Update app.py

app.py

@@ -467,13 +467,32 @@ def build_ui(evaluate_fn):
     def _ui_evaluate(description: str, code: str):
         verdict, metrics, issues, error = evaluate_fn(description, code)
         accuracy, summary = _split_metrics(metrics)
+    
+        if error:
+            # Error occurred — show only the error box, hide all result boxes
+            return (
+                gr.update(value="",       visible=False),  # verdict
+                gr.update(value="",       visible=False),  # accuracy
+                gr.update(value="",       visible=False),  # summary
+                gr.update(value="",       visible=False),  # issues
+                gr.update(value=error,    visible=True),   # error ← only this shows
+            )
+    
+        # No error — show all result boxes, hide the error box
         if "PASS" in verdict.upper():
             verdict_display = "✅  PASS"
         elif "FAIL" in verdict.upper():
             verdict_display = "❌  FAIL"
         else:
             verdict_display = verdict or ""
-        return verdict_display, accuracy, summary, issues, error
+    
+        return (
+            gr.update(value=verdict_display, visible=True),   # verdict
+            gr.update(value=accuracy,        visible=True),   # accuracy
+            gr.update(value=summary,         visible=True),   # summary
+            gr.update(value=issues,          visible=True),   # issues
+            gr.update(value="",              visible=False),  # error ← hidden when no error
+        )
 
     with gr.Blocks(title="Python Code Evaluator") as demo: