Upload supabase_storage.py

supabase_storage.py

@@ -239,31 +239,63 @@ def list_datasets_from_supabase() -> List[Dict[str, Any]]:
 
     try:
         storage = client.storage.from_(BUCKET_NAME)
+        dataset_names = []
 
-        # List all items in the bucket root (pass empty string for root path)
-        items = storage.list(path="")
-        print(f"[Supabase] Bucket contains {len(items)} items")
+        # Method 1: Try storage.list() with empty path
+        try:
+            items = storage.list(path="")
+            print(f"[Supabase] Method 1 (list path=''): {len(items)} items")
+            for item in items:
+                name = item.get('name', '')
+                if name and not name.startswith('_') and not name.startswith('.'):
+                    item_metadata = item.get('metadata')
+                    is_file = item_metadata and item_metadata.get('mimetype')
+                    if not is_file:
+                        dataset_names.append(name)
+                        print(f"[Supabase] Found via list(): {name}")
+        except Exception as e:
+            print(f"[Supabase] Method 1 failed: {e}")
 
-        # Debug: print what we got
-        for item in items:
-            print(f"[Supabase] Item: {item.get('name')} - metadata: {item.get('metadata')}")
+        # Method 2: If list() returned nothing, try listing without path argument
+        if not dataset_names:
+            try:
+                items = storage.list()
+                print(f"[Supabase] Method 2 (list no args): {len(items)} items")
+                for item in items:
+                    name = item.get('name', '')
+                    if name and not name.startswith('_') and not name.startswith('.'):
+                        item_metadata = item.get('metadata')
+                        is_file = item_metadata and item_metadata.get('mimetype')
+                        if not is_file:
+                            dataset_names.append(name)
+                            print(f"[Supabase] Found via list(): {name}")
+            except Exception as e:
+                print(f"[Supabase] Method 2 failed: {e}")
 
-        # Filter to get only folders (datasets), excluding files like _registry.json
-        # In Supabase Storage, folders typically have id=None or no mimetype
-        dataset_names = []
-        for item in items:
-            name = item.get('name', '')
-            # Skip hidden files and registry
-            if name.startswith('_') or name.startswith('.'):
-                continue
-            # Skip if it's a file (has mimetype in metadata)
-            # Folders in Supabase typically don't have mimetype
-            item_metadata = item.get('metadata')
-            is_file = item_metadata and item_metadata.get('mimetype')
-
-            if not is_file:
-                dataset_names.append(name)
-                print(f"[Supabase] Found dataset folder: {name}")
+        # Method 3: Fallback - read from registry file if exists
+        if not dataset_names:
+            print("[Supabase] List methods returned empty, trying registry fallback...")
+            try:
+                response = storage.download("_registry.json")
+                registry = json.loads(response.decode('utf-8'))
+                dataset_names = registry.get('datasets', [])
+                print(f"[Supabase] Found via registry: {dataset_names}")
+            except Exception as e:
+                print(f"[Supabase] Registry fallback failed: {e}")
+
+        # Method 4: Ultimate fallback - probe for known dataset names
+        if not dataset_names:
+            print("[Supabase] Trying direct probe for datasets...")
+            potential_names = ['testing', 'only_ariya___siglip', 'onlyariya_clip']
+            for name in potential_names:
+                try:
+                    storage.download(f"{name}/metadata.json")
+                    dataset_names.append(name)
+                    print(f"[Supabase] Found via probe: {name}")
+                except:
+                    pass
+
+        print(f"[Supabase] Total datasets found: {len(dataset_names)} - {dataset_names}")
 
         datasets = []
         for folder_name in dataset_names: