Update app.py

Fix the label issues so now it actually shows the "thinking" that the model is doing instead of just printing the answers

app.py

@@ -1,4 +1,5 @@
 import os
+import re
 os.environ["GRADIO_ENABLE_SSR"] = "0"
 
 import gradio as gr
@@ -28,6 +29,19 @@ def format_rules(rules):
     formatted_rules += "</rules>\n"
     return formatted_rules
 
+def format_output(text):
+    think_match = re.search(r"<think>(.*?)</think>", text, flags=re.DOTALL)
+    answers = re.findall(r"<answer>(.*?)</answer>", text, flags=re.DOTALL)
+
+    display = ""
+    if think_match:
+        display += "Reasoning:\n" + think_match.group(1).strip() + "\n\n"
+    if answers:
+        display += "Answers:\n" + "\n".join([a.strip() for a in answers])
+    return display.strip() if display else text.strip()
+
+    
+
 def format_transcript(transcript):
     formatted_transcript = f"<transcript>\n{transcript}\n</transcript>\n"
     return formatted_transcript
@@ -134,10 +148,10 @@ class ModelWrapper:
         try:
             remainder = output_text.split("Brief explanation\n</explanation>")[-1]
             thinking_answer_text = remainder.split("</transcript>")[-1]
-            return thinking_answer_text
+            return format_output(thinking_answer_text)
         except:
             input_length = len(message)
-            return output_text[input_length:] if len(output_text) > input_length else "No response generated."
+            return format_output(output_text[input_length:]) if len(output_text) > input_length else "No response generated."
 
 # --- Model Cache ---
 LOADED_MODELS = {}