Handling missing data

app.py

@@ -1,11 +1,12 @@
-import os
+import time
+
 import pandas as pd
 import gradio as gr
 from gradio.themes.utils import sizes
 from gradio_leaderboard import Leaderboard
 from dotenv import load_dotenv
 
-load_dotenv()  # Load environment variables from .env file
+load_dotenv()  # Load environment variables from .env file (before imports)
 
 from about import ABOUT_INTRO, ABOUT_TEXT, FAQS, SUBMIT_INSTRUCTIONS
 from constants import (
@@ -52,7 +53,12 @@ def get_leaderboard_object(assay: str | None = None):
         filter_columns.append("property")
     # Bug: Can't leave search_columns empty because then it says "Column None not found in headers"
     # Note(Lood): Would be nice to make it clear that the Search Column is searching on model name
-    current_dataframe = pd.read_csv("debug-current-results.csv")
+    try:
+        current_dataframe = pd.read_csv("debug-current-results.csv")
+    except (pd.errors.EmptyDataError, pd.errors.ParserError):
+        # Return empty dataframe as fallback
+        print("Fallback: Returning empty dataframe as it doesn't exist yet")
+        current_dataframe = pd.DataFrame(columns=LEADERBOARD_COLUMNS_RENAME.values())
     lb = Leaderboard(
         value=format_leaderboard_table(df_results=current_dataframe, assay=assay),
         datatype=["str", "str", "str", "number", "str"],
@@ -68,16 +74,18 @@ def get_leaderboard_object(assay: str | None = None):
 
 
 def refresh_overall_leaderboard():
-    if not os.path.exists("debug-current-results.csv"):
-        fetch_hf_results()  # Hope this doesn't cause race conditions with the main fetch_hf_results() thread
-    current_dataframe = pd.read_csv("debug-current-results.csv")
+    # Handle race condition where file is being written while we read it
+    try:
+        current_dataframe = pd.read_csv("debug-current-results.csv")
+    except (pd.errors.EmptyDataError, pd.errors.ParserError):
+        # Return empty dataframe with required columns as fallback
+        print("Fallback: Returning empty dataframe as it doesn't exist yet")
+        current_dataframe = pd.DataFrame(columns=LEADERBOARD_COLUMNS_RENAME.values())
     return format_leaderboard_table(df_results=current_dataframe)
 
 
 # Initialize global dataframe
-fetch_hf_results()
-current_dataframe = pd.read_csv("debug-current-results.csv")
-
+time.sleep(2)  # Give the outer thread time to create the file at the start
 
 # Make font size bigger using gradio theme
 with gr.Blocks(theme=gr.themes.Default(text_size=sizes.text_lg)) as demo: