Update app.py

app.py

@@ -28,7 +28,16 @@ def add_new_eval(
 ):
     representation_name = model_name_textbox if revision_name_textbox == '' else revision_name_textbox
     results = run_probe(benchmark_type, representation_name, human_file, skempi_file, similarity_tasks, function_prediction_aspect, function_prediction_dataset, family_prediction_dataset)
-    return None
+
+    for benchmark_type in results:
+        if benchmark_type == 'similarity':
+            save_similarity_output(results['similarity'], representation_name)
+        elif benchmark_type == 'function':
+            save_function_output(results['function'], representation_name)
+        elif benchmark_type == 'family':
+            save_family_output(results['family'], representation_name)
+        elif benchmark_type == "affinity":
+            save_affinity_output(results['affinity', representation_name])
 
 # Function to update leaderboard dynamically based on user selection
 def update_leaderboard(selected_methods, selected_metrics):
@@ -94,7 +103,7 @@ with block:
             method_selector = gr.CheckboxGroup(choices=method_names, label="Select methods to visualize", interactive=True, value=method_names)
             
             # Button to draw the plot for the selected benchmark
-            plot_button = gr.Button("Plot Visualization")
+            plot_button = gr.Button("Plot")
             plot_output = gr.Image(label="Plot")
 
             # Update metric selectors when benchmark type is chosen