feat: fetch_etf_with_nav 返回历史溢价率序列

修改内容：
1. universal_fetcher.py
   - fetch_etf_with_nav 返回三值：(price_df, nav_df, premium_series)
   - 新增 _calculate_premium_series 方法：计算每一天的溢价率
   - 溢价率 = (ETF收盘价 - ETF净值) / ETF净值
   - 净值用ffill对齐价格日期（处理T+1延迟）

2. flask_server.py
   - /api/v1/etf/nav 端点返回历史溢价率序列
   - 添加 premium_series 字段：[{date, premium}]
   - 添加 latest_premium: 最新溢价率
   - 添加 premium_stats: 统计数据（mean/std/min/max/median）

测试结果（513100.SH 纳指100 ETF）：
- 价格数据: 8条
- 净值数据: 8条
- 溢价率序列: 8条
- 最新溢价率: 0.1500%
- 溢价率均值: 1.1433%
- 溢价率范围: 0.15% ~ 1.69%

datasource/flask_server.py

@@ -484,11 +484,11 @@ def get_etf_nav():
             "hint": "Only A股ETF (codes starting with 51/52/15/16) supported",
         }), 400
     
-    # 获取净值
+    # 获取净值和溢价率
     f = get_fetcher()
     try:
         with f:
-            price_df, nav_df = f.fetch_etf_with_nav(code, start, end)
+            price_df, nav_df, premium_series = f.fetch_etf_with_nav(code, start, end)
         
         result = {
             "code": code,
@@ -496,14 +496,30 @@ def get_etf_nav():
             "nav": dataframe_to_json(nav_df) if nav_df else {"data": [], "count": 0},
         }
         
-        # 计算最新溢价率
-        if nav_df is not None and len(nav_df) > 0 and price_df is not None and len(price_df) > 0:
-            latest_nav = nav_df['nav'].iloc[-1]
-            latest_price = price_df['close'].iloc[-1]
-            if latest_nav > 0:
-                premium = (latest_price - latest_nav) / latest_nav
-                result['premium_rate'] = premium
-                result['premium_date'] = nav_df.index[-1].strftime('%Y-%m-%d')
+        # 添加历史溢价率序列
+        if premium_series is not None and len(premium_series) > 0:
+            # 转换为日期-溢价率列表
+            premium_data = [
+                {"date": date.strftime('%Y-%m-%d'), "premium": round(premium, 6)}
+                for date, premium in premium_series.items()
+            ]
+            result['premium_series'] = premium_data
+            
+            # 最新溢价率
+            latest_premium = premium_series.iloc[-1]
+            latest_date = premium_series.index[-1].strftime('%Y-%m-%d')
+            result['latest_premium'] = round(latest_premium, 6)
+            result['premium_date'] = latest_date
+            
+            # 溢价率统计
+            result['premium_stats'] = {
+                "mean": round(premium_series.mean(), 6),
+                "std": round(premium_series.std(), 6),
+                "min": round(premium_series.min(), 6),
+                "max": round(premium_series.max(), 6),
+                "median": round(premium_series.median(), 6),
+            }
+        
         
         return jsonify(result)