Upload folder using huggingface_hub

app.py

@@ -83,15 +83,30 @@ def predict_sales_batch():
     # Read the CSV file into a Pandas DataFrame
     input_data = pd.read_csv(file)
 
-    # Make predictions for all product details in the DataFrame
-    predicted_sales = model.predict(input_data).tolist()
-
-    # Create a dictionary of predictions with product IDs as keys
+    # Save Product_Id for mapping in output
     product_ids = input_data['Product_Id'].tolist()
-    output_dict = dict(zip(product_ids, predicted_sales))
-    
-    # Return the predictions dictionary as a JSON response
-    return jsonify(output_dict)
+
+    # Columns that your model does NOT need
+    drop_cols = [
+        'Product_Id',
+        'Store_Id',
+        'Store_Establishment_Year',
+        'Product_Store_Sales_Total'  # target column
+    ]
+
+     # Drop only columns that exist in the CSV
+    input_data = df.drop(columns=[c for c in drop_cols if c in input_data.columns])
+
+    # Now df contains ONLY the features used for training
+    predictions = model.predict(input_data)
+
+    # Convert NumPy types → Python float
+    predictions = [float(p) for p in predictions]
+
+    # Prepare output with Product_Id as keys
+    output = dict(zip(product_ids, predictions))
+
+    return jsonify(output)
 
 # Run the Flask application in debug mode if this script is executed directly
 if __name__ == '__main__':