Update app.py

app.py

@@ -36,6 +36,18 @@ def add_new_eval(
     save,
 ):
     """Validate inputs, run evaluation and (optionally) save results."""
+    
+    # map the user‐facing labels back to the original codes
+    try:
+        benchmark_types_mapped = [benchmark_type_map[b] for b in benchmark_types]
+        similarity_tasks_mapped = [similarity_tasks_map[s] for s in similarity_tasks]
+        function_prediction_aspect_mapped = function_prediction_aspect_map[function_prediction_aspect]
+        family_prediction_dataset_mapped = [family_prediction_dataset_map[f] for f in family_prediction_dataset]
+    except KeyError as e:
+        gr.Warning(f"Unrecognized option: {e.args[0]}")
+        return -1
+
+    # validate inputs
     if any(task in benchmark_types for task in ['similarity', 'family', 'function']) and human_file is None:
         gr.Warning("Human representations are required for similarity, family, or function benchmarks!")
         return -1
@@ -172,10 +184,10 @@ with block:
             metric_names = leaderboard.columns.tolist(); metric_names.remove('Method')
 
             benchmark_metric_mapping = {
-                "similarity": [m for m in metric_names if m.startswith('sim_')],
-                "function":  [m for m in metric_names if m.startswith('func')],
-                "family":    [m for m in metric_names if m.startswith('fam_')],
-                "affinity":  [m for m in metric_names if m.startswith('aff_')],
+                "Semantic Similarity": [m for m in metric_names if m.startswith('sim_')],
+                "GO-Based Function Prediction":  [m for m in metric_names if m.startswith('func')],
+                "Target Protein Family Prediction":    [m for m in metric_names if m.startswith('fam_')],
+                "Protein Protein Binding Affinity":  [m for m in metric_names if m.startswith('aff_')],
             }
 
             leaderboard_method_selector = gr.CheckboxGroup(