debug

app.py

@@ -405,9 +405,9 @@ 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[:-1])
+        models.add(model)
         print(f"benchmark is {benchmark} and model is {model}")
     
     # Create selection groups

src/display/utils.py

@@ -119,27 +119,37 @@ AutoEvalColumnMultimodal = make_dataclass("AutoEvalColumnMultimodal", auto_eval_
 
 ##############################################################################################################
 # Version 3
-auto_eval_column_dict_mib_subgraph = []
 
-# Method name column
-auto_eval_column_dict_mib_subgraph.append(["method", ColumnContent, ColumnContent("Method", "markdown", True, never_hidden=True)])
+# First, let's create field names that are valid Python identifiers
+auto_eval_column_dict_mib_subgraph = []
 
-# Add Task and Model columns for filtering
-auto_eval_column_dict_mib_subgraph.append(["Task", ColumnContent, ColumnContent("Task", "str", True)])
-auto_eval_column_dict_mib_subgraph.append(["Model", ColumnContent, ColumnContent("Model", "str", True)])
+# Method name column (always present)
+auto_eval_column_dict_mib_subgraph.append(
+    ["method", ColumnContent, ColumnContent("Method", "markdown", True, never_hidden=True)]
+)
 
-# For each task and model combination
+# Add columns for each task-model combination
 for task in TasksMib_Subgraph:
     for model in task.value.models:
-        col_name = f"{task.value.benchmark}({model})"
+        # Create a valid field name by using underscores
+        field_name = f"{task.value.benchmark}-{model}"
+        # The display name can be more descriptive
+        display_name = f"{task.value.benchmark}({model})"
+        
         auto_eval_column_dict_mib_subgraph.append([
-            col_name, 
-            ColumnContent, 
-            ColumnContent(col_name, "number", True)
+            field_name,  # This must be a valid Python identifier
+            ColumnContent,
+            ColumnContent(display_name, "number", True)
         ])
 
-# Average column
-auto_eval_column_dict_mib_subgraph.append(["average", ColumnContent, ColumnContent("Average", "number", True)])
+# Add the Average column
+auto_eval_column_dict_mib_subgraph.append(
+    ["average", ColumnContent, ColumnContent("Average", "number", True)]
+)
+
+print("Debug - Column field names:")
+for field in auto_eval_column_dict_mib_subgraph:
+    print(f"Field name: {field[0]}, Display name: {field[2].name}")