Update space

app.py

@@ -212,6 +212,93 @@ def sanitize_name(name):
     """Sanitize name for file system use"""
     return re.sub(r'[^a-zA-Z0-9]', '_', name)
 
+def scan_submissions_directory():
+    """
+    Scans the submissions directory and updates the leaderboard tables with all submitted results.
+    Returns a dictionary mapping split names to lists of submissions.
+    """
+    global df_synthesized_full, df_synthesized_10, df_human_generated
+    
+    try:
+        # Get submissions directory content from HuggingFace hub
+        submissions_content = hub_storage.list_repo_content("submissions")
+        if not submissions_content:
+            print("No submissions directory found or empty")
+            return
+            
+        # Track submissions for each split
+        submissions_by_split = {
+            'test': [],
+            'test-0.1': [],
+            'human_generated_eval': []
+        }
+        
+        # Iterate through team folders
+        for folder in submissions_content:
+            if not folder.endswith('/'):  # Skip files
+                continue
+                
+            try:
+                # Get latest.json to find most recent submission
+                latest_content = hub_storage.get_repo_content(f"{folder}latest.json")
+                if not latest_content:
+                    continue
+                    
+                latest_info = json.loads(latest_content)
+                if latest_info.get('status') != 'pending_review':  # Only include approved submissions
+                    timestamp = latest_info.get('latest_submission')
+                    if not timestamp:
+                        continue
+                        
+                    # Get metadata file for this submission
+                    metadata_path = f"{folder}metadata_{timestamp}.json"
+                    metadata_content = hub_storage.get_repo_content(metadata_path)
+                    if not metadata_content:
+                        continue
+                        
+                    submission_data = json.loads(metadata_content)
+                    split = submission_data.get('Split')
+                    if split in submissions_by_split:
+                        submissions_by_split[split].append(submission_data)
+                        
+                        # Update corresponding DataFrame
+                        update_leaderboard_data(submission_data)
+                        
+            except Exception as e:
+                print(f"Error processing folder {folder}: {str(e)}")
+                continue
+        
+        print("Leaderboard initialized with existing submissions:")
+        for split, submissions in submissions_by_split.items():
+            print(f"{split}: {len(submissions)} submissions")
+            
+        return submissions_by_split
+        
+    except Exception as e:
+        print(f"Error scanning submissions directory: {str(e)}")
+        return None
+
+def initialize_leaderboard():
+    """
+    Initialize the leaderboard with baseline results and submitted results.
+    """
+    global df_synthesized_full, df_synthesized_10, df_human_generated
+    
+    try:
+        # First, initialize with baseline results
+        df_synthesized_full = pd.DataFrame(data_synthesized_full)
+        df_synthesized_10 = pd.DataFrame(data_synthesized_10)
+        df_human_generated = pd.DataFrame(data_human_generated)
+        
+        # Then scan and add submitted results
+        scan_submissions_directory()
+        
+        print("Leaderboard initialization complete")
+        
+    except Exception as e:
+        print(f"Error initializing leaderboard: {str(e)}")
+
+
 def save_submission(submission_data, csv_file):
     """
     Save submission data and CSV file using model_name_team_name format
@@ -780,6 +867,9 @@ with gr.Blocks(css=css) as demo:
     submit_btn = gr.Button("Submit", variant="primary")
     result = gr.Textbox(label="Submission Status", interactive=False)
     
+    # Initialize leaderboard at startup
+    initialize_leaderboard()
+    
     # Set up event handlers
     model_type_filter.change(
         update_tables,
@@ -787,6 +877,13 @@ with gr.Blocks(css=css) as demo:
         outputs=all_dfs
     )
     
+    # Initial table update
+    demo.load(
+        update_tables,
+        inputs=[model_type_filter],
+        outputs=all_dfs
+    )
+    
     submit_btn.click(
         process_submission,
         inputs=[