fix: 缓存加载后Date列转datetime，修复类型比较错误

从CSV缓存加载时Date列是字符串，需要转为datetime才能与Timestamp比较

tradingagents/dataflows/stockstats_utils.py

@@ -296,6 +296,10 @@ def load_ohlcv(symbol: str, curr_date: str) -> pd.DataFrame:
     if os.path.exists(data_file):
         cached = pd.read_csv(data_file, on_bad_lines="skip", encoding="utf-8")
         if not cached.empty and "Close" in cached.columns:
+            # Ensure Date column is datetime after loading from cache
+            if "Date" in cached.columns:
+                cached["Date"] = pd.to_datetime(cached["Date"], errors="coerce")
+                cached = cached.dropna(subset=["Date"])
             data = cached
 
     if data is None: