crawlee 1.0.2b2__py3-none-any.whl → 1.0.2b4__py3-none-any.whl

crawlee/storages/_base.py

@@ -44,7 +44,9 @@ class Storage(ABC):
 
         Args:
             id: The storage ID.
-            name: The storage name (global scope, persists across runs).
+            name: The storage name (global scope, persists across runs). Name can only contain letters "a" through "z",
+                the digits "0" through "9", and the hyphen ("-") but only in the middle of the string
+                (e.g. "my-value-1").
             alias: The storage alias (run scope, creates unnamed storage).
             configuration: Configuration object used during the storage creation or restoration process.
             storage_client: Underlying storage client to use. If not provided, the default global storage client

crawlee/storages/_dataset.py

@@ -12,6 +12,7 @@ from crawlee._utils.file import export_csv_to_stream, export_json_to_stream
 
 from ._base import Storage
 from ._key_value_store import KeyValueStore
+from ._utils import validate_storage_name
 
 if TYPE_CHECKING:
     from collections.abc import AsyncIterator
@@ -75,6 +76,8 @@ class Dataset(Storage):
             id: The unique identifier of the storage.
             name: The name of the storage, if available.
         """
+        validate_storage_name(name)
+
         self._client = client
         self._id = id
         self._name = name

crawlee/storages/_key_value_store.py

@@ -15,6 +15,7 @@ from crawlee._utils.recoverable_state import RecoverableState
 from crawlee.storage_clients.models import KeyValueStoreMetadata
 
 from ._base import Storage
+from ._utils import validate_storage_name
 
 if TYPE_CHECKING:
     from collections.abc import AsyncIterator
@@ -84,6 +85,8 @@ class KeyValueStore(Storage):
             id: The unique identifier of the storage.
             name: The name of the storage, if available.
         """
+        validate_storage_name(name)
+
         self._client = client
         self._id = id
         self._name = name

crawlee/storages/_request_queue.py

@@ -13,6 +13,7 @@ from crawlee._utils.wait import wait_for_all_tasks_for_finish
 from crawlee.request_loaders import RequestManager
 
 from ._base import Storage
+from ._utils import validate_storage_name
 
 if TYPE_CHECKING:
     from collections.abc import Sequence
@@ -80,6 +81,8 @@ class RequestQueue(Storage, RequestManager):
             id: The unique identifier of the storage.
             name: The name of the storage, if available.
         """
+        validate_storage_name(name)
+
         self._client = client
         self._id = id
         self._name = name

crawlee/storages/_storage_instance_manager.py

@@ -8,6 +8,8 @@ from typing import TYPE_CHECKING, TypeVar
 from crawlee._utils.raise_if_too_many_kwargs import raise_if_too_many_kwargs
 from crawlee.storage_clients._base import DatasetClient, KeyValueStoreClient, RequestQueueClient
 
+from ._utils import validate_storage_name
+
 if TYPE_CHECKING:
     from ._base import Storage
 
@@ -90,7 +92,9 @@ class StorageInstanceManager:
         Args:
             cls: The storage class to instantiate.
             id: Storage ID.
-            name: Storage name. (global scope, persists across runs).
+            name: Storage name. (global scope, persists across runs). Name can only contain letters "a" through "z",
+                the digits "0" through "9", and the hyphen ("-") but only in the middle of the string
+                (e.g. "my-value-1").
             alias: Storage alias (run scope, creates unnamed storage).
             client_opener_coro: Coroutine to open the storage client when storage instance not found in cache.
             storage_client_cache_key: Additional optional key from storage client to differentiate cache entries.
@@ -146,6 +150,10 @@ class StorageInstanceManager:
                     f'Use a different name or drop the existing alias storage first.'
                 )
 
+            # Validate storage name
+            if name is not None:
+                validate_storage_name(name)
+
             # Create new instance
             client: KeyValueStoreClient | DatasetClient | RequestQueueClient
             client = await client_opener_coro

crawlee/storages/_utils.py

@@ -0,0 +1,11 @@
+import re
+
+NAME_REGEX = re.compile(r'^([a-zA-Z0-9]|[a-zA-Z0-9][a-zA-Z0-9-]*[a-zA-Z0-9])$')
+
+
+def validate_storage_name(name: str | None) -> None:
+    if name and not NAME_REGEX.match(name):
+        raise ValueError(
+            f'Invalid storage name "{name}". Name can only contain letters "a" through "z", the digits "0" through'
+            '"9", and the hyphen ("-") but only in the middle of the string (e.g. "my-value-1")'
+        )

{crawlee-1.0.2b2.dist-info → crawlee-1.0.2b4.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: crawlee
-Version: 1.0.2b2
+Version: 1.0.2b4
 Summary: Crawlee for Python
 Project-URL: Apify Homepage, https://apify.com
 Project-URL: Changelog, https://crawlee.dev/python/docs/changelog

{crawlee-1.0.2b2.dist-info → crawlee-1.0.2b4.dist-info}/RECORD

@@ -180,14 +180,15 @@ crawlee/storage_clients/_sql/_request_queue_client.py,sha256=iavp-G62ApPtPmKePYv
 crawlee/storage_clients/_sql/_storage_client.py,sha256=3xfgUcdW7Pu_j3SDYFzAdnU81jl1CmZ9Z5_NLvNi4P8,10913
 crawlee/storage_clients/_sql/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 crawlee/storages/__init__.py,sha256=wc2eioyCKAAYrg4N7cshpjC-UbE23OzGar9nK_kteSY,186
-crawlee/storages/_base.py,sha256=Ka9RdYnEYlNLiKJGWqSwEDOPv5AtJWXKZErSTesh42Y,2124
-crawlee/storages/_dataset.py,sha256=T2C8Q3gDUWg8p0JBS5MHnCuSd252oDU61JuNZGZzg-I,14573
-crawlee/storages/_key_value_store.py,sha256=3oI5hVoM_NpTQVKXCbQCmb0sZhW7vN2oXQo-Yxi7BLQ,10127
-crawlee/storages/_request_queue.py,sha256=jt-d-NkI9lAorLssoI2r_lZjeEipe-5Cn6z9bfQqY3k,13154
-crawlee/storages/_storage_instance_manager.py,sha256=iFX3ymsIXyTg8tMHtx5Wn9XyaC77dIf15GpuggsJPDM,7821
+crawlee/storages/_base.py,sha256=zUOcMJTg8MAzq-m9X1NJcWncCfxzI5mb5MyY35WAkMk,2310
+crawlee/storages/_dataset.py,sha256=l3VJCaJnaAEhJFpfRUOLzIbW332R8gdEPSSGhLq65pg,14652
+crawlee/storages/_key_value_store.py,sha256=ik--ZPCzOiG5hmm6k5LNH_FO9P3MoW0UvTKiGVf1RIY,10206
+crawlee/storages/_request_queue.py,sha256=bjBOGbpMaGUsqJPVB-JD2VShziPAYMI-GvWKKpylzDE,13233
+crawlee/storages/_storage_instance_manager.py,sha256=72n0YlPwNpSQDJSPf4TxnI2GvIK6L-ZiTmHRbFcoVU0,8164
+crawlee/storages/_utils.py,sha256=Yz-5tEBYKYCFJemYT29--uGJqoJLApLDLgPcsnbifRw,439
 crawlee/storages/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-crawlee-1.0.2b2.dist-info/METADATA,sha256=t0naR_tcjn5u_u1v5ksG0sa8WzOwOZ0HryOIJ-6tKvE,29314
-crawlee-1.0.2b2.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-crawlee-1.0.2b2.dist-info/entry_points.txt,sha256=1p65X3dA-cYvzjtlxLL6Kn1wpY-3uEDVqJLp53uNPeo,45
-crawlee-1.0.2b2.dist-info/licenses/LICENSE,sha256=AsFjHssKjj4LGd2ZCqXn6FBzMqcWdjQre1byPPSypVw,11355
-crawlee-1.0.2b2.dist-info/RECORD,,
+crawlee-1.0.2b4.dist-info/METADATA,sha256=YlX57uGdGcahWhpTusPjBchIq9y47R9WnXHSka5T0PU,29314
+crawlee-1.0.2b4.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+crawlee-1.0.2b4.dist-info/entry_points.txt,sha256=1p65X3dA-cYvzjtlxLL6Kn1wpY-3uEDVqJLp53uNPeo,45
+crawlee-1.0.2b4.dist-info/licenses/LICENSE,sha256=AsFjHssKjj4LGd2ZCqXn6FBzMqcWdjQre1byPPSypVw,11355
+crawlee-1.0.2b4.dist-info/RECORD,,