Update app.py

app.py

@@ -164,7 +164,6 @@ def update_leaderboard_pro(results):
 
 def load_leaderboard():
     if not os.path.exists(LEADERBOARD_FILE) or os.stat(LEADERBOARD_FILE).st_size == 0:
-        # Create an empty DataFrame with all expected columns
         return pd.DataFrame({
             "Model Name": [],
             "Overall Accuracy": [],
@@ -177,7 +176,28 @@ def load_leaderboard():
     # Read the CSV file
     df = pd.read_csv(LEADERBOARD_FILE)
     
-    # Ensure all columns exist
+    # Fill empty Team Name
+    df['Team Name'] = df['Team Name'].fillna('Unknown Team')
+    
+    # Fill empty Timestamp
+    df['Timestamp'] = df['Timestamp'].fillna(datetime.now().strftime("%Y-%m-%d %H:%M:%S"))
+    
+    # If Timestamp is an empty string, replace with current timestamp
+    df.loc[df['Timestamp'] == '', 'Timestamp'] = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
+    
+    # Fill Correct Predictions based on Accuracy and Total Questions if possible
+    # Assuming Total Questions is typically 16186 and Accuracy is a percentage
+    df['Correct Predictions'] = df.apply(
+        lambda row: round(16186 * (row['Overall Accuracy'] / 100)) 
+        if pd.isna(row['Correct Predictions']) or row['Correct Predictions'] == '' 
+        else row['Correct Predictions'], 
+        axis=1
+    )
+    
+    # Remove duplicate entries, keeping the last occurrence
+    df = df.drop_duplicates(subset='Model Name', keep='last')
+    
+    # Ensure all expected columns are present
     expected_columns = [
         "Model Name", 
         "Overall Accuracy", 
@@ -187,21 +207,9 @@ def load_leaderboard():
         "Team Name"
     ]
     
-    # Add missing columns with default values
-    for col in expected_columns:
-        if col not in df.columns:
-            if col == "Timestamp":
-                df[col] = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
-            elif col == "Team Name":
-                df[col] = "Unknown Team"
-            else:
-                df[col] = None
-    
-    # Remove duplicate entries based on Model Name
-    df = df.drop_duplicates(subset="Model Name", keep='last')
-    
-    # Reorder columns to match expected structure
+    # Reorder columns and save
     df = df[expected_columns]
+    df.to_csv(LEADERBOARD_FILE, index=False)
     
     return df