Make default Apify storages use alias mechanism

Author: Pijukatel

docs/04_upgrading/upgrading_to_v3.md

@@ -45,6 +45,13 @@ async def main():
 ## Removed Actor.config property
 - `Actor.config` property has been removed. Use `Actor.configuration` instead.
 
+## Default storage ids in configuration changed to None
+- `Configuration.default_key_value_store_id` changed from `'default'` to `None`.
+- `Configuration.default_dataset_id` changed from `'default'` to `None`.
+- `Configuration.default_request_queue_id` changed from `'default'` to `None`.
+
+As a consequence of this change, using default storage without specifying its `id` in `Configuration` will use unnamed storage.
+
 ## Storages
 
 <!-- TODO -->

src/apify/_configuration.py

@@ -142,37 +142,37 @@ class Configuration(CrawleeConfiguration):
     ] = None
 
     default_dataset_id: Annotated[
-        str,
+        str | None,
         Field(
             validation_alias=AliasChoices(
                 'actor_default_dataset_id',
                 'apify_default_dataset_id',
             ),
             description='Default dataset ID used by the Apify storage client when no ID or name is provided.',
         ),
-    ] = 'default'
+    ] = None
 
     default_key_value_store_id: Annotated[
-        str,
+        str | None,
         Field(
             validation_alias=AliasChoices(
                 'actor_default_key_value_store_id',
                 'apify_default_key_value_store_id',
             ),
             description='Default key-value store ID for the Apify storage client when no ID or name is provided.',
         ),
-    ] = 'default'
+    ] = None
 
     default_request_queue_id: Annotated[
-        str,
+        str | None,
         Field(
             validation_alias=AliasChoices(
                 'actor_default_request_queue_id',
                 'apify_default_request_queue_id',
             ),
             description='Default request queue ID for the Apify storage client when no ID or name is provided.',
         ),
-    ] = 'default'
+    ] = None
 
     disable_outdated_warning: Annotated[
         bool,

src/apify/storage_clients/_apify/_dataset_client.py

@@ -124,8 +124,10 @@ async def open(
         )
         apify_datasets_client = apify_client_async.datasets()
 
-        # Normalize 'default' alias to None
-        alias = None if alias == 'default' else alias
+        # Normalize unnamed default storage in cases where not defined in `configuration.default_dataset_id` to unnamed
+        # storage aliased as `__default__`
+        if not any([alias, name, id, configuration.default_dataset_id]):
+            alias = '__default__'
 
         if alias:
             # Check if there is pre-existing alias mapping in the default KVS.
@@ -150,6 +152,11 @@ async def open(
         # If none are provided, try to get the default storage ID from environment variables.
         elif id is None:
             id = configuration.default_dataset_id
+            if not id:
+                raise ValueError(
+                    'Dataset "id", "name", or "alias" must be specified, '
+                    'or a default dataset ID must be set in the configuration.'
+                )
 
         # Now create the client for the determined ID
         apify_dataset_client = apify_client_async.dataset(dataset_id=id)

src/apify/storage_clients/_apify/_key_value_store_client.py

@@ -115,8 +115,10 @@ async def open(
         )
         apify_kvss_client = apify_client_async.key_value_stores()
 
-        # Normalize 'default' alias to None
-        alias = None if alias == 'default' else alias
+        # Normalize unnamed default storage in cases where not defined in `configuration.default_key_value_store_id` to
+        # unnamed storage aliased as `__default__`
+        if not any([alias, name, id, configuration.default_key_value_store_id]):
+            alias = '__default__'
 
         if alias:
             # Check if there is pre-existing alias mapping in the default KVS.
@@ -142,6 +144,11 @@ async def open(
         # If none are provided, try to get the default storage ID from environment variables.
         elif id is None:
             id = configuration.default_key_value_store_id
+            if not id:
+                raise ValueError(
+                    'KeyValueStore "id", "name", or "alias" must be specified, '
+                    'or a default KeyValueStore ID must be set in the configuration.'
+                )
 
         # Now create the client for the determined ID
         apify_kvs_client = apify_client_async.key_value_store(key_value_store_id=id)

src/apify/storage_clients/_apify/_request_queue_client.py

@@ -200,8 +200,10 @@ async def open(
         )
         apify_rqs_client = apify_client_async.request_queues()
 
-        # Normalize 'default' alias to None
-        alias = None if alias == 'default' else alias
+        # Normalize unnamed default storage in cases where not defined in `configuration.default_request_queue_id` to
+        # unnamed storage aliased as `__default__`
+        if not any([alias, name, id, configuration.default_request_queue_id]):
+            alias = '__default__'
 
         if alias:
             # Check if there is pre-existing alias mapping in the default KVS.
@@ -226,6 +228,11 @@ async def open(
         # If none are provided, try to get the default storage ID from environment variables.
         elif id is None:
             id = configuration.default_request_queue_id
+            if not id:
+                raise ValueError(
+                    'RequestQueue "id", "name", or "alias" must be specified, '
+                    'or a default default_request_queue_id ID must be set in the configuration.'
+                )
 
         # Use suitable client_key to make `hadMultipleClients` response of Apify API useful.
         # It should persist across migrated or resurrected Actor runs on the Apify platform.

src/apify/storage_clients/_apify/_utils.py

@@ -76,7 +76,7 @@ async def _get_alias_map(cls) -> dict[str, str]:
         Returns:
             Map of aliases and storage ids.
         """
-        if not cls._alias_map:
+        if not cls._alias_map and Configuration.get_global_configuration().is_at_home:
             default_kvs_client = await _get_default_kvs_client()
 
             record = await default_kvs_client.get_record(cls._ALIAS_MAPPING_KEY)
@@ -156,7 +156,8 @@ async def _get_default_kvs_client() -> KeyValueStoreClientAsync:
         min_delay_between_retries_millis=500,
         timeout_secs=360,
     )
-
+    if not configuration.default_key_value_store_id:
+        raise ValueError("'Configuration.default_key_value_store_id' must be set.")
     return apify_client_async.key_value_store(key_value_store_id=configuration.default_key_value_store_id)
 
 

tests/integration/test_apify_storages.py

@@ -32,3 +32,20 @@ async def test_alias_concurrent_creation_local(
     except AssertionError:
         for storage in storages:
             await storage.drop()
+
+
+@pytest.mark.parametrize(
+    'storage_type',
+    [Dataset, KeyValueStore, RequestQueue],
+)
+async def test_unnamed_default_without_config(
+    storage_type: Dataset | KeyValueStore | RequestQueue, apify_token: str
+) -> None:
+    """Test that default Apify storage used locally is unnamed storage."""
+    service_locator.set_configuration(Configuration(token=apify_token))
+    service_locator.set_storage_client(ApifyStorageClient())
+
+    storage = await storage_type.open()
+    assert storage.name is None
+    assert storage.id
+    await storage.drop()