Update app.py

app.py

@@ -137,7 +137,7 @@ def update_table(start_date, end_date, window, user_text):
     global df  # Use global variable
     #global df_2
     df = df_predictions.copy()
-    max_row_index = df['predicted_revenue'].idxmax()
+    max_row_index = df['Predicted Revenue'].idxmax()
     df = df.style.apply(lambda x: ['background-color: #d9f7be; color: black' if x.name == max_row_index else '' for i in x], axis=1)
     #df_2 = df_payments.copy() 
     return df#, df_2
@@ -182,9 +182,9 @@ with gr.Blocks(fill_height=True) as demo:
                 range_input = gr.Slider(3, 12, 6, step=1, label="Window / Moving Average Period")
                 gr.Markdown("Window = 3-period/months, the predictive model reacts more quickly to recent monthly payment fluctuations, but it may also include more noise.<br><br>Window = 12-period/months, the Forecast adjusts more slowly and is less sensitive to small fluctuations, making it more reliable, but also slower to react to sharp changes.")
         btn_update = gr.Button("Run Forecast")
-        gr.HTML(f"<h3>Predictions</h3>")
-        gr.Markdown("It's recommended to make the prediction once the selected month has passed.")
-        table_1 = gr.DataFrame(value=df, label="Predictions - consult 'Prediction Revenue' column:")
+        gr.HTML(f"<h2>Predictions</h2>")
+        gr.Markdown("It's recommended to make the predictions after the selected month has passed.")
+        table_1 = gr.DataFrame(value=df, label="Predictions -> consult 'Prediction Revenue' column:")
         #table_2 = gr.DataFrame(value=df_2, label="Forecast Inputs:")
         btn_update.click(fn=update_table, inputs=[start_input, prediction_input, range_input, user_text], outputs=[table_1#,table_2
                                                                                                                   ])