scrapely-python 0.1.1__py3-none-any.whl → 0.1.2__py3-none-any.whl

scrapely/__init__.py

@@ -132,7 +132,7 @@ from .errors import (
     raise_for_status,
 )
 
-__version__ = '0.1.0'
+__version__ = '0.1.2'
 __all__ = [
     # Core
     'Actor',

scrapely/actor.py

@@ -1776,19 +1776,20 @@ class Actor(metaclass=_ActorMeta):
         """
         cls._ensure_initialized('create_proxy_configuration')
 
-        from .proxy_configuration import ProxyConfiguration
+        from .proxy_configuration import ProxyConfiguration, _input_use_platform_proxy
 
         # Handle actor_proxy_input (Node SDK normalizeProxyInput/createProxyConfiguration
-        # parity): `useApifyProxy: false` explicitly disables the PLATFORM proxy only —
+        # parity): `useScrapelyProxy: false` (or the legacy `useApifyProxy: false`; a
+        # boolean useScrapelyProxy wins) explicitly disables the PLATFORM proxy only —
         # custom proxies (proxyUrls / tiered lists / new_url_function) from the same
         # input must still be used. The standard "Custom proxy URLs" input shape is
-        # {useApifyProxy: false, proxyUrls: [...]}; returning None for it would
+        # {useScrapelyProxy: false, proxyUrls: [...]}; returning None for it would
         # silently drop the user's own proxies and crawl on their real IPs. We only
         # bail when no custom proxies are given either. A MISSING key falls back to
         # platform mode when the server injected credentials — input like
         # {apifyProxyGroups: [...]} must not silently disable proxying.
         if actor_proxy_input is not None:
-            if actor_proxy_input.get('useApifyProxy') is False and not (
+            if _input_use_platform_proxy(actor_proxy_input) is False and not (
                 proxy_urls
                 or new_url_function
                 or tiered_proxy_urls

scrapely/proxy_configuration.py

@@ -108,6 +108,19 @@ def _sanitize_fragment(value: str) -> str:
     return re.sub(r'[:@/?#\s]', '', value)
 
 
+def _input_use_platform_proxy(actor_proxy_input: Dict[str, Any]) -> Optional[bool]:
+    """The proxy input field's platform-proxy switch, or None when it has none.
+
+    `useScrapelyProxy` wins when it is a boolean; otherwise the legacy
+    `useApifyProxy` key of existing inputs and ported actors decides.
+    """
+    for key in ('useScrapelyProxy', 'useApifyProxy'):
+        value = actor_proxy_input.get(key)
+        if isinstance(value, bool):
+            return value
+    return None
+
+
 def _platform_env() -> Optional[Dict[str, Any]]:
     """Read platform proxy credentials from the environment (or None)."""
     hostname = os.environ.get('SCRAPELY_PROXY_HOSTNAME') or os.environ.get('APIFY_PROXY_HOSTNAME')
@@ -213,7 +226,7 @@ class ProxyConfiguration:
 
         # If actor_proxy_input is provided, extract settings from it
         if actor_proxy_input:
-            if actor_proxy_input.get('useApifyProxy', False):
+            if _input_use_platform_proxy(actor_proxy_input):
                 self._groups = actor_proxy_input.get('apifyProxyGroups', []) or []
                 self._country_code = actor_proxy_input.get('apifyProxyCountry')
                 self._subdivision_code = actor_proxy_input.get('apifyProxySubdivision')
@@ -291,7 +304,7 @@ class ProxyConfiguration:
         )
         if has_real_custom and (self._groups or self._country_code or self._subdivision_code or password):
             raise ValueError(
-                'Cannot combine custom proxies with Apify Proxy! It is not allowed to set '
+                'Cannot combine custom proxies with Scrapely Proxy! It is not allowed to set '
                 '"proxy_urls" or "new_url_function" combined with "groups", "country_code" '
                 'or "subdivision_code".'
             )
@@ -480,7 +493,8 @@ class ProxyConfiguration:
         Sends ``GET {status_url}/?format=json`` *through the platform gateway``
         so the proxy handshake itself authenticates the run's credentials.
         Mirrors the JS SDK: 4s timeout, 2 attempts, returns ``None`` when the
-        endpoint stays unreachable. Never raises.
+        endpoint stays unreachable. Never raises. ``fromPayload`` is True when
+        the gateway answered with its own status JSON rather than an HTTP error.
         """
         import urllib.error
         import urllib.request
@@ -514,6 +528,7 @@ class ProxyConfiguration:
                     'connected': payload.get('connected') is True,
                     'connectionError': payload.get('connectionError') or '',
                     'isManInTheMiddle': payload.get('isManInTheMiddle') is True,
+                    'fromPayload': True,
                 }
             except urllib.error.HTTPError as e:
                 if e.code == 407:
@@ -524,6 +539,7 @@ class ProxyConfiguration:
                             'SCRAPELY_PROXY_*/APIFY_PROXY_* credentials.'
                         ),
                         'isManInTheMiddle': False,
+                        'fromPayload': False,
                     }
                 if 590 <= e.code <= 599:
                     description = PROXY_STATUS_CODE_DESCRIPTIONS.get(e.code, 'upstream error')
@@ -531,6 +547,7 @@ class ProxyConfiguration:
                         'connected': False,
                         'connectionError': f'Platform proxy upstream error (HTTP {e.code}: {description}).',
                         'isManInTheMiddle': False,
+                        'fromPayload': False,
                     }
                 # Other HTTP errors behave like connection errors: retry.
             except Exception:  # noqa: BLE001 - retry connection errors (Apify behavior)
@@ -623,6 +640,9 @@ class ProxyConfiguration:
 
         Raises:
             ValueError: If proxy configuration is invalid
+            ConnectionError: If the platform proxy reports that this configuration
+                cannot be used (for example an unknown proxy group), with the
+                proxy's ``connectionError`` as the message (Apify parity)
         """
         if self._initialized:
             return
@@ -639,9 +659,10 @@ class ProxyConfiguration:
         if self._proxy_urls and self._new_url_function:
             raise ValueError('Cannot specify both proxy_urls and new_url_function')
 
-        # Apify-parity access check (best-effort, never fatal): GET
-        # {status_url}/?format=json *through the platform gateway* so the proxy
-        # handshake itself authenticates the run's credentials. Parity with the
+        # Apify-parity access check: GET {status_url}/?format=json *through the
+        # platform gateway* so the proxy handshake itself authenticates the run's
+        # credentials. A "connected": false answer raises like Apify's
+        # _check_access; timeouts and HTTP errors only warn. Parity with the
         # Apify SDKs (usesApifyProxy / uses_apify_proxy) and the JS Scrapely
         # port (platformEnv): the check only applies when this configuration
         # actually uses the platform proxy — BYOP-only configurations skip it
@@ -662,6 +683,8 @@ class ProxyConfiguration:
                 )
             else:
                 self._is_man_in_the_middle = status['isManInTheMiddle']
+                if not status['connected'] and status['fromPayload']:
+                    raise ConnectionError(status['connectionError'])
                 if not status['connected']:
                     logger.warning('Platform proxy access check failed: %s', status['connectionError'])
                 else:

{scrapely_python-0.1.1.dist-info → scrapely_python-0.1.2.dist-info}/METADATA

@@ -1,16 +1,15 @@
 Metadata-Version: 2.4
 Name: scrapely-python
-Version: 0.1.1
+Version: 0.1.2
 Summary: Scrapely Python SDK - A toolkit for building Actors on the Scrapely platform
 Author-email: Scrapely <support@scrape.ly>
-License: MIT
+License-Expression: MIT
 Project-URL: Homepage, https://scrape.ly
 Project-URL: Documentation, https://docs.scrape.ly
 Project-URL: Repository, https://github.com/SandeepDev1/scrapely
 Keywords: scraping,actor,serverless,apify,scrapely
 Classifier: Development Status :: 3 - Alpha
 Classifier: Intended Audience :: Developers
-Classifier: License :: OSI Approved :: MIT License
 Classifier: Operating System :: OS Independent
 Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.11
@@ -22,7 +21,7 @@ Description-Content-Type: text/markdown
 Requires-Dist: httpx>=0.25.0
 Requires-Dist: crawlee<2.0.0,>=1.8.0
 Requires-Dist: websockets>=12.0
-Requires-Dist: scrapely-client-python<2,>=1.0.0
+Requires-Dist: scrapely-client-python<2,>=1.0.2
 Requires-Dist: pydantic>=2.0.0
 Requires-Dist: cryptography>=42.0.0
 Provides-Extra: dev

{scrapely_python-0.1.1.dist-info → scrapely_python-0.1.2.dist-info}/RECORD

@@ -1,6 +1,6 @@
-scrapely/__init__.py,sha256=EqZTSyL6fkCFXkbp2iej7Umi6y_F19aC7MqxOaUbB68,5745
+scrapely/__init__.py,sha256=qBRp9YUgItQprvVgBepI-5VWc6OuQ2LQjCdSp5Xviv0,5745
 scrapely/_browser_locations.py,sha256=8QVVnkms2bDn3_LDYgYDR59MvFx8757l_JYnZ7xb3oI,145269
-scrapely/actor.py,sha256=WzvD7gvCKRMyIN_Ei_I-MfIBL7Darn_B40GQJIgAn04,74960
+scrapely/actor.py,sha256=wNifzbcvfH3wt0-aSaesJN_uZXNREhNfTgiORDGz7IM,75081
 scrapely/browser_configuration.py,sha256=jAZapNSHLFW05B96df4wNBtcyiiU9DA3ml0yTL1FwfA,5770
 scrapely/charging.py,sha256=ckqsGbGuzpdw5PgEt6Apx3cwR46sA8V2WwUDSqSSfz8,46887
 scrapely/charging_client.py,sha256=tDtlPwLSVFvoRa2NBgNjyfHkAUM77SSIyYT2ACgpDm8,3710
@@ -9,7 +9,7 @@ scrapely/errors.py,sha256=DCAlVvE3jToZuPEG2HVDL7i8D8hLboomQ5JpSOcsXoQ,4107
 scrapely/log.py,sha256=iVKoQdMLQXEWz1STzbSXrNUxsS_FNwiIdqoEg7CzeqU,1440
 scrapely/patch_browser_pool.py,sha256=sRAjuzKoOEgn0X8695c_9qioRALzkTWXoDlorVnK9jg,37426
 scrapely/pricing_types.py,sha256=J6wCddZd-U1MYeOtaaEdeHSsp1HBgiQLfvSInBOvhIo,5305
-scrapely/proxy_configuration.py,sha256=lvfOvuR8tOE9uiMOS_nw6TEj11xy7GYGa56iYvPwtmA,34654
+scrapely/proxy_configuration.py,sha256=i6icNhei8deVK9KgCC_hdeTuSawh8WiMUOrgjvkps1s,35845
 scrapely/py.typed,sha256=XtWlywqUjc6x_M2gAXlDtYXbAsYkCjKZEjFyJvaHvm4,105
 scrapely/types.py,sha256=ceVuHvhYbPai2CL3vFmXLTMoGaMR27IefJFgWhzd7ws,8900
 scrapely/webhook.py,sha256=4WOM42ypOSOs4lKtP5ofxAh6lX-7c8p5U9lOfuFzx-g,2830
@@ -37,7 +37,7 @@ scrapely/storage/_file_system/_key_value_store_client.py,sha256=4Ipgbh4qEqAkwQJ9
 scrapely/storage/_file_system/_storage_client.py,sha256=aaGJ4P0QAtIgEMl6yHpuobeWFD5M4rMC5HwhOOvuGuo,1794
 scrapely/storage_clients/__init__.py,sha256=cs1YX7JV2_GKBXeiiJgkQFali08b5_y_jH6Yw7iROAA,656
 scrapely/storages/__init__.py,sha256=4PAELhJNZ3exJ7Uouicn-bCQ58NafKjhLBixQ8ATQc4,173
-scrapely_python-0.1.1.dist-info/METADATA,sha256=MhoybcGCeSHaUbNsCX-tIuRZ5OYvmZRG8ysPNzunnPg,3144
-scrapely_python-0.1.1.dist-info/WHEEL,sha256=YVMoNqKzERt-wjUZwJ33xBGAwnFl-4cqbYkTtWa4itE,91
-scrapely_python-0.1.1.dist-info/top_level.txt,sha256=T4B9UMrs4q99QSAFyD1zdXhl40KnLhBKqrDkPhOTsmY,9
-scrapely_python-0.1.1.dist-info/RECORD,,
+scrapely_python-0.1.2.dist-info/METADATA,sha256=j0LJpAvLe7XGLBaGlKJkCq564TBZewFlGUDPgCbTB5U,3104
+scrapely_python-0.1.2.dist-info/WHEEL,sha256=YVMoNqKzERt-wjUZwJ33xBGAwnFl-4cqbYkTtWa4itE,91
+scrapely_python-0.1.2.dist-info/top_level.txt,sha256=T4B9UMrs4q99QSAFyD1zdXhl40KnLhBKqrDkPhOTsmY,9
+scrapely_python-0.1.2.dist-info/RECORD,,