debug

app.py

@@ -405,20 +405,21 @@ def init_leaderboard_mib_subgraph(dataframe, track):
     # Extract unique benchmarks and models from column names
     for col in result_columns:
         print(f"col is {col}")
-        benchmark, model = col.split('_')
+        benchmark, model = col.split('(')
         benchmarks.add(benchmark)
-        models.add(model)
+        models.add(model[:-1])
+        print(f"benchmark is {benchmark} and model is {model}")
     
     # Create selection groups
     benchmark_selections = {
         # For each benchmark, store which columns should be shown
-        benchmark: [col for col in result_columns if col.startswith(f"{benchmark}_")]
+        benchmark: [col for col in result_columns if col.startswith(f"{benchmark}(")]
         for benchmark in benchmarks
     }
     
     model_selections = {
         # For each model, store which columns should be shown
-        model: [col for col in result_columns if col.endswith(f"_{model}")]
+        model: [col for col in result_columns if col.endswith(f"({model})")]
         for model in models
     }
     

src/display/utils.py

@@ -131,7 +131,7 @@ auto_eval_column_dict_mib_subgraph.append(["Model", ColumnContent, ColumnContent
 # For each task and model combination
 for task in TasksMib_Subgraph:
     for model in task.value.models:
-        col_name = f"{task.value.benchmark}_{model}"
+        col_name = f"{task.value.benchmark}({model})"
         auto_eval_column_dict_mib_subgraph.append([
             col_name, 
             ColumnContent,