry) {
+            panel.style.display = 'none';
+            activeIndustry = null;
+            return;
+        }
+        activeIndustry = industry;
+
+        var sector = sectors.find(function(s) { return s.name === industry; });
+        var filtered = stocks.filter(function(s) { return s.industry === industry; });
+
+        var sign = sector.avg_pct_chg >= 0 ? '+' : '';
+        var html = '<div class="panel-header">' +
+            '<div><span class="panel-title">' + industry + '（' + filtered.length + '只）</span>' +
+            '<span class="panel-summary" style="margin-left:16px;">加权涨跌: ' +
+            '<span style="color:' + (sector.avg_pct_chg >= 0 ? '#ef4444' : '#22c55e') + '">' +
+            sign + sector.avg_pct_chg.toFixed(2) + '%</span></span></div>' +
+            '<button class="panel-close" onclick="closeSectorPanel()">&times;</button>' +
+            '</div>' +
+            '<table class="stock-table"><thead><tr>' +
+            '<th>名称</th><th>涨跌幅</th><th>收盘价</th><th>总市值(亿)</th>' +
+            '<th>净流入(万)</th><th>换手率</th>' +
+            '</tr></thead><tbody>';
+
+        filtered.forEach(function(s) {
+            var cls = s.pct_chg > 0 ? 'pct-up' : (s.pct_chg < 0 ? 'pct-down' : '');
+            var pSign = s.pct_chg > 0 ? '+' : '';
+            html += '<tr>' +
+                '<td>' + s.name + '</td>' +
+                '<td class="' + cls + '">' + pSign + (s.pct_chg != null ? s.pct_chg.toFixed(2) : '--') + '%</td>' +
+                '<td>' + (s.close != null ? s.close.toFixed(2) : '--') + '</td>' +
+                '<td>' + (s.total_mv != null ? (s.total_mv / 10000).toFixed(1) : '--') + '</td>' +
+                '<td class="' + (s.net_mf_amount > 0 ? 'pct-up' : (s.net_mf_amount < 0 ? 'pct-down' : '')) + '">' +
+                (s.net_mf_amount != null ? s.net_mf_amount.toFixed(0) : '--') + '</td>' +
+                '<td>' + (s.turnover_rate != null ? s.turnover_rate.toFixed(2) + '%' : '--') + '</td>' +
+                '</tr>';
+        });
+
+        html += '</tbody></table>';
+        panel.innerHTML = html;
+        panel.style.display = 'block';
+        window._activeIndustry = industry;
+    });
+
+    window.addEventListener('resize', function() { chart.resize(); });
+})();
+</script>
+{% endblock %}
diff --git a/app/templates/pattern_screen.html b/app/templates/pattern_screen.html
new file mode 100644
index 000000000..b2052e85e
--- /dev/null
+++ b/app/templates/pattern_screen.html
@@ -0,0 +1,374 @@
+{% extends "base.html" %}
+{% block title %}形态选股{% endblock %}
+{% block extra_css %}
+<style>
+.pattern-page { display: flex; min-height: calc(100vh - 56px); }
+
+/* ── Left Panel — uses Obsidian design tokens ── */
+.filter-panel {
+    width: 300px; min-width: 300px;
+    background: var(--obs-surface-alt);
+    border-right: 1px solid var(--obs-border-light);
+    overflow-y: auto; padding: 16px;
+    display: flex; flex-direction: column;
+}
+.filter-panel .search-box { margin-bottom: 12px; }
+.filter-panel .search-box input {
+    width: 100%; padding: 6px 10px; border-radius: var(--border-radius-sm);
+    border: 1px solid var(--obs-border);
+    background: var(--obs-bg); color: var(--obs-text);
+    font-size: 13px;
+}
+.filter-panel .search-box input:focus {
+    outline: none; border-color: var(--obs-primary);
+    box-shadow: 0 0 0 2px var(--obs-primary-bg);
+}
+.filter-panel .search-box input::placeholder { color: var(--obs-text-muted); }
+
+/* Accordion — override Bootstrap defaults completely */
+.accordion-item {
+    border: none !important; margin-bottom: 2px;
+    background: transparent !important;
+}
+.accordion-button {
+    background: transparent !important; color: var(--obs-text);
+    font-size: 13px; font-weight: 600;
+    padding: 8px 12px; border: none !important;
+    box-shadow: none !important;
+}
+.accordion-button:not(.collapsed) {
+    background: var(--obs-primary-bg) !important; color: var(--obs-primary-light);
+}
+.accordion-button::after { filter: invert(0.7); }
+.accordion-collapse { border: none !important; }
+.accordion-body {
+    padding: 4px 8px;
+    background: transparent !important;
+    border: none !important;
+}
+
+/* Checkbox row — card-like feel, matches card hover pattern */
+.pattern-check {
+    display: flex; align-items: center; padding: 4px 6px;
+    border-radius: var(--border-radius-sm);
+    font-size: 12px; color: var(--obs-text-secondary); cursor: pointer;
+    background: var(--obs-surface);
+    border: 1px solid transparent;
+    margin-bottom: 2px;
+    transition: var(--transition-fast);
+}
+.pattern-check:hover {
+    border-color: var(--obs-border-hover);
+    box-shadow: 0 0 12px rgba(99,102,241,0.06);
+    color: var(--obs-text);
+}
+.pattern-check:has(input:checked) {
+    background: var(--obs-primary-bg);
+    border-color: var(--obs-border-hover);
+    color: var(--obs-primary-light);
+}
+.pattern-check input[type="checkbox"] {
+    margin-right: 6px; accent-color: var(--obs-primary);
+    width: 14px; height: 14px;
+}
+.pattern-check .hit-count {
+    margin-left: auto; color: var(--obs-text-muted); font-size: 11px;
+}
+
+/* Reset button */
+.filter-actions { margin-top: auto; padding-top: 12px; display: flex; }
+.filter-actions .btn {
+    flex: 1; font-size: 13px;
+    border-color: var(--obs-border) !important;
+    color: var(--obs-text-secondary) !important;
+    background: transparent !important;
+}
+.filter-actions .btn:hover {
+    border-color: var(--obs-border-hover) !important;
+    color: var(--obs-primary-light) !important;
+    background: var(--obs-primary-bg) !important;
+}
+
+/* ── Right Area ── */
+.result-area { flex: 1; padding: 20px; overflow-x: auto; }
+.stats-bar {
+    display: flex; align-items: center; gap: 16px;
+    margin-bottom: 16px; color: var(--obs-text-secondary); font-size: 14px;
+}
+.stats-bar .badge { background: var(--obs-primary-bg); color: var(--obs-primary-light); }
+.result-table { width: 100%; font-size: 13px; color: var(--obs-text); }
+.result-table thead th {
+    background: var(--obs-surface); color: var(--obs-text-secondary);
+    font-weight: 600; padding: 10px 12px;
+    border-bottom: 1px solid var(--obs-border);
+    cursor: pointer; user-select: none; white-space: nowrap;
+}
+.result-table thead th:hover { color: var(--obs-primary-light); }
+.result-table thead th .sort-icon { margin-left: 4px; font-size: 10px; }
+.result-table tbody td {
+    padding: 8px 12px; border-bottom: 1px solid rgba(51,65,85,0.4);
+    white-space: nowrap;
+}
+.result-table tbody tr:hover { background: var(--obs-primary-bg); }
+.pct-up { color: var(--obs-danger); }
+.pct-down { color: var(--obs-success); }
+.pct-flat { color: var(--obs-text-secondary); }
+.pagination-bar {
+    display: flex; align-items: center; justify-content: center;
+    gap: 6px; margin-top: 16px;
+}
+.page-btn {
+    padding: 4px 10px; border-radius: var(--border-radius-sm);
+    border: 1px solid var(--obs-border);
+    background: transparent; color: var(--obs-text-secondary);
+    font-size: 12px; cursor: pointer;
+}
+.page-btn:hover { border-color: var(--obs-primary); color: var(--obs-text); }
+.page-btn.active { background: var(--obs-primary); color: #fff; border-color: var(--obs-primary); }
+.page-btn:disabled { opacity: 0.4; cursor: default; }
+.empty-state { text-align: center; padding: 60px 20px; color: var(--obs-text-muted); }
+.empty-state i { font-size: 48px; margin-bottom: 16px; }
+@media (max-width: 768px) {
+    .pattern-page { flex-direction: column; }
+    .filter-panel { width: 100%; min-width: auto; max-height: 40vh; }
+}
+</style>
+{% endblock %}
+
+{% block content %}
+<div class="pattern-page">
+  <div class="filter-panel">
+    <div class="search-box">
+      <input type="text" id="searchInput" placeholder="搜索形态名称..." oninput="filterBySearch()">
+    </div>
+    <div class="accordion" id="groupAccordion"></div>
+    <div class="filter-actions">
+      <button class="btn btn-outline-secondary" onclick="resetFilters()">
+        <i class="fas fa-undo me-1"></i>重置
+      </button>
+    </div>
+  </div>
+  <div class="result-area">
+    <div class="stats-bar" id="statsBar">
+      <span>交易日: <strong id="tradeDate">--</strong></span>
+      <span>|</span>
+      <span>已选: <span class="badge" id="selectedCount">0</span> 个形态</span>
+      <span>|</span>
+      <span>匹配: <strong id="totalCount">0</strong> 只</span>
+    </div>
+    <div id="tableContainer">
+      <div class="empty-state">
+        <i class="fas fa-crosshairs d-block"></i>
+        <div>加载中...</div>
+      </div>
+    </div>
+    <div class="pagination-bar" id="paginationBar"></div>
+  </div>
+</div>
+
+<script>
+let currentSort = 'pct_chg';
+let currentOrder = 'desc';
+let currentOffset = 0;
+let currentLimit = 50;
+let currentTotal = 0;
+let allGroups = [];
+
+document.addEventListener('DOMContentLoaded', () => {
+    loadGroups();
+    doScreen();
+});
+
+async function loadGroups() {
+    try {
+        const resp = await axios.get('/api/pattern-screen/groups');
+        const data = resp.data.data;
+        if (!data) return;
+        allGroups = data;
+        renderGroups(data);
+    } catch (e) {
+        console.error('加载分组失败:', e);
+    }
+}
+
+function renderGroups(groups) {
+    const container = document.getElementById('groupAccordion');
+    container.innerHTML = '';
+    groups.forEach((g, idx) => {
+        const show = idx < 2;
+        container.innerHTML += `
+        <div class="accordion-item">
+          <h2 class="accordion-header">
+            <button class="accordion-button ${show ? '' : 'collapsed'}" type="button"
+                    data-bs-toggle="collapse" data-bs-target="#group-${g.id}">
+              ${g.label} <span style="margin-left:auto;font-weight:400;font-size:11px;color:#64748b;">${g.fields.length}</span>
+            </button>
+          </h2>
+          <div id="group-${g.id}" class="accordion-collapse collapse ${show ? 'show' : ''}"
+               data-bs-parent="#groupAccordion">
+            <div class="accordion-body">
+              ${g.fields.map(f => `
+                <label class="pattern-check" data-label="${f.label}">
+                  <input type="checkbox" value="${f.key}" onchange="onPatternChange()">
+                  ${f.label}
+                  <span class="hit-count">${f.count}</span>
+                </label>
+              `).join('')}
+            </div>
+          </div>
+        </div>`;
+    });
+}
+
+function filterBySearch() {
+    const q = document.getElementById('searchInput').value.trim().toLowerCase();
+    document.querySelectorAll('.pattern-check').forEach(el => {
+        const label = el.getAttribute('data-label').toLowerCase();
+        el.style.display = (!q || label.includes(q)) ? '' : 'none';
+    });
+}
+
+function onPatternChange() {
+    const count = document.querySelectorAll('.pattern-check input:checked').length;
+    document.getElementById('selectedCount').textContent = count;
+    currentOffset = 0;
+    doScreen();
+}
+
+function getCheckedPatterns() {
+    return Array.from(document.querySelectorAll('.pattern-check input:checked'))
+        .map(cb => cb.value);
+}
+
+async function doScreen() {
+    const patterns = getCheckedPatterns();
+    try {
+        const resp = await axios.post('/api/pattern-screen/screen', {
+            patterns,
+            sort_by: currentSort,
+            order: currentOrder,
+            limit: currentLimit,
+            offset: currentOffset,
+        });
+        const result = resp.data.data;
+        if (!result) return;
+        currentTotal = result.total;
+        document.getElementById('tradeDate').textContent = formatDate(result.trade_date);
+        document.getElementById('totalCount').textContent = result.total;
+        renderTable(result.rows);
+        renderPagination(result.total, result.limit, result.offset);
+    } catch (e) {
+        console.error('筛选失败:', e);
+        document.getElementById('tableContainer').innerHTML =
+            '<div class="empty-state"><i class="fas fa-exclamation-triangle d-block"></i><div>筛选失败，请重试</div></div>';
+    }
+}
+
+function renderTable(rows) {
+    if (!rows || rows.length === 0) {
+        document.getElementById('tableContainer').innerHTML =
+            '<div class="empty-state"><i class="fas fa-inbox d-block"></i><div>暂无匹配结果</div></div>';
+        return;
+    }
+    const sortIcon = (col) => {
+        if (col !== currentSort) return '<i class="fas fa-sort sort-icon" style="opacity:0.3"></i>';
+        return currentOrder === 'desc'
+            ? '<i class="fas fa-sort-down sort-icon"></i>'
+            : '<i class="fas fa-sort-up sort-icon"></i>';
+    };
+    let html = `<table class="result-table">
+      <thead><tr>
+        <th onclick="sortTable('ts_code')">代码 ${sortIcon('ts_code')}</th>
+        <th>名称</th>
+        <th>行业</th>
+        <th onclick="sortTable('pct_chg')">涨跌幅 ${sortIcon('pct_chg')}</th>
+        <th onclick="sortTable('close')">现价 ${sortIcon('close')}</th>
+        <th onclick="sortTable('amount')">成交额 ${sortIcon('amount')}</th>
+        <th onclick="sortTable('total_mv')">总市值 ${sortIcon('total_mv')}</th>
+        <th onclick="sortTable('turnover_rate')">换手率 ${sortIcon('turnover_rate')}</th>
+        <th onclick="sortTable('vol_ratio_5')">量比 ${sortIcon('vol_ratio_5')}</th>
+      </tr></thead><tbody>`;
+    rows.forEach(r => {
+        const pctClass = r.pct_chg > 0 ? 'pct-up' : (r.pct_chg < 0 ? 'pct-down' : 'pct-flat');
+        const pctStr = r.pct_chg > 0 ? '+' + r.pct_chg.toFixed(2) : r.pct_chg.toFixed(2);
+        html += `<tr>
+          <td>${r.ts_code || ''}</td>
+          <td>${r.name || ''}</td>
+          <td>${r.industry || '--'}</td>
+          <td class="${pctClass}">${pctStr}%</td>
+          <td>${fmtNum(r.close, 2)}</td>
+          <td>${fmtAmount(r.amount)}</td>
+          <td>${fmtAmount(r.total_mv)}</td>
+          <td>${fmtNum(r.turnover_rate, 2)}%</td>
+          <td>${fmtNum(r.vol_ratio_5, 2)}</td>
+        </tr>`;
+    });
+    html += '</tbody></table>';
+    document.getElementById('tableContainer').innerHTML = html;
+}
+
+function sortTable(field) {
+    if (currentSort === field) {
+        currentOrder = currentOrder === 'desc' ? 'asc' : 'desc';
+    } else {
+        currentSort = field;
+        currentOrder = 'desc';
+    }
+    currentOffset = 0;
+    doScreen();
+}
+
+function renderPagination(total, limit, offset) {
+    const bar = document.getElementById('paginationBar');
+    if (total <= limit) { bar.innerHTML = ''; return; }
+    const pages = Math.ceil(total / limit);
+    const cur = Math.floor(offset / limit) + 1;
+    let html = `<button class="page-btn" onclick="gotoPage(${offset - limit})" ${cur === 1 ? 'disabled' : ''}>
+        <i class="fas fa-chevron-left"></i></button>`;
+    let start = Math.max(1, cur - 3);
+    let end = Math.min(pages, start + 6);
+    start = Math.max(1, end - 6);
+    for (let i = start; i <= end; i++) {
+        const off = (i - 1) * limit;
+        html += `<button class="page-btn ${i === cur ? 'active' : ''}" onclick="gotoPage(${off})">${i}</button>`;
+    }
+    html += `<button class="page-btn" onclick="gotoPage(${offset + limit})" ${cur === pages ? 'disabled' : ''}>
+        <i class="fas fa-chevron-right"></i></button>`;
+    bar.innerHTML = html;
+}
+
+function gotoPage(offset) {
+    currentOffset = Math.max(0, offset);
+    doScreen();
+}
+
+function resetFilters() {
+    document.querySelectorAll('.pattern-check input').forEach(cb => cb.checked = false);
+    document.getElementById('searchInput').value = '';
+    filterBySearch();
+    onPatternChange();
+    currentOffset = 0;
+    currentSort = 'pct_chg';
+    currentOrder = 'desc';
+    doScreen();
+}
+
+function fmtNum(v, digits) {
+    if (v == null) return '--';
+    return Number(v).toFixed(digits);
+}
+
+function fmtAmount(v) {
+    if (v == null) return '--';
+    v = Number(v);
+    if (Math.abs(v) >= 1e8) return (v / 1e8).toFixed(2) + '亿';
+    if (Math.abs(v) >= 1e4) return (v / 1e4).toFixed(2) + '万';
+    return v.toFixed(2);
+}
+
+function formatDate(d) {
+    if (!d || d.length !== 8) return d || '--';
+    return d.slice(0, 4) + '-' + d.slice(4, 6) + '-' + d.slice(6, 8);
+}
+</script>
+{% endblock %}
diff --git a/docs/superpowers/plans/2026-06-07-pattern-screen.md b/docs/superpowers/plans/2026-06-07-pattern-screen.md
new file mode 100644
index 000000000..96b91fb93
--- /dev/null
+++ b/docs/superpowers/plans/2026-06-07-pattern-screen.md
@@ -0,0 +1,1162 @@
+# 形态选股 Implementation Plan
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+
+**Goal:** Add a pattern-based stock screening page that reads `data/data.parquet` and provides a grouped checkbox filter panel + results table.
+
+**Architecture:** Independent service reads the pre-computed wide parquet into a cached DataFrame. Pure-AND filtering across 132 binary pattern fields. Flask blueprint exposes two API endpoints and one page route. Frontend is a single Jinja template with vanilla JS.
+
+**Tech Stack:** Flask Blueprint, pandas, Bootstrap 5, vanilla JS, ECharts theme from base template.
+
+---
+
+## File Structure
+
+| Action | File | Responsibility |
+|--------|------|----------------|
+| Create | `app/services/pattern_screen_service.py` | Data loading, group metadata, AND-filtering, sort, paginate |
+| Create | `app/api/pattern_screen_api.py` | REST endpoints: `/api/pattern-screen/groups` and `/api/pattern-screen/screen` |
+| Create | `app/routes/pattern_screen.py` | Page route: `/pattern-screen/` |
+| Create | `app/templates/pattern_screen.html` | Left panel (grouped checkboxes) + right table |
+| Create | `tests/services/test_pattern_screen_service.py` | Service unit tests |
+| Create | `tests/api/test_pattern_screen_api.py` | API contract tests |
+| Modify | `app/__init__.py:40,54` | Register new blueprints |
+
+---
+
+### Task 1: PatternScreenService — metadata and screening logic
+
+**Files:**
+- Create: `app/services/pattern_screen_service.py`
+- Test: `tests/services/test_pattern_screen_service.py`
+
+- [ ] **Step 1: Write the failing test**
+
+```python
+# tests/services/test_pattern_screen_service.py
+"""PatternScreenService unit tests."""
+import pytest
+from unittest.mock import patch, MagicMock
+import pandas as pd
+import numpy as np
+
+
+@pytest.fixture
+def sample_df():
+    """Minimal DataFrame mimicking data/data.parquet structure."""
+    return pd.DataFrame({
+        'ts_code': ['000001.SZ', '000002.SZ', '000003.SZ', '000004.SZ'],
+        'name': ['平安银行', '万科A', '国农科技', '国信证券'],
+        'industry': ['银行', '房地产', '综合', '证券'],
+        'trade_date': ['20260605'] * 4,
+        'pct_chg': [1.5, -0.5, 3.0, 0.0],
+        'close': [12.5, 8.0, 25.0, 15.0],
+        'amount': [1000000, 500000, 200000, 300000],
+        'total_mv': [2400000, 1200000, 600000, 900000],
+        'turnover_rate': [1.2, 0.8, 2.5, 0.5],
+        'vol_ratio_5': [1.8, 0.6, 2.1, 0.9],
+        'consec_up_days': [2, 0, 3, 1],
+        'pattern_golden_cross': [1, 0, 1, 0],
+        'pattern_ma_bull': [0, 0, 1, 1],
+        'pattern_bull_candle': [1, 0, 1, 0],
+        'pattern_bear_candle': [0, 1, 0, 1],
+        'break_high_20': [1, 0, 1, 0],
+    })
+
+
+@pytest.fixture
+def service(sample_df):
+    """Service with injected test DataFrame."""
+    with patch('app.services.pattern_screen_service.PatternScreenService._load_df', return_value=sample_df):
+        from app.services.pattern_screen_service import PatternScreenService
+        svc = PatternScreenService()
+        svc._df = sample_df
+        return svc
+
+
+class TestGetGroups:
+    def test_returns_groups_with_hits(self, service):
+        groups = service.get_groups()
+        assert isinstance(groups, list)
+        assert len(groups) > 0
+        # Each group has id, label, fields
+        g = groups[0]
+        assert 'id' in g
+        assert 'label' in g
+        assert 'fields' in g
+        for f in g['fields']:
+            assert 'key' in f
+            assert 'label' in f
+            assert 'count' in f
+            assert isinstance(f['count'], int)
+
+    def test_fields_not_in_dataframe_are_excluded(self, service):
+        groups = service.get_groups()
+        df_cols = set(service._df.columns)
+        for g in groups:
+            for f in g['fields']:
+                assert f['key'] in df_cols
+
+
+class TestScreen:
+    def test_no_patterns_returns_all(self, service, sample_df):
+        result = service.screen(patterns=[])
+        assert result['total'] == len(sample_df)
+        assert len(result['rows']) == len(sample_df)
+
+    def test_single_pattern_filters(self, service):
+        result = service.screen(patterns=['pattern_golden_cross'])
+        assert result['total'] == 2
+        codes = [r['ts_code'] for r in result['rows']]
+        assert '000001.SZ' in codes
+        assert '000003.SZ' in codes
+
+    def test_multiple_patterns_and_logic(self, service):
+        result = service.screen(patterns=['pattern_golden_cross', 'pattern_ma_bull'])
+        # Only 000003.SZ has both
+        assert result['total'] == 1
+        assert result['rows'][0]['ts_code'] == '000003.SZ'
+
+    def test_sort_desc(self, service):
+        result = service.screen(patterns=[], sort_by='pct_chg', order='desc')
+        pcts = [r['pct_chg'] for r in result['rows']]
+        assert pcts == sorted(pcts, reverse=True)
+
+    def test_sort_asc(self, service):
+        result = service.screen(patterns=[], sort_by='pct_chg', order='asc')
+        pcts = [r['pct_chg'] for r in result['rows']]
+        assert pcts == sorted(pcts)
+
+    def test_pagination(self, service):
+        result = service.screen(patterns=[], limit=2, offset=0)
+        assert len(result['rows']) == 2
+        assert result['limit'] == 2
+        assert result['offset'] == 0
+        assert result['total'] == 4
+
+    def test_offset_beyond_results(self, service):
+        result = service.screen(patterns=[], offset=100)
+        assert result['total'] == 4
+        assert len(result['rows']) == 0
+
+    def test_invalid_sort_by_raises(self, service):
+        with pytest.raises(ValueError, match='sort_by'):
+            service.screen(patterns=[], sort_by='invalid_column')
+
+    def test_invalid_order_raises(self, service):
+        with pytest.raises(ValueError, match='order'):
+            service.screen(patterns=[], order='random')
+
+    def test_invalid_pattern_key_raises(self, service):
+        with pytest.raises(ValueError, match='pattern.*not found'):
+            service.screen(patterns=['nonexistent_pattern'])
+
+    def test_limit_capped_at_500(self, service):
+        result = service.screen(patterns=[], limit=9999)
+        assert result['limit'] == 500
+
+    def test_includes_trade_date(self, service):
+        result = service.screen(patterns=[])
+        assert result['trade_date'] == '20260605'
+
+    def test_nan_converted_to_none(self, sample_df):
+        # Add NaN to test conversion
+        sample_df.loc[0, 'industry'] = np.nan
+        with patch('app.services.pattern_screen_service.PatternScreenService._load_df', return_value=sample_df):
+            from app.services.pattern_screen_service import PatternScreenService
+            svc = PatternScreenService()
+            svc._df = sample_df
+            result = svc.screen(patterns=[])
+            # industry for first row should be None, not NaN
+            assert result['rows'][0]['industry'] is None
+
+
+class TestInvalidateCache:
+    def test_clears_df(self, service):
+        assert service._df is not None
+        service.invalidate_cache()
+        assert service._df is None
+```
+
+- [ ] **Step 2: Run test to verify it fails**
+
+Run: `pytest tests/services/test_pattern_screen_service.py -v`
+Expected: FAIL — module `pattern_screen_service` not found
+
+- [ ] **Step 3: Write the service implementation**
+
+```python
+# app/services/pattern_screen_service.py
+"""形态选股服务 — 读取 data/data.parquet，提供分组元数据和纯AND筛选。"""
+import os
+import pandas as pd
+import numpy as np
+from flask import current_app
+from loguru import logger
+
+# ── 形态分组定义 ──────────────────────────────────────────
+# 从 stock_screener/backend/meta.py 前7组提取
+PATTERN_GROUPS = [
+    {
+        "id": "single_candle",
+        "label": "单K形态",
+        "fields": [
+            {"key": "pattern_bull_candle", "label": "阳线"},
+            {"key": "pattern_bear_candle", "label": "阴线"},
+            {"key": "pattern_hammer", "label": "锤头线"},
+            {"key": "pattern_doji", "label": "十字星"},
+            {"key": "pattern_spinning_top", "label": "纺锤线"},
+            {"key": "pattern_shooting_star", "label": "流星线"},
+            {"key": "pattern_long_upper_shadow", "label": "长上影线"},
+            {"key": "pattern_long_lower_shadow", "label": "长下影线"},
+            {"key": "pattern_gravestone_doji", "label": "墓碑十字"},
+            {"key": "pattern_dragonfly_doji", "label": "蜻蜓十字"},
+            {"key": "pattern_hanging_man", "label": "上吊线"},
+            {"key": "pattern_inverted_hammer", "label": "倒锤头"},
+            {"key": "pattern_big_bull", "label": "大阳线"},
+            {"key": "pattern_big_bear", "label": "大阴线"},
+            {"key": "pattern_medium_bull", "label": "中阳线"},
+            {"key": "pattern_medium_bear", "label": "中阴线"},
+            {"key": "pattern_small_bull", "label": "小阳线"},
+            {"key": "pattern_small_bear", "label": "小阴线"},
+            {"key": "pattern_no_body", "label": "无实体线"},
+            {"key": "pattern_no_upper_bull", "label": "光头阳线"},
+            {"key": "pattern_no_upper_bear", "label": "光头阴线"},
+            {"key": "pattern_no_lower_bull", "label": "光脚阳线"},
+            {"key": "pattern_no_lower_bear", "label": "光脚阴线"},
+            {"key": "pattern_t_shape", "label": "T字线"},
+            {"key": "pattern_inverted_t_shape", "label": "倒T字线"},
+            {"key": "pattern_low_open_high_close", "label": "低开高走"},
+            {"key": "pattern_high_open_low_close", "label": "高开低走"},
+            {"key": "pattern_gap_up", "label": "跳空高开"},
+            {"key": "pattern_gap_down", "label": "跳空低开"},
+            {"key": "pattern_close_above_prev_close", "label": "收盘站上前收"},
+            {"key": "pattern_close_below_prev_close", "label": "收盘跌破前收"},
+            {"key": "pattern_gap_reclaim_prev_close", "label": "低开收回前收"},
+            {"key": "pattern_gap_fade_below_prev_close", "label": "高开回落失守前收"},
+            {"key": "pattern_close_high", "label": "收盘近最高"},
+            {"key": "pattern_flat_open_high_close", "label": "平开高走"},
+            {"key": "pattern_flat_open_low_close", "label": "平开低走"},
+            {"key": "pattern_gap_up_close_bull", "label": "高开收阳"},
+            {"key": "pattern_gap_down_close_bear", "label": "低开收阴"},
+            {"key": "pattern_open_near_high_close_high", "label": "开盘即最高附近收盘"},
+            {"key": "pattern_open_near_low_close_low", "label": "开盘即最低附近收盘"},
+            {"key": "pattern_flat_open", "label": "平开"},
+            {"key": "pattern_gap_up_fill", "label": "高开补缺"},
+            {"key": "pattern_gap_down_fill", "label": "低开补缺"},
+            {"key": "pattern_pin_bar", "label": "Pin Bar"},
+            {"key": "pattern_reversal_prelude", "label": "反包前兆"},
+            {"key": "pattern_high_resistance", "label": "高位受阻"},
+            {"key": "pattern_low_stabilization", "label": "低位止跌"},
+            {"key": "pattern_break_prev_high", "label": "向上突破前高"},
+            {"key": "pattern_break_prev_low", "label": "向下跌破前低"},
+            {"key": "pattern_false_break", "label": "假突破回落"},
+            {"key": "pattern_false_breakdown_recovery", "label": "假跌破回升"},
+        ],
+    },
+    {
+        "id": "double_candle",
+        "label": "双K形态",
+        "fields": [
+            {"key": "pattern_bullish_engulfing", "label": "阳包阴"},
+            {"key": "pattern_bearish_engulfing", "label": "阴包阳"},
+            {"key": "pattern_inside_bar", "label": "孕线"},
+            {"key": "pattern_dark_cloud", "label": "乌云盖顶"},
+            {"key": "pattern_piercing", "label": "刺透形态"},
+            {"key": "pattern_tweezer_top", "label": "镊子顶"},
+            {"key": "pattern_tweezer_bottom", "label": "镊子底"},
+            {"key": "pattern_gap_break", "label": "跳空上攻"},
+            {"key": "pattern_gap_down_break", "label": "跳空下跌"},
+            {"key": "pattern_gap_up_no_fill", "label": "跳空不补上行"},
+            {"key": "pattern_gap_down_no_fill", "label": "跳空不补下行"},
+            {"key": "pattern_reversal_bar", "label": "反转包线"},
+            {"key": "pattern_flat_top", "label": "平头顶"},
+            {"key": "pattern_flat_bottom", "label": "平头底"},
+            {"key": "pattern_island_reversal", "label": "岛形反转"},
+            {"key": "pattern_t_limit", "label": "T字板"},
+            {"key": "pattern_limit_reversal_wrap", "label": "涨停反包"},
+            {"key": "pattern_vol_up", "label": "放量上涨"},
+            {"key": "pattern_vol_down", "label": "缩量下跌"},
+            {"key": "pattern_double_volume_bar", "label": "倍量柱"},
+            {"key": "pattern_breakout_volume_confirm", "label": "突破放量确认"},
+        ],
+    },
+    {
+        "id": "triple_candle",
+        "label": "三K形态",
+        "fields": [
+            {"key": "pattern_morning_star", "label": "早晨之星"},
+            {"key": "pattern_evening_star", "label": "黄昏之星"},
+            {"key": "pattern_morning_doji_star", "label": "启明星"},
+            {"key": "pattern_evening_doji_star", "label": "黄昏十字星"},
+            {"key": "pattern_three_black_crows", "label": "三只乌鸦"},
+            {"key": "pattern_red_three", "label": "红三兵"},
+            {"key": "pattern_three_outside_up", "label": "三外升"},
+            {"key": "pattern_three_outside_down", "label": "三外降"},
+            {"key": "pattern_rising_three_methods", "label": "上升三法"},
+            {"key": "pattern_falling_three_methods", "label": "下降三法"},
+            {"key": "pattern_three_up", "label": "三连阳"},
+            {"key": "pattern_three_down", "label": "三连阴"},
+            {"key": "pattern_three_yang_kaitai", "label": "三阳开泰"},
+            {"key": "pattern_three_yin_breakdown", "label": "三阴破位"},
+        ],
+    },
+    {
+        "id": "trend_structure",
+        "label": "趋势结构",
+        "fields": [
+            {"key": "pattern_up_trend", "label": "上升趋势"},
+            {"key": "pattern_down_trend", "label": "下降趋势"},
+            {"key": "pattern_sideways", "label": "横盘整理"},
+            {"key": "pattern_golden_cross", "label": "均线金叉"},
+            {"key": "pattern_duck_head", "label": "老鸭头"},
+            {"key": "pattern_double_bottom", "label": "双底"},
+            {"key": "pattern_arc_bottom", "label": "圆弧底"},
+            {"key": "pattern_ma_bull", "label": "均线多头排列"},
+            {"key": "pattern_high_tight", "label": "高位强势整理"},
+            {"key": "pattern_pullback_hold", "label": "回踩不破"},
+            {"key": "pattern_trend_continue", "label": "趋势中继"},
+            {"key": "pattern_ma_spread_bull", "label": "均线发散多头"},
+        ],
+    },
+    {
+        "id": "volume_price",
+        "label": "量价形态",
+        "fields": [
+            {"key": "pattern_box_breakout", "label": "箱体放量突破"},
+            {"key": "pattern_vol_price_up", "label": "量价齐升"},
+            {"key": "pattern_platform_break", "label": "平台突破"},
+            {"key": "pattern_triangle_squeeze", "label": "三角收敛突破"},
+            {"key": "pattern_limit_turnover_strong", "label": "涨停换手强"},
+            {"key": "pattern_price_volume_bear_divergence", "label": "价量顶背离"},
+            {"key": "pattern_price_volume_bull_divergence", "label": "价量底背离"},
+            {"key": "pattern_price_down_volume_up", "label": "价跌量增"},
+            {"key": "pattern_volume_staircase", "label": "量能阶梯"},
+            {"key": "pattern_pullback_volume_shrink", "label": "回调缩量"},
+            {"key": "pattern_high_turnover", "label": "高换手"},
+            {"key": "pattern_limit_up_volume_shrink", "label": "一字板缩量"},
+            {"key": "pattern_false_breakout_volume_weak", "label": "假突破量弱"},
+            {"key": "pattern_floor_volume_price", "label": "地量价稳"},
+            {"key": "pattern_blowoff_volume_price", "label": "天量滞涨"},
+            {"key": "pattern_v_reversal", "label": "V型反转"},
+        ],
+    },
+    {
+        "id": "compound",
+        "label": "复合形态",
+        "fields": [
+            {"key": "pattern_first_limit", "label": "首板"},
+            {"key": "pattern_multi_limit", "label": "连板"},
+            {"key": "pattern_one_word_limit", "label": "一字板"},
+            {"key": "pattern_limit_down_to_up", "label": "地天板"},
+            {"key": "pattern_lotus_breakout", "label": "莲花突破"},
+            {"key": "pattern_midway_refuel", "label": "中继加油"},
+            {"key": "pattern_consolidation_platform", "label": "整理平台"},
+            {"key": "pattern_n_breakout", "label": "N字突破"},
+            {"key": "pattern_gap_breakaway", "label": "跳空突破"},
+            {"key": "pattern_channel_breakout", "label": "通道突破"},
+            {"key": "pattern_flag_breakout", "label": "旗形突破"},
+        ],
+    },
+    {
+        "id": "momentum",
+        "label": "动量因子",
+        "fields": [
+            {"key": "break_high_20", "label": "突破20日新高"},
+            {"key": "break_high_60", "label": "突破60日新高"},
+            {"key": "break_high_120", "label": "突破120日新高"},
+            {"key": "break_high_250", "label": "突破250日新高"},
+            {"key": "consec_up_3", "label": "连续上涨3日"},
+            {"key": "consec_up_5", "label": "连续上涨5日"},
+        ],
+    },
+]
+
+SORT_WHITELIST = [
+    "ts_code", "pct_chg", "close", "amount", "total_mv",
+    "turnover_rate", "vol_ratio_5", "consec_up_days",
+]
+
+DISPLAY_COLUMNS = [
+    "ts_code", "name", "industry", "pct_chg", "close",
+    "amount", "total_mv", "turnover_rate", "vol_ratio_5",
+]
+
+
+class PatternScreenService:
+    """形态选股服务：读取宽表，提供分组元数据和筛选。"""
+
+    def __init__(self):
+        self._df = None
+
+    def _load_df(self) -> pd.DataFrame:
+        data_dir = current_app.config.get('DATA_DIR')
+        path = os.path.join(data_dir, 'data.parquet')
+        logger.info(f"PatternScreenService loading {path}")
+        return pd.read_parquet(path)
+
+    def _ensure_df(self) -> pd.DataFrame:
+        if self._df is None:
+            self._df = self._load_df()
+        return self._df
+
+    def get_groups(self) -> list:
+        """返回分组元数据，每个字段附带当日命中数。
+
+        Returns:
+            list[dict]: [{id, label, fields: [{key, label, count}]}]
+        """
+        df = self._ensure_df()
+        df_cols = set(df.columns)
+        result = []
+        for group in PATTERN_GROUPS:
+            fields = []
+            for f in group["fields"]:
+                if f["key"] not in df_cols:
+                    continue
+                count = int(df[f["key"]].fillna(0).sum())
+                fields.append({"key": f["key"], "label": f["label"], "count": count})
+            if fields:
+                result.append({"id": group["id"], "label": group["label"], "fields": fields})
+        return result
+
+    def screen(self, patterns=None, sort_by="pct_chg", order="desc", limit=50, offset=0):
+        """纯AND筛选 + 排序 + 分页。
+
+        Args:
+            patterns: list[str] of pattern field keys to filter (all must == 1).
+            sort_by: column to sort by (must be in SORT_WHITELIST).
+            order: "asc" or "desc".
+            limit: page size, capped at 500.
+            offset: page offset, must be >= 0.
+
+        Returns:
+            dict with total, offset, limit, trade_date, rows.
+
+        Raises:
+            ValueError: on invalid sort_by, order, or pattern keys.
+        """
+        # Validate
+        if sort_by not in SORT_WHITELIST:
+            raise ValueError(f"sort_by '{sort_by}' not in whitelist: {SORT_WHITELIST}")
+        if order not in ("asc", "desc"):
+            raise ValueError(f"order must be 'asc' or 'desc', got '{order}'")
+
+        patterns = patterns or []
+        df = self._ensure_df()
+
+        # Validate pattern keys
+        df_cols = set(df.columns)
+        for p in patterns:
+            if p not in df_cols:
+                raise ValueError(f"pattern '{p}' not found in data columns")
+
+        # Filter: pure AND
+        mask = pd.Series(True, index=df.index)
+        for p in patterns:
+            mask = mask & (df[p].fillna(0) == 1)
+        filtered = df.loc[mask]
+
+        # Sort
+        filtered = filtered.sort_values(by=sort_by, ascending=(order == "asc"))
+
+        # Paginate
+        total = len(filtered)
+        limit = max(1, min(500, int(limit)))
+        offset = max(0, int(offset))
+        page = filtered.iloc[offset:offset + limit]
+
+        # Build rows — convert NaN to None for JSON
+        rows_df = page[DISPLAY_COLUMNS].copy()
+        rows = rows_df.where(rows_df.notna(), None).to_dict(orient="records")
+
+        trade_date = str(df['trade_date'].iloc[0]) if 'trade_date' in df.columns else ''
+
+        return {
+            "total": total,
+            "offset": offset,
+            "limit": limit,
+            "trade_date": trade_date,
+            "rows": rows,
+        }
+
+    def invalidate_cache(self):
+        """清除缓存的 DataFrame，下次调用时重新读取。"""
+        self._df = None
+        logger.info("PatternScreenService cache invalidated")
+
+
+# ── 模块级单例 ──────────────────────────────────────────
+_service = None
+
+
+def get_pattern_screen_service() -> PatternScreenService:
+    """获取服务单例（缓存 DataFrame 跨请求复用）。"""
+    global _service
+    if _service is None:
+        _service = PatternScreenService()
+    return _service
+```
+
+- [ ] **Step 4: Run tests to verify they pass**
+
+Run: `pytest tests/services/test_pattern_screen_service.py -v`
+Expected: all PASS
+
+- [ ] **Step 5: Commit**
+
+```bash
+git add app/services/pattern_screen_service.py tests/services/test_pattern_screen_service.py
+git commit -m "feat: add PatternScreenService with metadata and AND-filtering"
+```
+
+---
+
+### Task 2: API endpoints — groups and screen
+
+**Files:**
+- Create: `app/api/pattern_screen_api.py`
+- Test: `tests/api/test_pattern_screen_api.py`
+
+- [ ] **Step 1: Write the failing test**
+
+```python
+# tests/api/test_pattern_screen_api.py
+"""Pattern screen API contract tests."""
+import pytest
+import json
+
+
+@pytest.fixture
+def client(app):
+    return app.test_client()
+
+
+@pytest.fixture
+def mock_service(app):
+    """Register a mock service on the app for API tests."""
+    from unittest.mock import MagicMock
+    svc = MagicMock()
+    svc.get_groups.return_value = [
+        {
+            "id": "trend_structure",
+            "label": "趋势结构",
+            "fields": [
+                {"key": "pattern_golden_cross", "label": "均线金叉", "count": 304},
+                {"key": "pattern_ma_bull", "label": "均线多头排列", "count": 259},
+            ],
+        }
+    ]
+    svc.screen.return_value = {
+        "total": 2,
+        "offset": 0,
+        "limit": 50,
+        "trade_date": "20260605",
+        "rows": [
+            {"ts_code": "000001.SZ", "name": "平安银行", "industry": "银行",
+             "pct_chg": 1.5, "close": 12.5, "amount": 1000000,
+             "total_mv": 2400000, "turnover_rate": 1.2, "vol_ratio_5": 1.8},
+            {"ts_code": "000003.SZ", "name": "国农科技", "industry": "综合",
+             "pct_chg": 3.0, "close": 25.0, "amount": 200000,
+             "total_mv": 600000, "turnover_rate": 2.5, "vol_ratio_5": 2.1},
+        ],
+    }
+
+    # Patch the singleton getter in the API module
+    from unittest.mock import patch
+    patcher = patch(
+        'app.api.pattern_screen_api.get_pattern_screen_service',
+        return_value=svc,
+    )
+    patcher.start()
+    yield svc
+    patcher.stop()
+
+
+class TestGetGroups:
+    def test_returns_200(self, client, mock_service):
+        resp = client.get('/api/pattern-screen/groups')
+        assert resp.status_code == 200
+        data = resp.get_json()
+        assert data['code'] == 200
+        assert isinstance(data['data'], list)
+
+    def test_group_structure(self, client, mock_service):
+        resp = client.get('/api/pattern-screen/groups')
+        data = resp.get_json()['data']
+        g = data[0]
+        assert 'id' in g
+        assert 'label' in g
+        assert 'fields' in g
+
+
+class TestScreen:
+    def test_returns_200(self, client, mock_service):
+        resp = client.post('/api/pattern-screen/screen',
+                           json={'patterns': ['pattern_golden_cross']})
+        assert resp.status_code == 200
+        data = resp.get_json()
+        assert data['code'] == 200
+        assert data['data']['total'] == 2
+
+    def test_empty_patterns(self, client, mock_service):
+        resp = client.post('/api/pattern-screen/screen', json={})
+        assert resp.status_code == 200
+
+    def test_invalid_sort_by_returns_400(self, client, mock_service):
+        from unittest.mock import patch
+        mock_service.screen.side_effect = ValueError("sort_by 'bad' not in whitelist")
+        resp = client.post('/api/pattern-screen/screen',
+                           json={'sort_by': 'bad'})
+        assert resp.status_code == 400
+        assert resp.get_json()['code'] == 400
+
+    def test_response_format(self, client, mock_service):
+        resp = client.post('/api/pattern-screen/screen',
+                           json={'patterns': ['pattern_golden_cross']})
+        data = resp.get_json()['data']
+        assert 'total' in data
+        assert 'offset' in data
+        assert 'limit' in data
+        assert 'trade_date' in data
+        assert 'rows' in data
+        row = data['rows'][0]
+        for col in ['ts_code', 'name', 'industry', 'pct_chg', 'close',
+                     'amount', 'total_mv', 'turnover_rate', 'vol_ratio_5']:
+            assert col in row
+```
+
+- [ ] **Step 2: Run test to verify it fails**
+
+Run: `pytest tests/api/test_pattern_screen_api.py -v`
+Expected: FAIL — module `pattern_screen_api` not found
+
+- [ ] **Step 3: Write the API blueprint**
+
+```python
+# app/api/pattern_screen_api.py
+"""形态选股 API 端点。"""
+from flask import Blueprint, request, jsonify
+from loguru import logger
+
+pattern_screen_api = Blueprint('pattern_screen_api', __name__, url_prefix='/api/pattern-screen')
+
+
+@pattern_screen_api.route('/groups', methods=['GET'])
+def get_groups():
+    """返回形态分组元数据（含命中数）。"""
+    try:
+        from app.services.pattern_screen_service import get_pattern_screen_service
+        svc = get_pattern_screen_service()
+        groups = svc.get_groups()
+        return jsonify({'code': 200, 'message': '成功', 'data': groups})
+    except Exception as e:
+        logger.error(f"获取形态分组失败: {e}")
+        return jsonify({'code': 500, 'message': f'服务器错误: {str(e)}', 'data': None}), 500
+
+
+@pattern_screen_api.route('/screen', methods=['POST'])
+def screen():
+    """执行形态筛选，返回结果表格。"""
+    try:
+        data = request.get_json() or {}
+        patterns = data.get('patterns', [])
+        sort_by = data.get('sort_by', 'pct_chg')
+        order = data.get('order', 'desc')
+        limit = data.get('limit', 50)
+        offset = data.get('offset', 0)
+
+        from app.services.pattern_screen_service import get_pattern_screen_service
+        svc = get_pattern_screen_service()
+        result = svc.screen(
+            patterns=patterns,
+            sort_by=sort_by,
+            order=order,
+            limit=limit,
+            offset=offset,
+        )
+        return jsonify({'code': 200, 'message': '成功', 'data': result})
+    except ValueError as e:
+        logger.warning(f"形态筛选参数错误: {e}")
+        return jsonify({'code': 400, 'message': str(e), 'data': None}), 400
+    except Exception as e:
+        logger.error(f"形态筛选失败: {e}")
+        return jsonify({'code': 500, 'message': f'服务器错误: {str(e)}', 'data': None}), 500
+```
+
+- [ ] **Step 4: Register the API blueprint in `app/__init__.py`**
+
+Add after the existing `heatmap_routes` import (around line 40):
+
+```python
+from app.api.pattern_screen_api import pattern_screen_api
+```
+
+Add after the `app.register_blueprint(heatmap_routes)` line (around line 54):
+
+```python
+app.register_blueprint(pattern_screen_api)
+```
+
+- [ ] **Step 5: Run tests to verify they pass**
+
+Run: `pytest tests/api/test_pattern_screen_api.py -v`
+Expected: all PASS
+
+- [ ] **Step 6: Commit**
+
+```bash
+git add app/api/pattern_screen_api.py tests/api/test_pattern_screen_api.py app/__init__.py
+git commit -m "feat: add pattern screen API endpoints (groups + screen)"
+```
+
+---
+
+### Task 3: Page route and template
+
+**Files:**
+- Create: `app/routes/pattern_screen.py`
+- Create: `app/templates/pattern_screen.html`
+- Modify: `app/__init__.py:40,54` — add page blueprint registration
+- Modify: `app/templates/base.html:335-338` — add nav link
+
+- [ ] **Step 1: Create page route**
+
+```python
+# app/routes/pattern_screen.py
+"""形态选股页面路由。"""
+from flask import Blueprint, render_template
+
+pattern_screen_bp = Blueprint('pattern_screen', __name__)
+
+
+@pattern_screen_bp.route('/pattern-screen/')
+def index():
+    """形态选股页面。"""
+    return render_template('pattern_screen.html')
+```
+
+- [ ] **Step 2: Register page blueprint in `app/__init__.py`**
+
+Add import near line 40:
+
+```python
+from app.routes.pattern_screen import pattern_screen_bp
+```
+
+Add registration near line 54:
+
+```python
+app.register_blueprint(pattern_screen_bp)
+```
+
+- [ ] **Step 3: Add nav link in `app/templates/base.html`**
+
+After the "板块热力图" nav item (around line 338), add:
+
+```html
+<li class="nav-item">
+    <a class="nav-link" href="{{ url_for('pattern_screen.index') }}">
+        <i class="fas fa-crosshairs me-1"></i>形态选股
+    </a>
+</li>
+```
+
+- [ ] **Step 4: Create the template**
+
+Create `app/templates/pattern_screen.html` with the following full content:
+
+```html
+{% extends "base.html" %}
+{% block title %}形态选股{% endblock %}
+{% block extra_css %}
+<style>
+.pattern-page { display: flex; min-height: calc(100vh - 56px); }
+.filter-panel {
+    width: 300px; min-width: 300px;
+    background: var(--obs-surface, #1e293b);
+    border-right: 1px solid var(--obs-border, #334155);
+    overflow-y: auto; padding: 16px;
+    display: flex; flex-direction: column;
+}
+.filter-panel .search-box {
+    margin-bottom: 12px;
+}
+.filter-panel .search-box input {
+    width: 100%; padding: 6px 10px; border-radius: 4px;
+    border: 1px solid var(--obs-border, #334155);
+    background: var(--obs-bg, #0f172a); color: var(--obs-text, #e2e8f0);
+    font-size: 13px;
+}
+.filter-panel .search-box input::placeholder { color: #64748b; }
+.accordion-item { border: none; margin-bottom: 4px; }
+.accordion-button {
+    background: transparent; color: #e2e8f0; font-size: 13px;
+    font-weight: 600; padding: 8px 12px; border: none; box-shadow: none;
+}
+.accordion-button:not(.collapsed) {
+    background: rgba(99, 102, 241, 0.1); color: #818cf8;
+}
+.accordion-button::after { filter: invert(0.7); }
+.accordion-body { padding: 4px 8px; }
+.pattern-check {
+    display: flex; align-items: center; padding: 3px 4px; border-radius: 3px;
+    font-size: 12px; color: #94a3b8; cursor: pointer;
+}
+.pattern-check:hover { background: rgba(99, 102, 241, 0.08); }
+.pattern-check input[type="checkbox"] {
+    margin-right: 6px; accent-color: #6366f1;
+}
+.pattern-check .hit-count {
+    margin-left: auto; color: #64748b; font-size: 11px;
+}
+.filter-actions {
+    margin-top: auto; padding-top: 12px;
+    display: flex; gap: 8px;
+}
+.filter-actions .btn { flex: 1; font-size: 13px; }
+.result-area { flex: 1; padding: 20px; overflow-x: auto; }
+.stats-bar {
+    display: flex; align-items: center; gap: 16px;
+    margin-bottom: 16px; color: #94a3b8; font-size: 14px;
+}
+.stats-bar .badge { background: rgba(99, 102, 241, 0.2); color: #818cf8; }
+.result-table {
+    width: 100%; font-size: 13px; color: #e2e8f0;
+}
+.result-table thead th {
+    background: var(--obs-surface, #1e293b); color: #94a3b8;
+    font-weight: 600; padding: 10px 12px; border-bottom: 1px solid #334155;
+    cursor: pointer; user-select: none; white-space: nowrap;
+}
+.result-table thead th:hover { color: #818cf8; }
+.result-table thead th .sort-icon { margin-left: 4px; font-size: 10px; }
+.result-table tbody td {
+    padding: 8px 12px; border-bottom: 1px solid rgba(51, 65, 85, 0.4);
+    white-space: nowrap;
+}
+.result-table tbody tr:hover { background: rgba(99, 102, 241, 0.05); }
+.pct-up { color: #ef4444; }
+.pct-down { color: #22c55e; }
+.pct-flat { color: #94a3b8; }
+.pagination-bar {
+    display: flex; align-items: center; justify-content: center;
+    gap: 6px; margin-top: 16px;
+}
+.page-btn {
+    padding: 4px 10px; border-radius: 4px; border: 1px solid #334155;
+    background: transparent; color: #94a3b8; font-size: 12px; cursor: pointer;
+}
+.page-btn:hover { border-color: #6366f1; color: #e2e8f0; }
+.page-btn.active { background: #6366f1; color: #fff; border-color: #6366f1; }
+.page-btn:disabled { opacity: 0.4; cursor: default; }
+.empty-state { text-align: center; padding: 60px 20px; color: #64748b; }
+.empty-state i { font-size: 48px; margin-bottom: 16px; }
+@media (max-width: 768px) {
+    .pattern-page { flex-direction: column; }
+    .filter-panel { width: 100%; min-width: auto; max-height: 40vh; }
+}
+</style>
+{% endblock %}
+
+{% block content %}
+<div class="pattern-page">
+  <!-- Left Panel -->
+  <div class="filter-panel">
+    <div class="search-box">
+      <input type="text" id="searchInput" placeholder="搜索形态名称..." oninput="filterBySearch()">
+    </div>
+    <div class="accordion" id="groupAccordion"></div>
+    <div class="filter-actions">
+      <button class="btn btn-outline-secondary" onclick="resetFilters()">
+        <i class="fas fa-undo me-1"></i>重置
+      </button>
+      <button class="btn btn-primary" onclick="doScreen()" style="background:#6366f1;border-color:#6366f1;">
+        <i class="fas fa-search me-1"></i>筛选
+      </button>
+    </div>
+  </div>
+
+  <!-- Right Area -->
+  <div class="result-area">
+    <div class="stats-bar" id="statsBar">
+      <span>交易日: <strong id="tradeDate">--</strong></span>
+      <span>|</span>
+      <span>已选: <span class="badge" id="selectedCount">0</span> 个形态</span>
+      <span>|</span>
+      <span>匹配: <strong id="totalCount">0</strong> 只</span>
+    </div>
+    <div id="tableContainer">
+      <div class="empty-state">
+        <i class="fas fa-crosshairs d-block"></i>
+        <div>加载中...</div>
+      </div>
+    </div>
+    <div class="pagination-bar" id="paginationBar"></div>
+  </div>
+</div>
+
+<script>
+// ── State ──
+let currentSort = 'pct_chg';
+let currentOrder = 'desc';
+let currentOffset = 0;
+let currentLimit = 50;
+let currentTotal = 0;
+let allGroups = [];
+
+// ── Init ──
+document.addEventListener('DOMContentLoaded', () => {
+    loadGroups();
+    doScreen();
+});
+
+// ── Load groups into left panel ──
+async function loadGroups() {
+    try {
+        const resp = await axios.get('/api/pattern-screen/groups');
+        const data = resp.data.data;
+        if (!data) return;
+        allGroups = data;
+        renderGroups(data);
+    } catch (e) {
+        console.error('加载分组失败:', e);
+    }
+}
+
+function renderGroups(groups) {
+    const container = document.getElementById('groupAccordion');
+    container.innerHTML = '';
+    groups.forEach((g, idx) => {
+        const show = idx < 2; // first two expanded by default
+        container.innerHTML += `
+        <div class="accordion-item">
+          <h2 class="accordion-header">
+            <button class="accordion-button ${show ? '' : 'collapsed'}" type="button"
+                    data-bs-toggle="collapse" data-bs-target="#group-${g.id}">
+              ${g.label} <span style="margin-left:auto;font-weight:400;font-size:11px;color:#64748b;">${g.fields.length}</span>
+            </button>
+          </h2>
+          <div id="group-${g.id}" class="accordion-collapse collapse ${show ? 'show' : ''}"
+               data-bs-parent="#groupAccordion">
+            <div class="accordion-body">
+              ${g.fields.map(f => `
+                <label class="pattern-check" data-label="${f.label}">
+                  <input type="checkbox" value="${f.key}" onchange="updateSelectedCount()">
+                  ${f.label}
+                  <span class="hit-count">${f.count}</span>
+                </label>
+              `).join('')}
+            </div>
+          </div>
+        </div>`;
+    });
+}
+
+// ── Search filter ──
+function filterBySearch() {
+    const q = document.getElementById('searchInput').value.trim().toLowerCase();
+    document.querySelectorAll('.pattern-check').forEach(el => {
+        const label = el.getAttribute('data-label').toLowerCase();
+        el.style.display = (!q || label.includes(q)) ? '' : 'none';
+    });
+}
+
+// ── Selected count ──
+function updateSelectedCount() {
+    const count = document.querySelectorAll('.pattern-check input:checked').length;
+    document.getElementById('selectedCount').textContent = count;
+}
+
+// ── Get checked patterns ──
+function getCheckedPatterns() {
+    return Array.from(document.querySelectorAll('.pattern-check input:checked'))
+        .map(cb => cb.value);
+}
+
+// ── Screen ──
+async function doScreen() {
+    const patterns = getCheckedPatterns();
+    try {
+        const resp = await axios.post('/api/pattern-screen/screen', {
+            patterns,
+            sort_by: currentSort,
+            order: currentOrder,
+            limit: currentLimit,
+            offset: currentOffset,
+        });
+        const result = resp.data.data;
+        if (!result) return;
+        currentTotal = result.total;
+        document.getElementById('tradeDate').textContent = formatDate(result.trade_date);
+        document.getElementById('totalCount').textContent = result.total;
+        renderTable(result.rows);
+        renderPagination(result.total, result.limit, result.offset);
+    } catch (e) {
+        console.error('筛选失败:', e);
+        document.getElementById('tableContainer').innerHTML =
+            '<div class="empty-state"><i class="fas fa-exclamation-triangle d-block"></i><div>筛选失败，请重试</div></div>';
+    }
+}
+
+// ── Render table ──
+function renderTable(rows) {
+    if (!rows || rows.length === 0) {
+        document.getElementById('tableContainer').innerHTML =
+            '<div class="empty-state"><i class="fas fa-inbox d-block"></i><div>暂无匹配结果</div></div>';
+        return;
+    }
+    const sortIcon = (col) => {
+        if (col !== currentSort) return '<i class="fas fa-sort sort-icon" style="opacity:0.3"></i>';
+        return currentOrder === 'desc'
+            ? '<i class="fas fa-sort-down sort-icon"></i>'
+            : '<i class="fas fa-sort-up sort-icon"></i>';
+    };
+    let html = `<table class="result-table">
+      <thead><tr>
+        <th onclick="sortTable('ts_code')">代码 ${sortIcon('ts_code')}</th>
+        <th>名称</th>
+        <th>行业</th>
+        <th onclick="sortTable('pct_chg')">涨跌幅 ${sortIcon('pct_chg')}</th>
+        <th onclick="sortTable('close')">现价 ${sortIcon('close')}</th>
+        <th onclick="sortTable('amount')">成交额 ${sortIcon('amount')}</th>
+        <th onclick="sortTable('total_mv')">总市值 ${sortIcon('total_mv')}</th>
+        <th onclick="sortTable('turnover_rate')">换手率 ${sortIcon('turnover_rate')}</th>
+        <th onclick="sortTable('vol_ratio_5')">量比 ${sortIcon('vol_ratio_5')}</th>
+      </tr></thead><tbody>`;
+    rows.forEach(r => {
+        const pctClass = r.pct_chg > 0 ? 'pct-up' : (r.pct_chg < 0 ? 'pct-down' : 'pct-flat');
+        const pctStr = r.pct_chg > 0 ? '+' + r.pct_chg.toFixed(2) : r.pct_chg.toFixed(2);
+        html += `<tr>
+          <td>${r.ts_code || ''}</td>
+          <td>${r.name || ''}</td>
+          <td>${r.industry || '--'}</td>
+          <td class="${pctClass}">${pctStr}%</td>
+          <td>${fmtNum(r.close, 2)}</td>
+          <td>${fmtAmount(r.amount)}</td>
+          <td>${fmtAmount(r.total_mv)}</td>
+          <td>${fmtNum(r.turnover_rate, 2)}%</td>
+          <td>${fmtNum(r.vol_ratio_5, 2)}</td>
+        </tr>`;
+    });
+    html += '</tbody></table>';
+    document.getElementById('tableContainer').innerHTML = html;
+}
+
+// ── Sort ──
+function sortTable(field) {
+    if (currentSort === field) {
+        currentOrder = currentOrder === 'desc' ? 'asc' : 'desc';
+    } else {
+        currentSort = field;
+        currentOrder = 'desc';
+    }
+    currentOffset = 0;
+    doScreen();
+}
+
+// ── Pagination ──
+function renderPagination(total, limit, offset) {
+    const bar = document.getElementById('paginationBar');
+    if (total <= limit) { bar.innerHTML = ''; return; }
+    const pages = Math.ceil(total / limit);
+    const cur = Math.floor(offset / limit) + 1;
+    let html = `<button class="page-btn" onclick="gotoPage(${offset - limit})" ${cur === 1 ? 'disabled' : ''}>
+        <i class="fas fa-chevron-left"></i></button>`;
+    // Show max 7 page buttons
+    let start = Math.max(1, cur - 3);
+    let end = Math.min(pages, start + 6);
+    start = Math.max(1, end - 6);
+    for (let i = start; i <= end; i++) {
+        const off = (i - 1) * limit;
+        html += `<button class="page-btn ${i === cur ? 'active' : ''}" onclick="gotoPage(${off})">${i}</button>`;
+    }
+    html += `<button class="page-btn" onclick="gotoPage(${offset + limit})" ${cur === pages ? 'disabled' : ''}>
+        <i class="fas fa-chevron-right"></i></button>`;
+    bar.innerHTML = html;
+}
+
+function gotoPage(offset) {
+    currentOffset = Math.max(0, offset);
+    doScreen();
+}
+
+// ── Reset ──
+function resetFilters() {
+    document.querySelectorAll('.pattern-check input').forEach(cb => cb.checked = false);
+    document.getElementById('searchInput').value = '';
+    filterBySearch();
+    updateSelectedCount();
+    currentOffset = 0;
+    currentSort = 'pct_chg';
+    currentOrder = 'desc';
+    doScreen();
+}
+
+// ── Formatters ──
+function fmtNum(v, digits) {
+    if (v == null) return '--';
+    return Number(v).toFixed(digits);
+}
+
+function fmtAmount(v) {
+    if (v == null) return '--';
+    v = Number(v);
+    if (Math.abs(v) >= 1e8) return (v / 1e8).toFixed(2) + '亿';
+    if (Math.abs(v) >= 1e4) return (v / 1e4).toFixed(2) + '万';
+    return v.toFixed(2);
+}
+
+function formatDate(d) {
+    if (!d || d.length !== 8) return d || '--';
+    return d.slice(0, 4) + '-' + d.slice(4, 6) + '-' + d.slice(6, 8);
+}
+</script>
+{% endblock %}
+```
+
+- [ ] **Step 5: Verify page loads**
+
+Run: `python run.py` and open `http://localhost:5000/pattern-screen/`
+Expected: page renders with left panel showing grouped checkboxes and right panel showing stock table
+
+- [ ] **Step 6: Verify nav link works**
+
+From any page, click "形态选股" in the nav bar.
+Expected: navigates to `/pattern-screen/`
+
+- [ ] **Step 7: Commit**
+
+```bash
+git add app/routes/pattern_screen.py app/templates/pattern_screen.html app/__init__.py app/templates/base.html
+git commit -m "feat: add pattern screening page with left panel + right table layout"
+```
+
+---
+
+### Task 4: Integration verification
+
+**Files:** None new — verification only.
+
+- [ ] **Step 1: Run all tests**
+
+Run: `pytest -v`
+Expected: all existing tests + new pattern screen tests pass
+
+- [ ] **Step 2: Manual smoke test**
+
+1. `python run.py`
+2. Open `/pattern-screen/`
+3. Verify: groups load with hit counts
+4. Check a pattern (e.g. "均线金叉") → click "筛选" → verify table filters
+5. Check a second pattern (e.g. "均线多头排列") → verify AND logic (fewer results)
+6. Click a table header → verify sorting changes
+7. Click page 2 → verify pagination
+8. Click "重置" → verify all checkboxes clear and table shows all stocks
+9. Type in search box → verify checkbox labels filter
+
+- [ ] **Step 3: Commit any fixes**
+
+```bash
+git add -A && git commit -m "fix: pattern screen integration fixes"
+```
diff --git a/docs/superpowers/plans/2026-06-07-sector-heatmap.md b/docs/superpowers/plans/2026-06-07-sector-heatmap.md
new file mode 100644
index 000000000..f974450c5
--- /dev/null
+++ b/docs/superpowers/plans/2026-06-07-sector-heatmap.md
@@ -0,0 +1,570 @@
+# 板块热力图 Implementation Plan
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+
+**Goal:** Add a Treemap heatmap page showing A-share sector performance with in-page drill-down to individual stocks.
+
+**Architecture:** Flask route → HeatmapService reads `data/data.parquet` → injects JSON into Jinja2 template → ECharts renders treemap + JS handles click-to-expand stock table. No new dependencies.
+
+**Tech Stack:** Flask, pandas, ECharts 5.4.3, Bootstrap 5.1.3, Jinja2
+
+**Spec:** `docs/superpowers/specs/2026-06-07-sector-heatmap-design.md`
+
+---
+
+## File Structure
+
+| Action | File | Responsibility |
+|--------|------|----------------|
+| Create | `app/services/heatmap_service.py` | Read parquet, aggregate by industry, return JSON |
+| Create | `app/routes/heatmap.py` | Blueprint with `GET /heatmap` route |
+| Create | `app/templates/heatmap.html` | ECharts treemap + expandable stock table |
+| Modify | `app/__init__.py:38-52` | Register `heatmap_routes` blueprint |
+| Modify | `app/templates/base.html:333` | Add nav link before 多因子模型 dropdown |
+| Create | `tests/services/test_heatmap_service.py` | Unit tests for HeatmapService |
+
+---
+
+### Task 1: HeatmapService — Failing Tests
+
+**Files:**
+- Create: `tests/services/test_heatmap_service.py`
+
+- [ ] **Step 1: Write failing tests**
+
+```python
+"""Tests for HeatmapService — sector aggregation logic."""
+import json
+import pytest
+from unittest.mock import patch, MagicMock
+import pandas as pd
+import numpy as np
+
+
+@pytest.fixture
+def sample_df():
+    """Minimal DataFrame matching data/data.parquet schema."""
+    return pd.DataFrame({
+        'ts_code': ['000001.SZ', '000002.SZ', '000003.SZ', '000004.SZ', '000005.SZ'],
+        'name': ['平安银行', '万科Ａ', '测试银行', '测试地产', '停牌股'],
+        'industry': ['银行', '全国地产', '银行', '全国地产', '全国地产'],
+        'pct_chg': [1.5, -2.0, 0.5, -1.0, 0.0],
+        'close': [11.0, 3.5, 5.0, 8.0, 0.0],
+        'total_mv': [21300, 3900, 5000, 2000, 100],
+        'net_mf_amount': [14000, -5000, 3000, -2000, 0],
+        'turnover_rate': [0.5, 1.7, 1.0, 2.0, 0.0],
+        'trade_date': ['20260605'] * 5,
+    })
+
+
+@pytest.fixture
+def service():
+    from app.services.heatmap_service import HeatmapService
+    return HeatmapService()
+
+
+class TestGetHeatmapData:
+    """Test HeatmapService.get_heatmap_data()."""
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_returns_two_lists(self, mock_read, service, sample_df):
+        mock_read.return_value = sample_df
+        sectors, stocks = service.get_heatmap_data()
+        assert isinstance(sectors, list)
+        assert isinstance(stocks, list)
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_filters_suspended_stocks(self, mock_read, service, sample_df):
+        """Stocks with close == 0 should be excluded."""
+        mock_read.return_value = sample_df
+        sectors, stocks = service.get_heatmap_data()
+        # 停牌股 (close=0) should not appear in stocks
+        stock_names = [s['name'] for s in stocks]
+        assert '停牌股' not in stock_names
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_sector_count(self, mock_read, service, sample_df):
+        mock_read.return_value = sample_df
+        sectors, _ = service.get_heatmap_data()
+        industry_names = [s['name'] for s in sectors]
+        assert len(industry_names) == 2  # 银行 + 全国地产
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_sector_weighted_pct_chg(self, mock_read, service, sample_df):
+        """avg_pct_chg should be market-cap weighted average."""
+        mock_read.return_value = sample_df
+        sectors, _ = service.get_heatmap_data()
+        bank = next(s for s in sectors if s['name'] == '银行')
+        # Weighted: (1.5*21300 + 0.5*5000) / (21300+5000) = 34450/26300 ≈ 1.3103
+        assert abs(bank['avg_pct_chg'] - 1.3103) < 0.01
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_sector_stock_count(self, mock_read, service, sample_df):
+        mock_read.return_value = sample_df
+        sectors, _ = service.get_heatmap_data()
+        bank = next(s for s in sectors if s['name'] == '银行')
+        assert bank['stock_count'] == 2  # 2 银行 after filtering
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_sector_up_down_count(self, mock_read, service, sample_df):
+        mock_read.return_value = sample_df
+        sectors, _ = service.get_heatmap_data()
+        realestate = next(s for s in sectors if s['name'] == '全国地产')
+        # 万科 -2.0, 测试地产 -1.0 (停牌股 filtered out)
+        assert realestate['down_count'] == 2
+        assert realestate['up_count'] == 0
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_stocks_have_required_fields(self, mock_read, service, sample_df):
+        mock_read.return_value = sample_df
+        _, stocks = service.get_heatmap_data()
+        required = {'name', 'ts_code', 'pct_chg', 'close', 'total_mv',
+                    'net_mf_amount', 'turnover_rate', 'industry'}
+        for s in stocks:
+            assert required.issubset(s.keys()), f"Missing keys: {required - s.keys()}"
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_trade_date_returned(self, mock_read, service, sample_df):
+        mock_read.return_value = sample_df
+        sectors, _ = service.get_heatmap_data()
+        assert sectors[0]['trade_date'] == '20260605'
+```
+
+- [ ] **Step 2: Run tests to verify they fail**
+
+Run: `pytest tests/services/test_heatmap_service.py -v`
+Expected: FAIL — `ModuleNotFoundError: No module named 'app.services.heatmap_service'`
+
+---
+
+### Task 2: HeatmapService — Implementation
+
+**Files:**
+- Create: `app/services/heatmap_service.py`
+
+- [ ] **Step 1: Write HeatmapService**
+
+```python
+"""板块热力图数据服务 — 读取 data/data.parquet 并按行业聚合。"""
+import os
+import pandas as pd
+import numpy as np
+from loguru import logger
+
+
+class HeatmapService:
+    """板块热力图数据聚合服务。"""
+
+    def __init__(self):
+        self._data_path = os.path.join(
+            os.path.dirname(os.path.dirname(os.path.dirname(__file__))),
+            'data', 'data.parquet'
+        )
+
+    def get_heatmap_data(self):
+        """读取 parquet，返回 (sectors_json, stocks_json)。
+
+        Returns:
+            tuple: (list[dict], list[dict])
+                - sectors: 板块聚合数据，按 total_mv 降序
+                - stocks: 全量个股数据（已过滤停牌）
+        """
+        df = pd.read_parquet(self._data_path)
+        trade_date = str(df['trade_date'].iloc[0])
+
+        # 过滤停牌/退市
+        df = df[df['close'] > 0].copy()
+
+        # 板块聚合
+        sectors = []
+        for industry, group in df.groupby('industry'):
+            total_mv_sum = group['total_mv'].sum()
+            if total_mv_sum == 0:
+                continue
+            avg_pct_chg = np.average(group['pct_chg'], weights=group['total_mv'])
+            sectors.append({
+                'name': industry,
+                'avg_pct_chg': round(avg_pct_chg, 2),
+                'total_mv': round(total_mv_sum, 2),
+                'stock_count': len(group),
+                'up_count': int((group['pct_chg'] > 0).sum()),
+                'down_count': int((group['pct_chg'] < 0).sum()),
+                'net_mf_amount': round(group['net_mf_amount'].sum(), 2),
+                'trade_date': trade_date,
+            })
+
+        sectors.sort(key=lambda x: x['total_mv'], reverse=True)
+
+        # 个股明细
+        stock_cols = ['ts_code', 'name', 'industry', 'pct_chg', 'close',
+                      'total_mv', 'net_mf_amount', 'turnover_rate']
+        stocks_df = df[stock_cols].sort_values('pct_chg', ascending=False)
+        stocks = stocks_df.where(stocks_df.notna(), None).to_dict(orient='records')
+
+        return sectors, stocks
+```
+
+- [ ] **Step 2: Run tests to verify they pass**
+
+Run: `pytest tests/services/test_heatmap_service.py -v`
+Expected: All 8 tests PASS
+
+- [ ] **Step 3: Commit**
+
+```bash
+git add app/services/heatmap_service.py tests/services/test_heatmap_service.py
+git commit -m "feat: add HeatmapService with sector aggregation logic"
+```
+
+---
+
+### Task 3: Heatmap Route
+
+**Files:**
+- Create: `app/routes/heatmap.py`
+
+- [ ] **Step 1: Create route blueprint**
+
+```python
+"""板块热力图页面路由。"""
+from flask import Blueprint, render_template
+from loguru import logger
+from app.services.heatmap_service import HeatmapService
+
+heatmap_routes = Blueprint('heatmap_routes', __name__, url_prefix='/heatmap')
+
+
+@heatmap_routes.route('/')
+def index():
+    """板块热力图页面。"""
+    try:
+        service = HeatmapService()
+        sectors, stocks = service.get_heatmap_data()
+        trade_date = sectors[0]['trade_date'] if sectors else ''
+        return render_template(
+            'heatmap.html',
+            sectors_json=sectors,
+            stocks_json=stocks,
+            trade_date=trade_date,
+        )
+    except Exception as e:
+        logger.error(f"热力图加载失败: {e}")
+        return render_template(
+            'heatmap.html',
+            sectors_json=[],
+            stocks_json=[],
+            trade_date='',
+            error='数据加载失败，请确认 data/data.parquet 是否存在',
+        )
+```
+
+- [ ] **Step 2: Commit**
+
+```bash
+git add app/routes/heatmap.py
+git commit -m "feat: add heatmap route blueprint"
+```
+
+---
+
+### Task 4: Register Blueprint + Nav Link
+
+**Files:**
+- Modify: `app/__init__.py:38-52`
+- Modify: `app/templates/base.html:333`
+
+- [ ] **Step 1: Register heatmap_routes in app/__init__.py**
+
+Add import at line 39 (after `realtime_analysis_routes`):
+```python
+from app.routes.heatmap import heatmap_routes
+```
+
+Add registration at line 52 (after `app.register_blueprint(realtime_analysis_routes)`):
+```python
+app.register_blueprint(heatmap_routes)
+```
+
+- [ ] **Step 2: Add nav link in base.html**
+
+Insert after line 333 (after the 选股筛选 nav-item, before 多因子模型 dropdown):
+```html
+                    <li class="nav-item">
+                        <a class="nav-link" href="/heatmap">
+                            <i class="fas fa-th-large me-1"></i>板块热力图
+                        </a>
+                    </li>
+```
+
+- [ ] **Step 3: Verify app starts**
+
+Run: `python -c "from app import create_app; app = create_app('development'); print('OK')" `
+Expected: prints `OK`
+
+- [ ] **Step 4: Commit**
+
+```bash
+git add app/__init__.py app/templates/base.html
+git commit -m "feat: register heatmap blueprint and add nav link"
+```
+
+---
+
+### Task 5: Heatmap Template
+
+**Files:**
+- Create: `app/templates/heatmap.html`
+
+- [ ] **Step 1: Create the template**
+
+```html
+{% extends "base.html" %}
+
+{% block title %}板块热力图{% endblock %}
+
+{% block extra_css %}
+<style>
+.heatmap-page { padding: 20px 0; }
+.heatmap-header {
+    display: flex; align-items: center; justify-content: space-between;
+    margin-bottom: 16px; flex-wrap: wrap; gap: 12px;
+}
+.heatmap-title { font-size: 20px; font-weight: 600; color: #e2e8f0; }
+.heatmap-date { color: #94a3b8; font-size: 14px; }
+.heatmap-legend {
+    display: flex; align-items: center; gap: 8px; font-size: 12px; color: #94a3b8;
+}
+.legend-bar {
+    width: 120px; height: 12px; border-radius: 3px;
+    background: linear-gradient(to right, #27ae60, #a9dfbf, #bdc3c7, #f5b7b1, #c0392b);
+}
+#treemap-chart { width: 100%; height: 520px; }
+.stock-detail-panel {
+    margin-top: 16px; display: none;
+    animation: slideDown 0.3s ease;
+}
+@keyframes slideDown {
+    from { opacity: 0; transform: translateY(-10px); }
+    to { opacity: 1; transform: translateY(0); }
+}
+.panel-header {
+    display: flex; align-items: center; justify-content: space-between;
+    padding: 12px 16px; background: #1e293b; border-radius: 8px 8px 0 0;
+    border-bottom: 1px solid #334155;
+}
+.panel-title { font-size: 16px; font-weight: 600; color: #e2e8f0; }
+.panel-summary { font-size: 13px; color: #94a3b8; }
+.panel-close {
+    background: none; border: none; color: #94a3b8; cursor: pointer;
+    font-size: 18px; padding: 0 4px;
+}
+.panel-close:hover { color: #e2e8f0; }
+.stock-table { width: 100%; }
+.stock-table thead th {
+    background: #0f172a; color: #94a3b8; font-size: 12px; font-weight: 500;
+    padding: 8px 12px; border-bottom: 1px solid #1e293b; position: sticky; top: 0;
+    text-transform: uppercase; letter-spacing: 0.5px;
+}
+.stock-table tbody tr { border-bottom: 1px solid #1e293b; }
+.stock-table tbody tr:hover { background: rgba(99, 102, 241, 0.08); }
+.stock-table td { padding: 8px 12px; color: #e2e8f0; font-size: 13px; }
+.stock-table td:first-child { font-weight: 500; }
+.pct-up { color: #ef4444; }
+.pct-down { color: #22c55e; }
+.pct-flat { color: #94a3b8; }
+.error-banner {
+    padding: 20px; text-align: center; color: #ef4444;
+    background: #1e293b; border-radius: 8px; margin: 20px 0;
+}
+</style>
+{% endblock %}
+
+{% block content %}
+<div class="container-fluid heatmap-page">
+    {% if error is defined and error %}
+    <div class="error-banner">{{ error }}</div>
+    {% else %}
+    <div class="heatmap-header">
+        <div>
+            <span class="heatmap-title">板块热力图</span>
+            <span class="heatmap-date" style="margin-left:12px;">{{ trade_date }}</span>
+        </div>
+        <div class="heatmap-legend">
+            <span>跌</span>
+            <div class="legend-bar"></div>
+            <span>涨</span>
+        </div>
+    </div>
+    <div id="treemap-chart"></div>
+    <div id="stock-detail" class="stock-detail-panel"></div>
+    {% endif %}
+</div>
+{% endblock %}
+
+{% block extra_js %}
+<script>
+(function() {
+    var sectors = {{ sectors_json | tojson }};
+    var stocks = {{ stocks_json | tojson }};
+
+    // ---- Treemap ----
+    var chart = echarts.init(document.getElementById('treemap-chart'), 'financial');
+
+    function pctColor(val) {
+        if (val > 0) {
+            var t = Math.min(val / 5, 1);
+            return 'rgb(' + Math.round(192 + (245-192)*(1-t)) + ',' +
+                            Math.round(57 + (183-57)*(1-t)) + ',' +
+                            Math.round(43 + (177-43)*(1-t)) + ')';
+        } else if (val < 0) {
+            var t = Math.min(Math.abs(val) / 5, 1);
+            return 'rgb(' + Math.round(39 + (169-39)*(1-t)) + ',' +
+                            Math.round(174 + (223-174)*(1-t)) + ',' +
+                            Math.round(96 + (191-96)*(1-t)) + ')';
+        }
+        return '#bdc3c7';
+    }
+
+    var treemapData = sectors.map(function(s) {
+        return {
+            name: s.name,
+            value: s.total_mv,
+            avg_pct_chg: s.avg_pct_chg,
+            stock_count: s.stock_count,
+            up_count: s.up_count,
+            down_count: s.down_count,
+            net_mf_amount: s.net_mf_amount,
+            itemStyle: { color: pctColor(s.avg_pct_chg) }
+        };
+    });
+
+    var option = {
+        tooltip: {
+            formatter: function(info) {
+                var d = info.data;
+                return '<b>' + d.name + '</b><br/>' +
+                    '加权涨跌: <b>' + d.avg_pct_chg.toFixed(2) + '%</b><br/>' +
+                    '涨/跌: ' + d.up_count + '/' + d.down_count +
+                    ' (共' + d.stock_count + '只)<br/>' +
+                    '净流入: ' + (d.net_mf_amount / 10000).toFixed(2) + '亿';
+            }
+        },
+        series: [{
+            type: 'treemap',
+            data: treemapData,
+            roam: false,
+            nodeClick: false,
+            breadcrumb: { show: false },
+            levels: [{
+                itemStyle: { borderColor: '#0f172a', borderWidth: 2, gapWidth: 2 },
+                upperLabel: { show: false }
+            }],
+            label: {
+                show: true,
+                formatter: function(params) {
+                    var d = params.data;
+                    var sign = d.avg_pct_chg >= 0 ? '+' : '';
+                    return d.name + '\n' + sign + d.avg_pct_chg.toFixed(2) + '%';
+                },
+                fontSize: 13,
+                color: '#fff',
+                textShadowColor: 'rgba(0,0,0,0.5)',
+                textShadowBlur: 4
+            }
+        }]
+    };
+
+    chart.setOption(option);
+
+    // ---- Sector Click → Stock Table ----
+    var activeIndustry = null;
+
+    chart.on('click', function(params) {
+        var industry = params.data.name;
+        var panel = document.getElementById('stock-detail');
+
+        if (activeIndustry === industry) {
+            panel.style.display = 'none';
+            activeIndustry = null;
+            return;
+        }
+        activeIndustry = industry;
+
+        var sector = sectors.find(function(s) { return s.name === industry; });
+        var filtered = stocks.filter(function(s) { return s.industry === industry; });
+
+        var sign = sector.avg_pct_chg >= 0 ? '+' : '';
+        var html = '<div class="panel-header">' +
+            '<div><span class="panel-title">' + industry + '（' + filtered.length + '只）</span>' +
+            '<span class="panel-summary" style="margin-left:16px;">加权涨跌: ' +
+            '<span style="color:' + (sector.avg_pct_chg >= 0 ? '#ef4444' : '#22c55e') + '">' +
+            sign + sector.avg_pct_chg.toFixed(2) + '%</span></span></div>' +
+            '<button class="panel-close" onclick="document.getElementById(\'stock-detail\').style.display=\'none\';' +
+            'window._activeIndustry=null;">&times;</button>' +
+            '</div>' +
+            '<table class="stock-table"><thead><tr>' +
+            '<th>名称</th><th>涨跌幅</th><th>收盘价</th><th>总市值(亿)</th>' +
+            '<th>净流入(万)</th><th>换手率</th>' +
+            '</tr></thead><tbody>';
+
+        filtered.forEach(function(s) {
+            var cls = s.pct_chg > 0 ? 'pct-up' : (s.pct_chg < 0 ? 'pct-down' : 'pct-flat');
+            var pSign = s.pct_chg > 0 ? '+' : '';
+            html += '<tr>' +
+                '<td>' + s.name + '</td>' +
+                '<td class="' + cls + '">' + pSign + (s.pct_chg != null ? s.pct_chg.toFixed(2) : '--') + '%</td>' +
+                '<td>' + (s.close != null ? s.close.toFixed(2) : '--') + '</td>' +
+                '<td>' + (s.total_mv != null ? (s.total_mv / 10000).toFixed(1) : '--') + '</td>' +
+                '<td class="' + (s.net_mf_amount > 0 ? 'pct-up' : (s.net_mf_amount < 0 ? 'pct-down' : '')) + '">' +
+                (s.net_mf_amount != null ? s.net_mf_amount.toFixed(0) : '--') + '</td>' +
+                '<td>' + (s.turnover_rate != null ? s.turnover_rate.toFixed(2) + '%' : '--') + '</td>' +
+                '</tr>';
+        });
+
+        html += '</tbody></table>';
+        panel.innerHTML = html;
+        panel.style.display = 'block';
+        window._activeIndustry = industry;
+    });
+
+    window.addEventListener('resize', function() { chart.resize(); });
+})();
+</script>
+{% endblock %}
+```
+
+- [ ] **Step 2: Commit**
+
+```bash
+git add app/templates/heatmap.html
+git commit -m "feat: add heatmap template with ECharts treemap and stock table"
+```
+
+---
+
+### Task 6: End-to-End Verification
+
+- [ ] **Step 1: Run all tests**
+
+Run: `pytest -v`
+Expected: All existing + new tests PASS
+
+- [ ] **Step 2: Start the app and verify**
+
+Run: `python run.py`
+Open browser → `http://localhost:5000/heatmap`
+
+Verify:
+- [ ] Treemap renders with colored rectangles sized by market cap
+- [ ] Hover shows tooltip with sector details
+- [ ] Click sector → stock table expands below
+- [ ] Click same sector → table collapses
+- [ ] Click different sector → table switches
+- [ ] Nav bar shows "板块热力图" link and it navigates correctly
+- [ ] Red = up, green = down
+
+- [ ] **Step 3: Final commit (if any fixes needed)**
+
+```bash
+git add -A && git commit -m "fix: heatmap end-to-end adjustments"
+```
diff --git a/docs/superpowers/specs/2026-06-07-pattern-screen-design.md b/docs/superpowers/specs/2026-06-07-pattern-screen-design.md
new file mode 100644
index 000000000..414f6685e
--- /dev/null
+++ b/docs/superpowers/specs/2026-06-07-pattern-screen-design.md
@@ -0,0 +1,180 @@
+# 形态选股功能设计文档
+
+## 概述
+
+为 quantitative_analysis 项目添加形态选股（Pattern Screening）功能。基于 `data/data.parquet` 宽表中的 132 个 `pattern_*` / `break_high_*` / `consec_up_*` 二值字段，提供分组筛选界面和 API。
+
+参考项目：`/Users/henrylin/vscode_space/stock_screener/backend`
+
+## 方案选型
+
+**Approach A: 独立页面 + 独立服务**
+
+理由：`data/data.parquet` 是预计算宽表（5524 行 × 132 pattern 列），与现有 `ParquetDataReader` 的分区表体系不同，直接 pandas 读取最简单高效，无需耦合现有选股模块。
+
+## 后端架构
+
+### 文件结构
+
+```
+app/services/pattern_screen_service.py   # 服务层
+app/api/pattern_screen_api.py            # API 层
+app/routes/pattern_screen.py             # 页面路由
+app/templates/pattern_screen.html        # 页面模板
+```
+
+### 数据层
+
+- 通过 `current_app.config['DATA_DIR']` 解析路径，读取 `{DATA_DIR}/data.parquet` 到内存，缓存 DataFrame
+- 132 个二值形态字段 + 基础字段（ts_code, name, industry, pct_chg, close, amount, total_mv, turnover_rate, vol_ratio_5 等）
+- 提供字段元数据（分组定义、中文标签、当日命中数）
+- 提供 `invalidate_cache()` 方法，在宽表重建 data job 完成后调用以刷新缓存
+
+### 服务层 — PatternScreenService
+
+```python
+class PatternScreenService:
+    _df: pd.DataFrame          # 缓存宽表
+    _field_meta: list[dict]    # 分组元数据
+
+    def get_groups() -> list[dict]
+        # 返回 [{id, label, fields: [{key, label, count}]}]
+
+    def screen(patterns, sort_by, order, limit, offset) -> dict
+        # 纯 AND 筛选：所有勾选字段必须 == 1
+        # patterns: list[str]，每个 key 必须存在于 DataFrame 列中，否则返回 400
+        # sort_by: 必须在白名单内: ["pct_chg", "close", "amount", "total_mv",
+        #          "turnover_rate", "vol_ratio_5", "consec_up_days"]，默认 "pct_chg"
+        # order: 仅接受 "asc" 或 "desc"，默认 "desc"
+        # limit: 1-500，默认 50
+        # offset: >= 0，默认 0
+        # 返回 {total, offset, limit, trade_date, rows: [...]}
+
+    def invalidate_cache()
+        # 清除缓存的 DataFrame，下次调用时重新读取 parquet
+```
+
+### API 端点
+
+| 端点 | 方法 | 用途 |
+|---|---|---|
+| `/api/pattern-screen/groups` | GET | 返回分组元数据（含命中数） |
+| `/api/pattern-screen/screen` | POST | 执行筛选，返回结果表格 |
+
+POST `/api/pattern-screen/screen` 请求体：
+
+```json
+{
+  "patterns": ["pattern_golden_cross", "pattern_ma_bull"],
+  "sort_by": "pct_chg",
+  "order": "desc",
+  "limit": 50,
+  "offset": 0
+}
+```
+
+响应格式（遵循 `analysis_api.py` 的 `{code, message, data}` 约定）：
+
+```json
+{
+  "code": 200,
+  "message": "成功",
+  "data": {
+    "total": 42,
+    "offset": 0,
+    "limit": 50,
+    "trade_date": "20260605",
+    "rows": [
+      {
+        "ts_code": "000001.SZ",
+        "name": "平安银行",
+        "industry": "银行",
+        "pct_chg": 2.5,
+        "close": 12.5,
+        "amount": 1500000,
+        "total_mv": 24000000,
+        "turnover_rate": 1.2,
+        "vol_ratio_5": 1.8
+      }
+    ]
+  }
+}
+```
+
+#### 参数校验规则
+
+- `patterns`: 可选，默认 `[]`（返回全部股票）；若提供，每个 key 必须在 DataFrame 列中，否则 400
+- `sort_by`: 可选，默认 `"pct_chg"`；必须在白名单中，否则 400
+- `order`: 可选，默认 `"desc"`；仅接受 `"asc"` / `"desc"`，否则 400
+- `limit`: 可选，默认 50；范围 1-500，超出截断
+- `offset`: 可选，默认 0；必须 >= 0
+
+### 形态分组定义
+
+直接从参考项目 `meta.py` 的 `_PATTERN_GROUPS` 提取前 7 组（legacy groups）：
+
+| 分组 ID | 中文名 | 字段数 |
+|---|---|---|
+| single_candle | 单K形态 | 50 |
+| double_candle | 双K形态 | 21 |
+| triple_candle | 三K形态 | 14 |
+| trend_structure | 趋势结构 | 12 |
+| volume_price | 量价形态 | 16 |
+| compound | 复合形态 | 11 |
+| momentum | 动量因子 | 6 |
+
+运行时自动过滤掉 DataFrame 中不存在的字段。
+
+### Blueprint 注册
+
+在 `app/__init__.py` 中注册（Pattern A：url_prefix 在 Blueprint 构造函数中声明）：
+- API blueprint: `pattern_screen_api = Blueprint('pattern_screen_api', __name__, url_prefix='/api/pattern-screen')`，然后 `app.register_blueprint(pattern_screen_api)`
+- 页面 blueprint: `pattern_screen_bp = Blueprint('pattern_screen', __name__)`，路由 `@pattern_screen_bp.route('/pattern-screen/')`
+
+## 前端设计
+
+### 页面布局
+
+继承 `base.html`，左右分栏：
+
+- **左侧面板**（固定宽度 300px）：
+  - 顶部搜索框（按中文名过滤形态）
+  - 分组手风琴（可折叠），每组显示 checkbox + 中文名 + 命中数
+  - 底部"重置"和"筛选"按钮
+
+- **右侧内容区**：
+  - 统计栏（已选形态数 + 匹配结果数）
+  - 结果表格（代码、名称、行业、涨跌幅、现价、成交额、总市值、换手率、量比）
+  - 表头可点击排序
+  - 底部分页控件
+
+### 交互流程
+
+1. 页面加载 → GET `/api/pattern-screen/groups` → 渲染分组面板
+2. 勾选形态 → 点击"筛选" → POST `/api/pattern-screen/screen` → 更新表格
+3. 点击表头 → 重新筛选（带 sort_by 参数）
+4. 点击页码 → 重新筛选（带 offset 参数）
+5. 点击"重置" → 清空勾选，显示全部
+
+### 技术栈
+
+- Bootstrap 5（与项目一致）
+- 原生 JavaScript（无额外框架）
+- 项目已有 CSS 主题 (`financial-theme.css`)
+
+## 筛选逻辑
+
+- **纯 AND**：所有勾选的形态字段必须同时为 1
+- 无勾选时（`patterns` 为空或省略）返回全部股票（仅排序和分页）
+- 不在分组定义中的 DataFrame 列（如 `consec_up_days`）仍可作为 `sort_by` 使用，但不显示在筛选面板中
+
+## 导航集成
+
+在 `base.html` 导航栏中添加"形态选股"链接，使用 `url_for('pattern_screen.index')` 生成 URL。
+
+## 不做的事情
+
+- 不做形态回测（与现有 backtest 功能不同）
+- 不做自然语言查询
+- 不做形态组合的 AND/OR 混合逻辑（纯 AND）
+- 不修改现有选股模块
diff --git a/docs/superpowers/specs/2026-06-07-sector-heatmap-design.md b/docs/superpowers/specs/2026-06-07-sector-heatmap-design.md
new file mode 100644
index 000000000..2076a4616
--- /dev/null
+++ b/docs/superpowers/specs/2026-06-07-sector-heatmap-design.md
@@ -0,0 +1,129 @@
+# 板块热力图功能设计
+
+> Date: 2026-06-07
+> Status: Approved
+
+## 概述
+
+新增 Treemap 热力图页面，展示 A 股 110 个行业板块的涨跌分布。板块面积 = 总市值占比，颜色 = 加权涨跌幅（红涨绿跌）。点击板块后页面内展开个股明细表格。
+
+数据源：`data/data.parquet`（每日由数据下载任务更新，~5500 行，276 列）。
+
+## 数据层
+
+### 数据源
+
+`data/data.parquet` — 每日全市场宽表快照，包含 industry、pct_chg、total_mv、net_mf_amount 等字段。
+
+### 板块聚合逻辑（后端 Python）
+
+1. 读取 parquet，过滤掉 `close == 0` 的停牌/退市股
+2. 按 `industry` 分组，计算每个板块：
+   - `avg_pct_chg`：加权平均涨跌幅（权重 = `total_mv`）
+   - `total_mv`：板块总市值
+   - `stock_count`：个股数量
+   - `up_count` / `down_count`：涨/跌家数
+   - `net_mf_amount`：板块主力净流入额合计
+3. 按板块总市值降序排列
+
+### 个股明细
+
+前端 JS 从全量数据中按 `industry` 过滤，展示字段：
+
+| 字段 | 说明 |
+|---|---|
+| `name` | 股票简称 |
+| `pct_chg` | 涨跌幅 % |
+| `close` | 收盘价 |
+| `total_mv` | 总市值（亿元） |
+| `net_mf_amount` | 主力净流入（万元） |
+| `turnover_rate` | 换手率 % |
+
+按 `pct_chg` 降序排列。
+
+### 数据传递
+
+后端将两份数据注入 Jinja2 模板：`sectors_json`（板块聚合）和 `stocks_json`（全量个股）。前端 JS 直接使用，无需额外 API 调用。
+
+## 前端页面结构
+
+### 页面布局（从上到下）
+
+1. **页面标题栏**：板块热力图 · {trade_date}，含排序选项和图例说明
+2. **Treemap 主区域**：ECharts treemap，每个矩形 = 一个行业板块
+3. **个股展开区域**：点击板块后动态展开/收起的表格
+
+### 交互行为
+
+- **点击板块矩形**：treemap 下方展开/切换个股表格（带折叠动画），再次点击同板块则收起
+- **悬停矩形**：ECharts tooltip 显示板块详情（涨跌家数、净流入、市值排名）
+- **表格行点击**：不处理（保持简洁）
+
+### 配色（A 股惯例）
+
+- 涨：红色渐变 `#c0392b`（大涨）→ `#f5b7b1`（微涨）
+- 跌：绿色渐变 `#27ae60`（大跌）→ `#a9dfbf`（微跌）
+- 平盘：`#bdc3c7` 灰色
+
+### ECharts Treemap 配置
+
+- `visualMap` 连续型，范围取当日实际涨跌幅 min/max
+- `leafDepth = 1`（只展示板块层级）
+- `roam: false`（禁止缩放平移）
+
+## 文件结构与集成
+
+### 新增文件（3 个）
+
+| 文件 | 用途 |
+|---|---|
+| `app/services/heatmap_service.py` | 数据聚合服务：读 parquet → 板块汇总 + 个股列表 |
+| `app/templates/heatmap.html` | 页面模板：ECharts treemap + 个股展开表格 |
+| `app/routes/heatmap.py` | 路由：`GET /heatmap` |
+
+### 修改文件（2 个）
+
+| 文件 | 改动 |
+|---|---|
+| `app/__init__.py` | 注册新 blueprint `heatmap_bp` |
+| `app/templates/base.html` | 导航栏追加"板块热力图"链接 |
+
+### HeatmapService 接口
+
+```python
+class HeatmapService:
+    def get_heatmap_data(self) -> tuple[list[dict], list[dict]]:
+        """返回 (sectors_json, stocks_json)"""
+```
+
+### Route
+
+```python
+heatmap_bp = Blueprint('heatmap', __name__)
+
+@heatmap_bp.route('/heatmap')
+def heatmap_page():
+    sectors, stocks = HeatmapService().get_heatmap_data()
+    return render_template('heatmap.html',
+                           sectors_json=sectors,
+                           stocks_json=stocks,
+                           trade_date=sectors[0]['trade_date'])
+```
+
+### 前端 JS（内联，~150 行）
+
+- `initTreemap(sectors)` — 初始化 ECharts treemap
+- `onSectorClick(industry)` — 过滤 stocks 数据，渲染/切换下方表格
+- 表格用原生 HTML `<table>` + Bootstrap 样式
+
+### 依赖
+
+无新依赖。使用项目已有的 pandas、ECharts 5.4.3、Bootstrap 5.1.3、Jinja2。
+
+## 验收标准
+
+1. 访问 `/heatmap` 能看到板块 Treemap 热力图，颜色和面积正确
+2. 点击任意板块，下方展开该板块个股表格；再次点击收起
+3. 点击不同板块，表格切换为对应板块的个股
+4. 页面顶部导航有"板块热力图"入口
+5. `data/data.parquet` 更新后刷新页面即可看到新数据
diff --git a/run.py b/run.py
index 0fbcd2c7d..c5a7c0d23 100644
--- a/run.py
+++ b/run.py
@@ -44,4 +44,4 @@ def inspect_runtime_health(flask_app):
         debug=False,
         use_reloader=False,
         allow_unsafe_werkzeug=True
-    ) 
+    )
diff --git a/tests/api/test_pattern_screen_api.py b/tests/api/test_pattern_screen_api.py
new file mode 100644
index 000000000..a4f76a0f2
--- /dev/null
+++ b/tests/api/test_pattern_screen_api.py
@@ -0,0 +1,98 @@
+"""Pattern screen API contract tests."""
+import pytest
+from unittest.mock import patch, MagicMock
+
+
+@pytest.fixture
+def client(app):
+    return app.test_client()
+
+
+@pytest.fixture
+def mock_service():
+    """Mock the singleton getter for API tests."""
+    svc = MagicMock()
+    svc.get_groups.return_value = [
+        {
+            "id": "trend_structure",
+            "label": "趋势结构",
+            "fields": [
+                {"key": "pattern_golden_cross", "label": "均线金叉", "count": 304},
+                {"key": "pattern_ma_bull", "label": "均线多头排列", "count": 259},
+            ],
+        }
+    ]
+    svc.screen.return_value = {
+        "total": 2,
+        "offset": 0,
+        "limit": 50,
+        "trade_date": "20260605",
+        "rows": [
+            {"ts_code": "000001.SZ", "name": "平安银行", "industry": "银行",
+             "pct_chg": 1.5, "close": 12.5, "amount": 1000000,
+             "total_mv": 2400000, "turnover_rate": 1.2, "vol_ratio_5": 1.8},
+            {"ts_code": "000003.SZ", "name": "国农科技", "industry": "综合",
+             "pct_chg": 3.0, "close": 25.0, "amount": 200000,
+             "total_mv": 600000, "turnover_rate": 2.5, "vol_ratio_5": 2.1},
+        ],
+    }
+
+    patcher = patch(
+        'app.services.pattern_screen_service.get_pattern_screen_service',
+        return_value=svc,
+    )
+    patcher.start()
+    yield svc
+    patcher.stop()
+
+
+class TestGetGroups:
+    def test_returns_200(self, client, mock_service):
+        resp = client.get('/api/pattern-screen/groups')
+        assert resp.status_code == 200
+        data = resp.get_json()
+        assert data['code'] == 200
+        assert isinstance(data['data'], list)
+
+    def test_group_structure(self, client, mock_service):
+        resp = client.get('/api/pattern-screen/groups')
+        data = resp.get_json()['data']
+        g = data[0]
+        assert 'id' in g
+        assert 'label' in g
+        assert 'fields' in g
+
+
+class TestScreen:
+    def test_returns_200(self, client, mock_service):
+        resp = client.post('/api/pattern-screen/screen',
+                           json={'patterns': ['pattern_golden_cross']})
+        assert resp.status_code == 200
+        data = resp.get_json()
+        assert data['code'] == 200
+        assert data['data']['total'] == 2
+
+    def test_empty_patterns(self, client, mock_service):
+        resp = client.post('/api/pattern-screen/screen', json={})
+        assert resp.status_code == 200
+
+    def test_invalid_sort_by_returns_400(self, client, mock_service):
+        mock_service.screen.side_effect = ValueError("sort_by 'bad' not in whitelist")
+        resp = client.post('/api/pattern-screen/screen',
+                           json={'sort_by': 'bad'})
+        assert resp.status_code == 400
+        assert resp.get_json()['code'] == 400
+
+    def test_response_format(self, client, mock_service):
+        resp = client.post('/api/pattern-screen/screen',
+                           json={'patterns': ['pattern_golden_cross']})
+        data = resp.get_json()['data']
+        assert 'total' in data
+        assert 'offset' in data
+        assert 'limit' in data
+        assert 'trade_date' in data
+        assert 'rows' in data
+        row = data['rows'][0]
+        for col in ['ts_code', 'name', 'industry', 'pct_chg', 'close',
+                     'amount', 'total_mv', 'turnover_rate', 'vol_ratio_5']:
+            assert col in row
diff --git a/tests/services/test_heatmap_service.py b/tests/services/test_heatmap_service.py
new file mode 100644
index 000000000..f0b37c6bd
--- /dev/null
+++ b/tests/services/test_heatmap_service.py
@@ -0,0 +1,93 @@
+"""Tests for HeatmapService — sector aggregation logic."""
+import json
+import pytest
+from unittest.mock import patch, MagicMock
+import pandas as pd
+import numpy as np
+
+
+@pytest.fixture
+def sample_df():
+    """Minimal DataFrame matching data/data.parquet schema."""
+    return pd.DataFrame({
+        'ts_code': ['000001.SZ', '000002.SZ', '000003.SZ', '000004.SZ', '000005.SZ'],
+        'name': ['平安银行', '万科Ａ', '测试银行', '测试地产', '停牌股'],
+        'industry': ['银行', '全国地产', '银行', '全国地产', '全国地产'],
+        'pct_chg': [1.5, -2.0, 0.5, -1.0, 0.0],
+        'close': [11.0, 3.5, 5.0, 8.0, 0.0],
+        'total_mv': [21300, 3900, 5000, 2000, 100],
+        'net_mf_amount': [14000, -5000, 3000, -2000, 0],
+        'turnover_rate': [0.5, 1.7, 1.0, 2.0, 0.0],
+        'trade_date': ['20260605'] * 5,
+    })
+
+
+@pytest.fixture
+def service():
+    from app.services.heatmap_service import HeatmapService
+    return HeatmapService()
+
+
+class TestGetHeatmapData:
+    """Test HeatmapService.get_heatmap_data()."""
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_returns_two_lists(self, mock_read, service, sample_df):
+        mock_read.return_value = sample_df
+        sectors, stocks = service.get_heatmap_data()
+        assert isinstance(sectors, list)
+        assert isinstance(stocks, list)
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_filters_suspended_stocks(self, mock_read, service, sample_df):
+        """Stocks with close == 0 should be excluded."""
+        mock_read.return_value = sample_df
+        sectors, stocks = service.get_heatmap_data()
+        stock_names = [s['name'] for s in stocks]
+        assert '停牌股' not in stock_names
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_sector_count(self, mock_read, service, sample_df):
+        mock_read.return_value = sample_df
+        sectors, _ = service.get_heatmap_data()
+        industry_names = [s['name'] for s in sectors]
+        assert len(industry_names) == 2  # 银行 + 全国地产
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_sector_weighted_pct_chg(self, mock_read, service, sample_df):
+        """avg_pct_chg should be market-cap weighted average."""
+        mock_read.return_value = sample_df
+        sectors, _ = service.get_heatmap_data()
+        bank = next(s for s in sectors if s['name'] == '银行')
+        # Weighted: (1.5*21300 + 0.5*5000) / (21300+5000) = 34450/26300 ≈ 1.3103
+        assert abs(bank['avg_pct_chg'] - 1.3103) < 0.01
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_sector_stock_count(self, mock_read, service, sample_df):
+        mock_read.return_value = sample_df
+        sectors, _ = service.get_heatmap_data()
+        bank = next(s for s in sectors if s['name'] == '银行')
+        assert bank['stock_count'] == 2  # 2 银行 after filtering
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_sector_up_down_count(self, mock_read, service, sample_df):
+        mock_read.return_value = sample_df
+        sectors, _ = service.get_heatmap_data()
+        realestate = next(s for s in sectors if s['name'] == '全国地产')
+        assert realestate['down_count'] == 2
+        assert realestate['up_count'] == 0
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_stocks_have_required_fields(self, mock_read, service, sample_df):
+        mock_read.return_value = sample_df
+        _, stocks = service.get_heatmap_data()
+        required = {'name', 'ts_code', 'pct_chg', 'close', 'total_mv',
+                    'net_mf_amount', 'turnover_rate', 'industry'}
+        for s in stocks:
+            assert required.issubset(s.keys()), f"Missing keys: {required - s.keys()}"
+
+    @patch('app.services.heatmap_service.pd.read_parquet')
+    def test_trade_date_returned(self, mock_read, service, sample_df):
+        mock_read.return_value = sample_df
+        sectors, _ = service.get_heatmap_data()
+        assert sectors[0]['trade_date'] == '20260605'
diff --git a/tests/services/test_pattern_screen_service.py b/tests/services/test_pattern_screen_service.py
new file mode 100644
index 000000000..5d284b1e2
--- /dev/null
+++ b/tests/services/test_pattern_screen_service.py
@@ -0,0 +1,140 @@
+"""PatternScreenService unit tests."""
+import pytest
+from unittest.mock import patch, MagicMock
+import pandas as pd
+import numpy as np
+
+
+@pytest.fixture
+def sample_df():
+    """Minimal DataFrame mimicking data/data.parquet structure."""
+    return pd.DataFrame({
+        'ts_code': ['000001.SZ', '000002.SZ', '000003.SZ', '000004.SZ'],
+        'name': ['平安银行', '万科A', '国农科技', '国信证券'],
+        'industry': ['银行', '房地产', '综合', '证券'],
+        'trade_date': ['20260605'] * 4,
+        'pct_chg': [1.5, -0.5, 3.0, 0.0],
+        'close': [12.5, 8.0, 25.0, 15.0],
+        'amount': [1000000, 500000, 200000, 300000],
+        'total_mv': [2400000, 1200000, 600000, 900000],
+        'turnover_rate': [1.2, 0.8, 2.5, 0.5],
+        'vol_ratio_5': [1.8, 0.6, 2.1, 0.9],
+        'consec_up_days': [2, 0, 3, 1],
+        'pattern_golden_cross': [1, 0, 1, 0],
+        'pattern_ma_bull': [0, 0, 1, 1],
+        'pattern_bull_candle': [1, 0, 1, 0],
+        'pattern_bear_candle': [0, 1, 0, 1],
+        'break_high_20': [1, 0, 1, 0],
+    })
+
+
+@pytest.fixture
+def service(sample_df):
+    """Service with injected test DataFrame."""
+    with patch('app.services.pattern_screen_service.PatternScreenService._load_df', return_value=sample_df):
+        from app.services.pattern_screen_service import PatternScreenService
+        svc = PatternScreenService()
+        svc._df = sample_df
+        return svc
+
+
+class TestGetGroups:
+    def test_returns_groups_with_hits(self, service):
+        groups = service.get_groups()
+        assert isinstance(groups, list)
+        assert len(groups) > 0
+        g = groups[0]
+        assert 'id' in g
+        assert 'label' in g
+        assert 'fields' in g
+        for f in g['fields']:
+            assert 'key' in f
+            assert 'label' in f
+            assert 'count' in f
+            assert isinstance(f['count'], int)
+
+    def test_fields_not_in_dataframe_are_excluded(self, service):
+        groups = service.get_groups()
+        df_cols = set(service._df.columns)
+        for g in groups:
+            for f in g['fields']:
+                assert f['key'] in df_cols
+
+
+class TestScreen:
+    def test_no_patterns_returns_all(self, service, sample_df):
+        result = service.screen(patterns=[])
+        assert result['total'] == len(sample_df)
+        assert len(result['rows']) == len(sample_df)
+
+    def test_single_pattern_filters(self, service):
+        result = service.screen(patterns=['pattern_golden_cross'])
+        assert result['total'] == 2
+        codes = [r['ts_code'] for r in result['rows']]
+        assert '000001.SZ' in codes
+        assert '000003.SZ' in codes
+
+    def test_multiple_patterns_and_logic(self, service):
+        result = service.screen(patterns=['pattern_golden_cross', 'pattern_ma_bull'])
+        assert result['total'] == 1
+        assert result['rows'][0]['ts_code'] == '000003.SZ'
+
+    def test_sort_desc(self, service):
+        result = service.screen(patterns=[], sort_by='pct_chg', order='desc')
+        pcts = [r['pct_chg'] for r in result['rows']]
+        assert pcts == sorted(pcts, reverse=True)
+
+    def test_sort_asc(self, service):
+        result = service.screen(patterns=[], sort_by='pct_chg', order='asc')
+        pcts = [r['pct_chg'] for r in result['rows']]
+        assert pcts == sorted(pcts)
+
+    def test_pagination(self, service):
+        result = service.screen(patterns=[], limit=2, offset=0)
+        assert len(result['rows']) == 2
+        assert result['limit'] == 2
+        assert result['offset'] == 0
+        assert result['total'] == 4
+
+    def test_offset_beyond_results(self, service):
+        result = service.screen(patterns=[], offset=100)
+        assert result['total'] == 4
+        assert len(result['rows']) == 0
+
+    def test_invalid_sort_by_raises(self, service):
+        with pytest.raises(ValueError, match='sort_by'):
+            service.screen(patterns=[], sort_by='invalid_column')
+
+    def test_invalid_order_raises(self, service):
+        with pytest.raises(ValueError, match='order'):
+            service.screen(patterns=[], order='random')
+
+    def test_invalid_pattern_key_raises(self, service):
+        with pytest.raises(ValueError, match='pattern.*not found'):
+            service.screen(patterns=['nonexistent_pattern'])
+
+    def test_limit_capped_at_500(self, service):
+        result = service.screen(patterns=[], limit=9999)
+        assert result['limit'] == 500
+
+    def test_includes_trade_date(self, service):
+        result = service.screen(patterns=[])
+        assert result['trade_date'] == '20260605'
+
+    def test_nan_converted_to_none(self, sample_df):
+        sample_df.loc[0, 'industry'] = np.nan
+        with patch('app.services.pattern_screen_service.PatternScreenService._load_df', return_value=sample_df):
+            from app.services.pattern_screen_service import PatternScreenService
+            svc = PatternScreenService()
+            svc._df = sample_df
+            result = svc.screen(patterns=[])
+            # Find the row with ts_code='000001.SZ' (index 0 in original df)
+            row = next(r for r in result['rows'] if r['ts_code'] == '000001.SZ')
+            assert row['industry'] is None
+
+
+class TestInvalidateCache:
+    def test_clears_df(self, service):
+        assert service._df is not None
+        service.invalidate_cache()
+        assert service._df is None