Update space

app.py

@@ -220,9 +220,21 @@ def scan_submissions_directory():
     global df_synthesized_full, df_synthesized_10, df_human_generated
     
     try:
+        # Initialize HuggingFace API
+        api = HfApi()
+        
         # Get submissions directory content from HuggingFace hub
-        submissions_content = hub_storage.list_repo_content("submissions")
-        if not submissions_content:
+        try:
+            repo_files = api.list_files_info(
+                repo_id=REPO_ID,
+                repo_type="space",
+                paths=["submissions"]
+            )
+        except Exception as e:
+            print(f"Error listing repository contents: {str(e)}")
+            return
+            
+        if not repo_files:
             print("No submissions directory found or empty")
             return
             
@@ -233,39 +245,70 @@ def scan_submissions_directory():
             'human_generated_eval': []
         }
         
-        # Iterate through team folders
-        for folder in submissions_content:
-            if not folder.endswith('/'):  # Skip files
+        # Group files by team folders
+        folder_files = {}
+        for file_info in repo_files:
+            path = file_info.path
+            if not path.startswith('submissions/'):
+                continue
+                
+            parts = path.split('/')
+            if len(parts) < 3:  # submissions/folder_name/file
                 continue
                 
+            folder_name = parts[1]
+            if folder_name not in folder_files:
+                folder_files[folder_name] = []
+            folder_files[folder_name].append(path)
+        
+        # Process each team folder
+        for folder_name, files in folder_files.items():
             try:
-                # Get latest.json to find most recent submission
-                latest_content = hub_storage.get_repo_content(f"{folder}latest.json")
-                if not latest_content:
+                # Look for latest.json
+                latest_file = next((f for f in files if f.endswith('latest.json')), None)
+                if not latest_file:
                     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
-                        
+                # Read latest.json
+                try:
+                    latest_content = hub_storage.get_file_content(latest_file)
+                    latest_info = json.loads(latest_content)
+                except Exception as e:
+                    print(f"Error reading latest.json for {folder_name}: {str(e)}")
+                    continue
+                
+                if latest_info.get('status') != 'approved':
+                    continue
+                    
+                timestamp = latest_info.get('latest_submission')
+                if not timestamp:
+                    continue
+                    
+                # Find corresponding metadata file
+                metadata_file = next(
+                    (f for f in files if f.endswith(f'metadata_{timestamp}.json')), 
+                    None
+                )
+                if not metadata_file:
+                    continue
+                    
+                # Read metadata file
+                try:
+                    metadata_content = hub_storage.get_file_content(metadata_file)
                     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 reading metadata for {folder_name}: {str(e)}")
+                    continue
+                
+                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)}")
+                print(f"Error processing folder {folder_name}: {str(e)}")
                 continue
         
         print("Leaderboard initialized with existing submissions:")
@@ -290,6 +333,8 @@ def initialize_leaderboard():
         df_synthesized_10 = pd.DataFrame(data_synthesized_10)
         df_human_generated = pd.DataFrame(data_human_generated)
         
+        print("Initialized with baseline results")
+        
         # Then scan and add submitted results
         scan_submissions_directory()
         
@@ -298,6 +343,22 @@ def initialize_leaderboard():
     except Exception as e:
         print(f"Error initializing leaderboard: {str(e)}")
 
+# Utility function to get file content
+def get_file_content(file_path):
+    """
+    Helper function to safely read file content from HuggingFace repository
+    """
+    try:
+        api = HfApi()
+        content = api.file_download(
+            repo_id=REPO_ID,
+            repo_type="space",
+            filename=file_path
+        )
+        return content.read().decode('utf-8')
+    except Exception as e:
+        print(f"Error reading file {file_path}: {str(e)}")
+        return None
 
 def save_submission(submission_data, csv_file):
     """
@@ -893,12 +954,6 @@ with gr.Blocks(css=css) as demo:
         outputs=result
     )
     
-    # Initial table update
-    demo.load(
-        update_tables,
-        inputs=[model_type_filter],
-        outputs=all_dfs
-    )
 
 # Launch the application
 demo.launch()

utils/hub_storage.py

@@ -27,4 +27,20 @@ class HubStorage:
             filename=path_in_repo,
             repo_type="space",
             token=self.token
-        )
+        )
+    
+    def get_file_content(self, file_path):
+        """
+        Get content of a file from the repository
+        """
+        try:
+            api = HfApi()
+            content = api.file_download(
+                repo_id=self.repo_id,
+                repo_type="space",
+                filename=file_path
+            )
+            return content.read().decode('utf-8')
+        except Exception as e:
+            print(f"Error reading file {file_path}: {str(e)}")
+            return None