Improve DataFrame handling in leaderboard functions; add debug print for empty DataFrame case

app.py

@@ -1,6 +1,5 @@
 import gradio as gr
 from gradio_leaderboard import Leaderboard, ColumnFilter, SelectColumns
-import pandas as pd
 from apscheduler.schedulers.background import BackgroundScheduler
 from huggingface_hub import snapshot_download
 
@@ -58,8 +57,8 @@ LEADERBOARD_DF = get_leaderboard_df(EVAL_RESULTS_PATH, EVAL_REQUESTS_PATH, COLS,
 ) = get_evaluation_queue_df(EVAL_REQUESTS_PATH, EVAL_COLS)
 
 def init_leaderboard(dataframe):
-    if dataframe is None or dataframe.empty:
-        raise ValueError("Leaderboard DataFrame is empty or None.")
+    # if dataframe is None or dataframe.empty:
+    #     raise ValueError("Leaderboard DataFrame is empty or None.")
     return Leaderboard(
         value=dataframe,
         datatype=[c.type for c in fields(AutoEvalColumn())],

src/populate.py

@@ -13,9 +13,11 @@ def get_leaderboard_df(results_path: str, requests_path: str, cols: list, benchm
     raw_data = get_raw_eval_results(results_path, requests_path)
     all_data_json = [v.to_dict() for v in raw_data]
     df = pd.DataFrame.from_records(all_data_json)
+    if df.empty:
+        print("No evaluation results found. Returning empty DataFrame with correct columns.")
+        return pd.DataFrame(columns=cols)
     df = df.sort_values(by=[AutoEvalColumn().average.name], ascending=False)
     df = df[cols].round(decimals=2)
-    # filter out if any of the benchmarks have not been produced
     df = df[has_no_nan_values(df, benchmark_cols)]
     return df