Upload folder using huggingface_hub

app.py

@@ -48,11 +48,8 @@ def predict_total_sales():
     # 🚨 Use original feature names here as used during training
     input_data = pd.DataFrame([sample])
 
-    # Predict log sales using full pipeline (preprocessing + model)
-    predicted_log_sales = model.predict(input_data)[0]
-
     # Convert log sales to actual sales
-    predicted_sale = round(float(np.exp(predicted_log_sales)), 2)
+    predicted_sale = round(model.predict(input_data).tolist()[0], 2) 
 
     return jsonify({'Predicted Product Store Sales Total': predicted_sale})
 
@@ -62,7 +59,7 @@ def predict_total_sales_batch():
     input_data = pd.read_csv(file)
 
     # Predict log sales using full pipeline
-    predicted_log_sales = model.predict(input_data)
+    predicted_log_sales = model.predict(input_data).tolist()
 
     # Convert log sales to actual sales
     predicted_sales = np.exp(predicted_log_sales)
@@ -70,6 +67,13 @@ def predict_total_sales_batch():
 
     return jsonify({'Predicted Sales for Each Row': predicted_sales})
 
+    # Create a dictionary of predictions with property IDs as keys
+    property_ids = input_data['id'].tolist()  # Assuming 'id' is the property ID column
+    output_dict = dict(zip(Product_Id, predicted_sales))  # Use actual sales total
+
+    # Return the predictions dictionary as a JSON response
+    return output_dict
+
 
 # Run the Flask application in debug mode if this script is executed directly
 if __name__ == '__main__':