Refactor get_leaderboard_df to handle multiple subsets and improve column selection and hiding

src/populate.py

@@ -12,9 +12,9 @@ from datasets import load_dataset
 def get_leaderboard_df(results_repo, results_path):
     subsets = load_dataset(results_repo, split='all', cache_dir=results_path)
     subset_dfs = {}
-    print("Subsets:", subsets.keys())
-    for subset_name in subsets.keys():
-        subset_df = subsets[subset_name].to_pandas()
+    print("Subsets:", subsets)
+    for subset_name, subset_data in subsets.items():
+        subset_df = subset_data.to_pandas()
         subset_df = subset_df.sort_values(by="Total", ascending=False)
         subset_df = subset_df.round(2)  # Round all numeric columns to two decimal places
         print(subset_df)