Update app.py

app.py

@@ -12,13 +12,8 @@ app = Flask(__name__, template_folder='templates')
 
 auth = HTTPBasicAuth()
 
-# Load the trained model
-try:
-    model = joblib.load('stroke_prediction_model.pkl')
-    print("Model loaded successfully.")
-except Exception as e:
-    print(f"Error loading model: {str(e)}")
-    model = None
+# Global variable to hold the model (lazy-loaded)
+model = None
 
 # Admin credentials for report access
 users = {
@@ -33,7 +28,7 @@ def verify_password(username, password):
 
 # Initialize SQLite database
 def init_db():
-    conn = sqlite3.connect('/tmp/submissions.db')  # Use /tmp for writable storage
+    conn = sqlite3.connect('/tmp/submissions.db')
     c = conn.cursor()
     c.execute('''
         CREATE TABLE IF NOT EXISTS submissions (
@@ -64,18 +59,22 @@ def home():
 
 @app.route('/predict', methods=['POST'])
 def predict():
+    global model
     if model is None:
-        return jsonify({'success': False, 'error': 'Model failed to load'}), 500
+        try:
+            model = joblib.load('stroke_prediction_model.pkl')
+            print("Model loaded successfully during first predict request.")
+        except Exception as e:
+            print(f"Error loading model: {str(e)}")
+            return jsonify({'success': False, 'error': 'Model failed to load'}), 500
 
     try:
-        # Get form data
         data = request.json
         if not data or 'input' not in data:
             return jsonify({'success': False, 'error': 'Invalid input data'}), 400
         input_data = data['input']
         print(f"Received input data: {input_data}")
 
-        # Prepare data for model
         features = {
             'age': float(input_data.get('age', 0)),
             'gender': input_data.get('gender', 'Unknown'),
@@ -90,14 +89,10 @@ def predict():
         }
         print(f"Processed features: {features}")
 
-        # Convert to DataFrame
         df = pd.DataFrame([features])
-
-        # One-hot encode categorical variables
         categorical_cols = ['gender', 'ever_married', 'Residence_type', 'work_type', 'smoking_status']
         df = pd.get_dummies(df, columns=categorical_cols, drop_first=True)
 
-        # Get expected columns from the model
         expected_columns = model.feature_names_in_ if hasattr(model, 'feature_names_in_') else [
             'age', 'hypertension', 'heart_disease', 'avg_glucose_level', 'bmi',
             'gender_Male', 'gender_Other', 'ever_married_Yes', 'Residence_type_Urban',
@@ -108,17 +103,14 @@ def predict():
         ]
         print(f"Expected columns: {expected_columns}")
 
-        # Align DataFrame with expected columns
         for col in expected_columns:
             if col not in df.columns:
                 df[col] = 0
         df = df[expected_columns]
 
-        # Make prediction
-        probability = model.predict_proba(df)[0][1] * 100  # Probability of stroke
+        probability = model.predict_proba(df)[0][1] * 100
         risk_level = 'High' if probability > 50 else 'Moderate' if probability > 25 else 'Low'
 
-        # Determine contributing factors
         contributing_factors = {
             'age': features['age'] > 45,
             'hypertension': features['hypertension'] == 1,
@@ -127,7 +119,6 @@ def predict():
             'smoking': features['smoking_status'] in ['smokes', 'formerly smoked']
         }
 
-        # Log submission to database
         conn = sqlite3.connect('/tmp/submissions.db')
         c = conn.cursor()
         c.execute('''
@@ -154,7 +145,6 @@ def predict():
         conn.commit()
         conn.close()
 
-        # Return prediction result
         return jsonify({
             'success': True,
             'probability': round(probability),
@@ -177,7 +167,6 @@ def admin_report():
         if df.empty:
             return "No submissions for today.", 200
         
-        # Prepare data for Excel
         df['hypertension'] = df['hypertension'].apply(lambda x: 'Yes' if x == 1 else 'No')
         df['heart_disease'] = df['heart_disease'].apply(lambda x: 'Yes' if x == 1 else 'No')
         df = df[[
@@ -191,7 +180,6 @@ def admin_report():
             'Probability (%)', 'Risk Level'
         ]
         
-        # Create Excel file in memory
         output = io.BytesIO()
         with pd.ExcelWriter(output, engine='xlsxwriter') as writer:
             df.to_excel(writer, sheet_name='Daily Report', index=False)
@@ -201,7 +189,6 @@ def admin_report():
                 worksheet.set_column(idx, idx, max_len)
         output.seek(0)
         
-        # Send file
         return send_file(
             output,
             download_name=f'report_{today}.xlsx',