Update app.py

app.py

@@ -49,6 +49,7 @@ description and a Python code snippet, then produce a structured evaluation repo
 - Do NOT execute the code.
 - Base your evaluation solely on static code analysis and logical reasoning.
 - Keep feedback concise, clear, and actionable.
+- Pay close attention to operators (+, -, *, /) and ensure they match the description exactly.
 - Your response MUST follow the output format exactly.
 ### OUTPUT FORMAT
 Respond only with the following structure — no extra text before or after:
@@ -266,7 +267,6 @@ def validate_and_evaluate(description: str, code: str):
 # ---------------------------------------------------------------------------
 # MODULE 1 — UI Module
 # ---------------------------------------------------------------------------
-
 CUSTOM_CSS = """
 @import url('https://fonts.googleapis.com/css2?family=Space+Mono:wght@400;700&family=DM+Sans:ital,wght@0,300;0,400;0,500;0,600;1,400&display=swap');
 *, *::before, *::after { box-sizing: border-box; }
@@ -477,29 +477,31 @@ 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:
+            # Top row stays visible but empty — keeps full width
             return (
-                gr.update(value="",  visible=True),           # verdict   — empty but visible
-                gr.update(value="",  visible=True),           # accuracy  — empty but visible
-                gr.update(value="",  visible=True),           # summary   — empty but visible
-                gr.update(value="",  visible=False),          # issues    — hidden
-                gr.update(value=error, visible=True),         # error     — shown
+                gr.update(value="",    visible=True),        # verdict
+                gr.update(value="",    visible=True),        # accuracy
+                gr.update(value="",    visible=True),        # summary
+                gr.update(value="",    visible=False),       # issues — hidden
+                gr.update(value=error, visible=True),        # error  — shown
             )
-    
+
+        # Success — show results, hide error
         if "PASS" in verdict.upper():
             verdict_display = "✅  PASS"
         elif "FAIL" in verdict.upper():
             verdict_display = "❌  FAIL"
         else:
             verdict_display = verdict or ""
-    
+
         return (
-            gr.update(value=verdict_display, visible=True),
-            gr.update(value=accuracy,        visible=True),
-            gr.update(value=summary,         visible=True),
-            gr.update(value=issues,          visible=True),   # issues    — shown
-            gr.update(value="",              visible=False),  # error     — hidden
+            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 — shown
+            gr.update(value="",              visible=False), # error  — hidden
         )
 
     with gr.Blocks(title="Python Code Evaluator") as demo:
@@ -540,24 +542,32 @@ def build_ui(evaluate_fn):
 
         with gr.Row():
             eval_btn  = gr.Button("Evaluate", variant="primary",
-                                  elem_id="eval-btn", scale=0)
-            clear_btn = gr.Button("Clear", variant="secondary",
+                                  elem_id="eval-btn",  scale=0)
+            clear_btn = gr.Button("Clear",    variant="secondary",
                                   elem_id="clear-btn", scale=0)
 
         gr.HTML(RESULTS_HEADING_HTML)
 
+        # Top row — always visible to preserve full width
         with gr.Row(equal_height=True):
-            verdict_out  = gr.Textbox(label="Verdict",   interactive=False, elem_id="verdict-out",  scale=1)
-            accuracy_out = gr.Textbox(label="Accuracy",  interactive=False, elem_id="accuracy-out", scale=1)
-            summary_out  = gr.Textbox(label="Summary",   interactive=False, elem_id="summary-out",  scale=2, lines=3)
-
-        issues_out = gr.Textbox(label="Issues Detected", interactive=False, lines=5, elem_id="issues-out")
+            verdict_out  = gr.Textbox(label="Verdict",  interactive=False, elem_id="verdict-out",  scale=1)
+            accuracy_out = gr.Textbox(label="Accuracy", interactive=False, elem_id="accuracy-out", scale=1)
+            summary_out  = gr.Textbox(label="Summary",  interactive=False, elem_id="summary-out",  scale=2, lines=3)
+
+        # Issues and error swap in the same spot
+        issues_out = gr.Textbox(
+            label="Issues Detected",
+            interactive=False,
+            lines=5,
+            visible=True,
+            elem_id="issues-out",
+        )
 
         with gr.Row():
             error_out = gr.Textbox(
                 label="Status / Errors",
                 interactive=False,
-                visible=False,          # ← hidden by default, only shown on error
+                visible=False,          # hidden by default
                 elem_id="error-out",
                 scale=1,
             )