fix dataset upload and download

sync_storage.py

@@ -223,10 +223,25 @@ This repository is automatically managed by the Open WebUI sync system.
                 return
             # Try to download the latest data archive
             try:
+                metadata = self._get_metadata()
+
+                logger.info(f"Found {len(current_files)} files to preserve.")
+                archive_filenames = [i["filename"] for i in metadata["archives"]]
+                current_files = list_repo_files(self.repo_id, repo_type="dataset", token=self.token)
+                logger.info(f"Archive filenames: {archive_filenames}")
+                logger.info(f"Current filenames: {current_files}")
+                latest_link = self.latest_link
+                if (latest_link not in current_files) and (len(archive_filenames) > 0):
+                    latest_link = archive_filenames[0]
+                    logger.info(f"Latest link not found, falling back to first archive: {latest_link}")
+                else:
+                    logger.error("No archives found in repository")
+                    return
+                logger.info(f"Downloading latest data archive: {latest_link}")
                 # First try the latest link
                 file_path = self.api.hf_hub_download(
                     repo_id=self.repo_id,
-                    filename=self.latest_link,
+                    filename=latest_link,
                     repo_type="dataset",
                     token=self.token
                 )
@@ -379,6 +394,7 @@ This repository is automatically managed by the Open WebUI sync system.
 
             logger.info(f"Found {len(current_files)} files to preserve.")
             archive_filenames = [i["filename"] for i in metadata["archives"]]
+            archive_filenames.append(self.latest_link)  # Also preserve latest link
             logger.info(f"Archive filenames: {archive_filenames}")
 
             # 2. Download all current files to a temporary directory