apify 2.0.1__py3-none-any.whl → 2.0.1b1__py3-none-any.whl

apify/_configuration.py

@@ -5,10 +5,8 @@ from datetime import datetime, timedelta
 from typing import Annotated
 
 from pydantic import AliasChoices, BeforeValidator, Field
-from typing_extensions import deprecated
 
 from crawlee._utils.models import timedelta_ms
-from crawlee._utils.urls import validate_http_url
 from crawlee.configuration import Configuration as CrawleeConfiguration
 
 
@@ -162,16 +160,15 @@ class Configuration(CrawleeConfiguration):
         str | None,
         Field(
             alias='apify_sdk_latest_version',
+            deprecated=True,
             description='Specifies the most recent release version of the Apify SDK for Javascript. Used for '
             'checking for updates.',
         ),
-        deprecated('SDK version checking is not supported for the Python SDK'),
     ] = None
 
     log_format: Annotated[
         str | None,
-        Field(alias='apify_log_format'),
-        deprecated('Adjust the log format in code instead'),
+        Field(alias='apify_log_format', deprecated=True),
     ] = None
 
     max_paid_dataset_items: Annotated[
@@ -251,7 +248,6 @@ class Configuration(CrawleeConfiguration):
             ),
             description='Date when the Actor will time out',
         ),
-        BeforeValidator(lambda val: val if val != '' else None),  # We should accept empty environment variables as well
     ] = None
 
     standby_port: Annotated[
@@ -260,17 +256,7 @@ class Configuration(CrawleeConfiguration):
             alias='actor_standby_port',
             description='TCP port for the Actor to start an HTTP server to receive messages in the Actor Standby mode',
         ),
-        deprecated('Use `web_server_port` instead'),
-    ] = 4321
-
-    standby_url: Annotated[
-        str,
-        BeforeValidator(validate_http_url),
-        Field(
-            alias='actor_standby_url',
-            description='URL for accessing web servers of Actor runs in Standby mode',
-        ),
-    ] = 'http://localhost'
+    ] = 4322
 
     token: Annotated[
         str | None,
@@ -307,7 +293,7 @@ class Configuration(CrawleeConfiguration):
                 'actor_web_server_url',
                 'apify_container_url',
             ),
-            description='Unique public URL for accessing a specific Actor run web server from the outside world',
+            description='Unique public URL for accessing the Actor run web server from the outside world',
         ),
     ] = 'http://localhost:4321'
 

apify/_crypto.py

@@ -114,7 +114,6 @@ def private_decrypt(
     return decipher_bytes.decode('utf-8')
 
 
-@ignore_docs
 def load_private_key(private_key_file_base64: str, private_key_password: str) -> rsa.RSAPrivateKey:
     private_key = serialization.load_pem_private_key(
         base64.b64decode(private_key_file_base64.encode('utf-8')),
@@ -134,7 +133,6 @@ def _load_public_key(public_key_file_base64: str) -> rsa.RSAPublicKey:
     return public_key
 
 
-@ignore_docs
 def decrypt_input_secrets(private_key: rsa.RSAPrivateKey, input_data: Any) -> Any:
     """Decrypt input secrets."""
     if not isinstance(input_data, dict):

apify/_platform_event_manager.py

@@ -8,6 +8,7 @@ import websockets.client
 from pydantic import BaseModel, Discriminator, Field, TypeAdapter
 from typing_extensions import Self, Unpack, override
 
+from apify_shared.utils import ignore_docs
 from crawlee.events._event_manager import EventManager, EventManagerOptions
 from crawlee.events._local_event_manager import LocalEventManager
 from crawlee.events._types import (
@@ -125,6 +126,7 @@ event_data_adapter: TypeAdapter[EventMessage | DeprecatedEvent | UnknownEvent] =
 )
 
 
+@ignore_docs
 class PlatformEventManager(EventManager):
     """A class for managing Actor events.
 

apify/_proxy_configuration.py

@@ -27,7 +27,6 @@ COUNTRY_CODE_REGEX = re.compile(r'^[A-Z]{2}$')
 SESSION_ID_MAX_LENGTH = 50
 
 
-@ignore_docs
 def is_url(url: str) -> bool:
     """Check if the given string is a valid URL."""
     try:

apify/_utils.py

@@ -11,7 +11,6 @@ def get_system_info() -> dict:
     system_info: dict[str, str | bool] = {
         'apify_sdk_version': metadata.version('apify'),
         'apify_client_version': metadata.version('apify-client'),
-        'crawlee_version': metadata.version('crawlee'),
         'python_version': python_version,
         'os': sys.platform,
     }

apify/apify_storage_client/_request_queue_client.py

@@ -2,6 +2,7 @@ from __future__ import annotations
 
 from typing import TYPE_CHECKING
 
+from more_itertools import chunked
 from typing_extensions import override
 
 from crawlee import Request
@@ -157,8 +158,11 @@ class RequestQueueClient(BaseRequestQueueClient):
         *,
         forefront: bool = False,
     ) -> BatchRequestsOperationResponse:
-        return BatchRequestsOperationResponse.model_validate(
-            await self._client.batch_add_requests(
+        processed = []
+        unprocessed = []
+
+        for chunk in chunked(requests, 25):  # The API endpoint won't accept more than 25 requests at once
+            response = await self._client.batch_add_requests(
                 requests=[
                     r.model_dump(
                         by_alias=True,
@@ -170,10 +174,18 @@ class RequestQueueClient(BaseRequestQueueClient):
                             'data',
                         },
                     )
-                    for r in requests
+                    for r in chunk
                 ],
                 forefront=forefront,
             )
+            processed.extend(response['processedRequests'])
+            unprocessed.extend(response['unprocessedRequests'])
+
+        return BatchRequestsOperationResponse.model_validate(
+            {
+                'processedRequests': processed,
+                'unprocessedRequests': unprocessed,
+            }
         )
 
     @override

apify/log.py

@@ -3,7 +3,6 @@ from __future__ import annotations
 import logging
 from typing import TYPE_CHECKING
 
-from apify_shared.utils import ignore_docs
 from crawlee._log_config import CrawleeLogFormatter, configure_logger, get_configured_log_level
 
 if TYPE_CHECKING:
@@ -16,7 +15,6 @@ logger_name = __name__.split('.')[0]
 logger = logging.getLogger(logger_name)
 
 
-@ignore_docs
 class ActorLogFormatter(CrawleeLogFormatter):  # noqa: D101 Inherited from parent class
     pass
 

apify/scrapy/requests.py

@@ -4,8 +4,6 @@ import codecs
 import pickle
 from typing import Any, cast
 
-from apify_shared.utils import ignore_docs
-
 try:
     from scrapy import Request, Spider
     from scrapy.http.headers import Headers
@@ -16,7 +14,6 @@ except ImportError as exc:
     ) from exc
 
 from crawlee import Request as CrawleeRequest
-from crawlee._types import HttpHeaders
 from crawlee._utils.crypto import crypto_random_object_id
 from crawlee._utils.requests import compute_unique_key, unique_key_to_request_id
 
@@ -31,7 +28,6 @@ def _is_request_produced_by_middleware(scrapy_request: Request) -> bool:
     return bool(scrapy_request.meta.get('redirect_times')) or bool(scrapy_request.meta.get('retry_times'))
 
 
-@ignore_docs
 def to_apify_request(scrapy_request: Request, spider: Spider) -> CrawleeRequest | None:
     """Convert a Scrapy request to an Apify request.
 
@@ -78,9 +74,9 @@ def to_apify_request(scrapy_request: Request, spider: Spider) -> CrawleeRequest
             id=request_id,
         )
 
-        # Convert Scrapy's headers to a HttpHeaders and store them in the apify_request
+        # Convert Scrapy's headers to a dictionary and store them in the apify_request
         if isinstance(scrapy_request.headers, Headers):
-            apify_request.headers = HttpHeaders(scrapy_request.headers.to_unicode_dict())
+            apify_request.headers = dict(scrapy_request.headers.to_unicode_dict())
         else:
             Actor.log.warning(
                 f'Invalid scrapy_request.headers type, not scrapy.http.headers.Headers: {scrapy_request.headers}'
@@ -102,7 +98,6 @@ def to_apify_request(scrapy_request: Request, spider: Spider) -> CrawleeRequest
     return apify_request
 
 
-@ignore_docs
 def to_scrapy_request(apify_request: CrawleeRequest, spider: Spider) -> Request:
     """Convert an Apify request to a Scrapy request.
 
@@ -165,7 +160,13 @@ def to_scrapy_request(apify_request: CrawleeRequest, spider: Spider) -> Request:
 
     # Add optional 'headers' field
     if apify_request.headers:
-        scrapy_request.headers |= Headers(apify_request.headers)
+        if isinstance(cast(Any, apify_request.headers), dict):
+            scrapy_request.headers = Headers(apify_request.headers)
+        else:
+            Actor.log.warning(
+                'apify_request[headers] is not an instance of the dict class, '
+                f'apify_request[headers] = {apify_request.headers}',
+            )
 
     # Add optional 'userData' field
     if apify_request.user_data:

apify/scrapy/utils.py

@@ -4,8 +4,6 @@ import asyncio
 from base64 import b64encode
 from urllib.parse import unquote
 
-from apify_shared.utils import ignore_docs
-
 try:
     from scrapy.settings import Settings  # noqa: TCH002
     from scrapy.utils.project import get_project_settings
@@ -20,7 +18,6 @@ except ImportError as exc:
 nested_event_loop: asyncio.AbstractEventLoop = asyncio.new_event_loop()
 
 
-@ignore_docs
 def get_basic_auth_header(username: str, password: str, auth_encoding: str = 'latin-1') -> bytes:
     """Generate a basic authentication header for the given username and password."""
     string = f'{unquote(username)}:{unquote(password)}'
@@ -28,7 +25,6 @@ def get_basic_auth_header(username: str, password: str, auth_encoding: str = 'la
     return b'Basic ' + b64encode(user_pass)
 
 
-@ignore_docs
 def get_running_event_loop_id() -> int:
     """Get the ID of the currently running event loop.
 

{apify-2.0.1.dist-info → apify-2.0.1b1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: apify
-Version: 2.0.1
+Version: 2.0.1b1
 Summary: Apify SDK for Python
 License: Apache-2.0
 Keywords: apify,sdk,automation,chrome,crawlee,crawler,headless,scraper,scraping
@@ -16,12 +16,11 @@ Classifier: Programming Language :: Python :: 3.9
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
-Classifier: Programming Language :: Python :: 3.13
 Classifier: Topic :: Software Development :: Libraries
 Provides-Extra: scrapy
-Requires-Dist: apify-client (>=1.8.1)
+Requires-Dist: apify-client (>=1.7.1)
 Requires-Dist: apify-shared (>=1.1.2)
-Requires-Dist: crawlee (>=0.3.8)
+Requires-Dist: crawlee (>=0.3.5)
 Requires-Dist: cryptography (>=42.0.0)
 Requires-Dist: httpx (>=0.27.0)
 Requires-Dist: lazy-object-proxy (>=1.10.0)

{apify-2.0.1.dist-info → apify-2.0.1b1.dist-info}/RECORD

@@ -1,22 +1,22 @@
 apify/__init__.py,sha256=ikoi2EpDYl6y-XSVtlU8UsdQdMEyOiIJCRRAaZFDOP8,550
 apify/_actor.py,sha256=oPgQ3rxxIEzVcZ9XtI3lf1a_6gwIMgxihNuYGjJpGww,41816
-apify/_configuration.py,sha256=n67sK9g8llsG8ZqIWhi1BEHliplGs5PTB6Ig7_0PaMU,9629
+apify/_configuration.py,sha256=gf7YOun32Whc9DamhoWDLmcUeNwtWVmmBPrl4oq6s4I,8997
 apify/_consts.py,sha256=_Xq4hOfOA1iZ3n1P967YWdyncKivpbX6RTlp_qanUoE,330
-apify/_crypto.py,sha256=e0_aM3l9_5Osk-jszYOOjrAKK60OggSHbiw5c30QnsU,5638
+apify/_crypto.py,sha256=b4Czs1NLPkaNzkPjovObjSIbsKnRrgtBkM9JvOysUMA,5612
 apify/_models.py,sha256=oYlTEr-DyQAE-V2rrYD5PhUxTXVPdAig7QV-u6CJw3E,5571
-apify/_platform_event_manager.py,sha256=K4cHabbQ7_ex7vkX-c-VhAOp8Efw3HDn5Wp4lfA-qAU,7571
-apify/_proxy_configuration.py,sha256=2z4VV_NrnIp6pDpgQKlKpcHM2pPyXiOpFedpPWje48A,13087
-apify/_utils.py,sha256=Ghho3Gf11zYN8qhjhAkZRvQ--A9Js36GHB0YSDGUK58,694
+apify/_platform_event_manager.py,sha256=h5fBmXtKD4t-yCdOSiLM1-DnCrIbGEmYmz2mOU3A8bA,7627
+apify/_proxy_configuration.py,sha256=VdKh_AyCwaCUlpCyaCe30L2S9OZ-vL1SN1g8oLwSeYA,13074
+apify/_utils.py,sha256=x4lnR9RNulySiEQTft-GeQqUcJsRr0k8p0Sv9NTeWFg,638
 apify/apify_storage_client/__init__.py,sha256=-UbR68bFsDR6ln8OFs4t50eqcnY36hujO-SeOt-KmcA,114
 apify/apify_storage_client/_apify_storage_client.py,sha256=xi4OFchxhe-1-sykanH6Zcya4OcBhn2uf7OQ1pV4Ins,2338
 apify/apify_storage_client/_dataset_client.py,sha256=j9seF2OKvbSMD9R9XF9fpa1vtr_1w4JcRV--WCmvU4E,5501
 apify/apify_storage_client/_dataset_collection_client.py,sha256=fkYvYGQCigHD2CDzpWk0swNAkfvAinAhMGpYqllle3E,1445
 apify/apify_storage_client/_key_value_store_client.py,sha256=uyeQgb75sGFsqIS4sq4hEZ3QP81COLfS3tmTqHc0tso,3340
 apify/apify_storage_client/_key_value_store_collection_client.py,sha256=vCtMTI-jx89Qp5WHILDNkCthwLuv0MAwm1J_5E4aypU,1519
-apify/apify_storage_client/_request_queue_client.py,sha256=jAiFkaJ38_myHFGTw-Rk21wmpbN0UCR2w2SFoimFGFc,5826
+apify/apify_storage_client/_request_queue_client.py,sha256=P8ws8jEzi2PWpp-cvYfV7kwuKbgH813BpNQ_wMSVtTA,6278
 apify/apify_storage_client/_request_queue_collection_client.py,sha256=NnO73UJ9ZrjV8xoudo30wfaM-SojRkG0guhxDyB-K1g,1527
 apify/apify_storage_client/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-apify/log.py,sha256=zIVjrqQ1DNWNQQOAmdmR9oAbf4nJH7CSMB6u4OOUf6I,1448
+apify/log.py,sha256=pX6ppIvds8OKqjFpIcshqG4zp_5DiOUU31ksyfSExto,1392
 apify/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 apify/scrapy/__init__.py,sha256=qDPV_zTRFaUqoFOyS5g4uBfz-UCkmWYJ82VXQ_3Cw6k,348
 apify/scrapy/middlewares/__init__.py,sha256=tfW-d3WFWLeNEjL8fTmon6NwgD-OXx1Bw2fBdU-wPy4,114
@@ -26,12 +26,12 @@ apify/scrapy/pipelines/__init__.py,sha256=GWPeLN_Zwj8vRBWtXW6DaxdB7mvyQ7Jw5Tz1cc
 apify/scrapy/pipelines/actor_dataset_push.py,sha256=QERmmExQOGIKQ70-p-lCj5qyE-c-fnYplEqd4mgaB1Q,953
 apify/scrapy/pipelines/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 apify/scrapy/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-apify/scrapy/requests.py,sha256=F4VNaX2fGqybJKbhcRcz0_m6dXse5LzKll4gtMuTRko,7480
+apify/scrapy/requests.py,sha256=pmm2M-cwrTXyI3t1nRBo9pS6nHfc4zkzS25-NXxzd9I,7637
 apify/scrapy/scheduler.py,sha256=AAIKY5i1QxkC1mtmix6n3M2eQaOw-d1T56Noue9xToc,6013
-apify/scrapy/utils.py,sha256=0XdFxi1qlUa6gHXG96e1FU9gW0N5Rsu0sVZklFYfC2U,2884
+apify/scrapy/utils.py,sha256=tz_Y8CTqe6KbyMMhLF3m7qqR46jtNH5U7Ty7e19roPU,2814
 apify/storages/__init__.py,sha256=-9tEYJVabVs_eRVhUehxN58GH0UG8OfuGjGwuDieP2M,122
 apify/storages/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-apify-2.0.1.dist-info/LICENSE,sha256=AsFjHssKjj4LGd2ZCqXn6FBzMqcWdjQre1byPPSypVw,11355
-apify-2.0.1.dist-info/METADATA,sha256=yD0LEyfMGH3_ePU2-BCBtCJMwFwTOvUVOP730cXjrE0,8655
-apify-2.0.1.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
-apify-2.0.1.dist-info/RECORD,,
+apify-2.0.1b1.dist-info/LICENSE,sha256=AsFjHssKjj4LGd2ZCqXn6FBzMqcWdjQre1byPPSypVw,11355
+apify-2.0.1b1.dist-info/METADATA,sha256=GPwXUjHARouYHQTkQ8HCOgMn-rJcf8taaugNu5aoVkA,8606
+apify-2.0.1b1.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+apify-2.0.1b1.dist-info/RECORD,,

{apify-2.0.1.dist-info → apify-2.0.1b1.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: poetry-core 1.9.1
+Generator: poetry-core 1.9.0
 Root-Is-Purelib: true
 Tag: py3-none-any