chore: Allow opening pre-existing unnamed storage

src/crawlee/storages/_storage_instance_manager.py

@@ -158,18 +158,22 @@ async def open_storage_instance(
             instance = cls(client, metadata.id, metadata.name)  # type: ignore[call-arg]
             instance_name = getattr(instance, 'name', None)
 
-            # Cache the instance. Always cache by id and cache named or unnamed (alias).
+            # Cache the instance.
+            # Always cache by id.
             self._cache_by_storage_client[storage_client_type].by_id[cls][instance.id][additional_cache_key] = instance
+
+            # Cache named storage.
             if instance_name is not None:
                 self._cache_by_storage_client[storage_client_type].by_name[cls][instance_name][additional_cache_key] = (
                     instance
                 )
-            elif alias is not None:
+
+            # Cache unnamed storage.
+            if alias is not None:
                 self._cache_by_storage_client[storage_client_type].by_alias[cls][alias][additional_cache_key] = instance
-            else:
-                raise RuntimeError('Storage instance must have either a name or an alias.')
 
             return instance
+
         finally:
             # Make sure the client opener is closed.
             # If it was awaited, then closing is no operation, if it was not awaited, this is the cleanup.

tests/unit/storages/test_storage_instance_manager.py

@@ -128,3 +128,16 @@ async def test_identical_storage_remove_from_cache(storage_type: type[Storage])
     service_locator.storage_instance_manager.remove_from_cache(storage_1)
     storage_2 = await storage_type.open()
     assert storage_1 is not storage_2
+
+
+async def test_preexisting_unnamed_storage_open_by_id(storage_type: type[Storage]) -> None:
+    """Test that persisted pre-existing unnamed storage can be opened by ID."""
+    storage_client = FileSystemStorageClient()
+    storage_1 = await storage_type.open(alias='custom_name', storage_client=storage_client)
+
+    # Make service_locator unaware of this storage
+    service_locator.storage_instance_manager.clear_cache()
+
+    storage_1_again = await storage_type.open(id=storage_1.id, storage_client=storage_client)
+
+    assert storage_1.id == storage_1_again.id