Update app.py

app.py

@@ -30,8 +30,8 @@ def add_new_eval(
     family_prediction_dataset,
 ):
     representation_name = model_name_textbox if revision_name_textbox == '' else revision_name_textbox
-    return None
     results = run_probe(benchmark_type, representation_name, human_file, skempi_file, similarity_tasks, function_prediction_aspect, function_prediction_dataset, family_prediction_dataset)
+    return results
 
     for benchmark_type in results:
         if benchmark_type == 'similarity':
@@ -97,6 +97,9 @@ with block:
                 outputs=data_component
             )
 
+            # Dropdown for benchmark type
+            benchmark_type_selector = gr.Dropdown(choices=list(benchmark_specific_metrics.keys()), label="Select Benchmark Type")
+
             # Dynamic selectors
             x_metric_selector = gr.Dropdown(choices=[], label="Select X-axis Metric", visible=False)
             y_metric_selector = gr.Dropdown(choices=[], label="Select Y-axis Metric", visible=False)
@@ -112,50 +115,6 @@ with block:
             plot_button = gr.Button("Plot")
             plot_output = gr.Image(label="Plot")
             
-            # Update metric selectors based on benchmark type
-            def update_metric_choices(benchmark_type):
-                if benchmark_type == 'similarity':
-                    # Show x and y metric selectors for similarity
-                    metric_names = benchmark_specific_metrics.get(benchmark_type, [])
-                    return (
-                        gr.update(choices=metric_names, value=metric_names[0], visible=True),
-                        gr.update(choices=metric_names, value=metric_names[1], visible=True),
-                        gr.update(visible=False), gr.update(visible=False), 
-                        gr.update(visible=False), gr.update(visible=False)
-                    )
-                elif benchmark_type == 'function':
-                    # Show aspect and dataset type selectors for function
-                    aspect_types = benchmark_specific_metrics[benchmark_type]['aspect_types']
-                    dataset_types = benchmark_specific_metrics[benchmark_type]['dataset_types']
-                    return (
-                        gr.update(visible=False), gr.update(visible=False),
-                        gr.update(choices=aspect_types, value=aspect_types[0], visible=True),
-                        gr.update(choices=dataset_types, value=dataset_types[0], visible=True),
-                        gr.update(visible=False), gr.update(visible=False)
-                    )
-                elif benchmark_type == 'family':
-                    # Show dataset and metric selectors for family
-                    datasets = benchmark_specific_metrics[benchmark_type]['datasets']
-                    metrics = benchmark_specific_metrics[benchmark_type]['metrics']
-                    return (
-                        gr.update(visible=False), gr.update(visible=False),
-                        gr.update(visible=False), gr.update(visible=False),
-                        gr.update(choices=datasets, value=datasets[0], visible=True),
-                        gr.update(choices=metrics, value=metrics[0], visible=True)
-                    )
-                elif benchmark_type == 'affinity':
-                    # Show single metric selector for affinity
-                    metrics = benchmark_specific_metrics[benchmark_type]
-                    return (
-                        gr.update(visible=False), gr.update(visible=False),
-                        gr.update(visible=False), gr.update(visible=False),
-                        gr.update(visible=False), gr.update(choices=metrics, value=metrics[0], visible=True)
-                    )
-                return gr.update(visible=False), gr.update(visible=False), gr.update(visible=False), gr.update(visible=False), gr.update(visible=False), gr.update(visible=False)
-            
-            # Dropdown for benchmark type
-            benchmark_type_selector = gr.Dropdown(choices=list(benchmark_specific_metrics.keys()), label="Select Benchmark Type")
-            
             # Update selectors when benchmark type changes
             benchmark_type_selector.change(
                 update_metric_choices,