From f7d6d5d136c66c62688cf38166df84de05dc3e45 Mon Sep 17 00:00:00 2001
From: "Bobby Abellana (aider)" <bobby@abellana.work>
Date: Thu, 6 Feb 2025 23:39:58 -0800
Subject: [PATCH] feat: Add price data validation and debug logging to stock
 data retrieval

---
 src/screener/t_sunnyband.py | 34 ++++++++++++++++++++++++++++++++++
 1 file changed, 34 insertions(+)

diff --git a/src/screener/t_sunnyband.py b/src/screener/t_sunnyband.py
index 4260068..dc9d126 100644
--- a/src/screener/t_sunnyband.py
+++ b/src/screener/t_sunnyband.py
@@ -34,6 +34,29 @@ def get_stock_data(ticker: str, start_date: datetime, end_date: datetime, interv
     """Fetch stock data from the database"""
     client = create_client()
     
+    # Add data validation
+    def validate_price_data(df: pd.DataFrame) -> pd.DataFrame:
+        if df.empty:
+            return df
+            
+        # Remove rows with suspicious values
+        df = df[
+            (df['open'] > 0) & 
+            (df['high'] > 0) & 
+            (df['low'] > 0) & 
+            (df['close'] > 0) &
+            (df['high'] >= df['low']) &
+            (df['high'] >= df['open']) &
+            (df['high'] >= df['close']) &
+            (df['low'] <= df['open']) &
+            (df['low'] <= df['close'])
+        ]
+        
+        # Sort by date and remove duplicates
+        df = df.sort_values('date').drop_duplicates(subset=['date'])
+        
+        return df
+    
     # Calculate proper date range (looking back from today)
     end_date = datetime.now()
     start_date = end_date - timedelta(days=60)  # 60 days of history
@@ -94,6 +117,17 @@ def get_stock_data(ticker: str, start_date: datetime, end_date: datetime, interv
             columns=['date', 'open', 'high', 'low', 'close', 'volume']
         )
         
+        # Add validation step
+        df = validate_price_data(df)
+        
+        if df.empty:
+            print(f"No valid data found for {ticker} after validation")
+            return df
+            
+        # Add debug output for last price
+        last_price = df.iloc[-1]
+        print(f"Debug: Last valid price for {ticker}: ${last_price['close']:.2f}")
+        
         if interval != "daily" and interval != "5min":
             # Resample to desired interval
             df.set_index('date', inplace=True)