feat: Migrate dataset backups to gzipped tar archives and ensure backward compatibility during restore and rotation.

scripts/restore_from_dataset.py

@@ -28,8 +28,8 @@ def main() -> None:
         api = HfApi(token=token)
         files = api.list_repo_files(repo_id=repo_id, repo_type="dataset")
         
-        # Filter for our backup pattern
-        backups = sorted([f for f in files if f.startswith("state/backup-") and f.endswith(".tar")], reverse=True)
+        # Filter for our backup pattern (support both .tar and .tar.gz)
+        backups = sorted([f for f in files if f.startswith("state/backup-") and (f.endswith(".tar") or f.endswith(".tar.gz"))], reverse=True)
         
         if not backups:
             # Fallback to legacy filename if no rolling backups exist
@@ -51,7 +51,8 @@ def main() -> None:
                     token=token,
                 )
                 
-                with tarfile.open(tar_path, "r") as tf:
+                # Auto-detect compression based on file extension or header (r:*)
+                with tarfile.open(tar_path, "r:*") as tf:
                     tf.extractall(state_dir)
                 
                 print(f"[restore_from_dataset] Successfully restored from {backup_file}")

scripts/save_to_dataset.py

@@ -65,13 +65,13 @@ def main() -> None:
 
     # Generate timestamped filename
     timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
-    filename = f"state/backup-{timestamp}.tar"
+    filename = f"state/backup-{timestamp}.tar.gz"
 
     with tempfile.TemporaryDirectory() as tmpdir:
-        tar_path = os.path.join(tmpdir, "openclaw.tar")
+        tar_path = os.path.join(tmpdir, "openclaw.tar.gz")
 
         try:
-            with tarfile.open(tar_path, "w") as tf:
+            with tarfile.open(tar_path, "w:gz") as tf:
                 # Filter to exclude lock files or temp files if needed, but allow extensions
                 def exclude_filter(info: tarfile.TarInfo) -> tarfile.TarInfo | None:
                     if info.name.endswith(".lock"):
@@ -98,7 +98,8 @@ def main() -> None:
     # 2. Rotation: Delete old backups, keep last 5
     try:
         files = api.list_repo_files(repo_id=repo_id, repo_type="dataset")
-        backups = sorted([f for f in files if f.startswith("state/backup-") and f.endswith(".tar")])
+        # Match both .tar and .tar.gz for backward compatibility during transition
+        backups = sorted([f for f in files if f.startswith("state/backup-") and (f.endswith(".tar") or f.endswith(".tar.gz"))])
         
         if len(backups) > 5:
             # Delete oldest