scrapely-python 0.1.26__py3-none-any.whl → 0.1.28__py3-none-any.whl

scrapely/__init__.py

@@ -19,7 +19,7 @@ from typing import TYPE_CHECKING, Any
 # the Crawlee patches are applied when Crawlee's browser modules load.
 from .patch_browser_pool import set_cdp_url  # noqa: F401
 
-# Crawlee is NOT imported here: it costs ~1 s per run under gVisor and many
+# Crawlee is NOT imported here: it costs ~1 s per run on the platform and many
 # actors never touch it. Everything that needs Crawlee (the re-exported Crawlee
 # classes, storage clients, events, request loaders) is resolved on first
 # attribute access by ``__getattr__`` below; ``Actor`` imports what it needs on
@@ -227,7 +227,7 @@ if TYPE_CHECKING:  # Static analysers and IDEs see the lazy names as plain impor
         set_api_storage,
     )
 
-__version__ = '0.1.26'
+__version__ = '0.1.28'
 __all__ = [
     # Core
     'Actor',

scrapely/actor.py

@@ -68,7 +68,7 @@ if TYPE_CHECKING:
 logger = logging.getLogger(__name__)
 
 # Names that need Crawlee, resolved on first use instead of at import: importing
-# Crawlee costs ~1 s per run under gVisor. They stay module attributes
+# Crawlee costs ~1 s per run on the platform. They stay module attributes
 # (``scrapely.actor.Dataset`` works and can be monkeypatched): ``_lazy`` returns
 # the module-global value when one is set, else imports and caches it.
 _LAZY_ATTRS: Dict[str, Tuple[str, str]] = {

scrapely/patch_browser_pool.py

@@ -27,10 +27,10 @@ Mirrors the Node SDK ``patch-browser-pool.ts``:
 - Unpatched builds (``SCRAPELY_BROWSER_PATCH=0``, set by the worker when the
   actor's browser patch was off at build time) keep their browser binaries:
   every patched entry point falls through to the original local launch, with
-  gVisor-safe Chromium args under ``SCRAPELY_CONTAINER_RUNTIME=runsc``.
+  sandbox-safe Chromium args when running on the platform.
 - Soft no-op when Playwright/Crawlee browsers are not installed (e.g. HTTP-only
   actors).
-- Crawlee is never imported by this module (it costs ~1 s under gVisor). The
+- Crawlee is never imported by this module (it costs ~1 s on the platform). The
   Crawlee patches are applied at once when Crawlee's browser / Playwright
   crawler modules are already loaded, otherwise by a post-import hook the
   moment those modules finish loading — before actor code can use the classes.
@@ -121,8 +121,8 @@ def _browser_patch_disabled() -> bool:
     return os.environ.get('SCRAPELY_BROWSER_PATCH') == '0'
 
 
-# gVisor lacks crashpad support; ReadonlyRootfs + tmpfs /tmp loses crash dirs.
-_GVISOR_CHROME_ARGS = (
+# Platform sandbox: no crash reporter; read-only rootfs + tmpfs /tmp loses crash dirs.
+_SANDBOX_CHROME_ARGS = (
     '--disable-breakpad',
     '--disable-crash-reporter',
     '--disable-dev-shm-usage',
@@ -130,13 +130,13 @@ _GVISOR_CHROME_ARGS = (
 )
 
 
-def _with_gvisor_args(options: Optional[dict]) -> dict:
-    """Copy of launch options with gVisor-safe Chromium args when under runsc."""
+def _with_sandbox_args(options: Optional[dict]) -> dict:
+    """Copy of launch options with sandbox-safe Chromium args when on the platform."""
     out = dict(options or {})
-    if os.environ.get('SCRAPELY_CONTAINER_RUNTIME') != 'runsc':
+    if not _is_at_home():
         return out
     args = list(out.get('args') or [])
-    for arg in _GVISOR_CHROME_ARGS:
+    for arg in _SANDBOX_CHROME_ARGS:
         if arg not in args:
             args.append(arg)
     out['args'] = args
@@ -678,7 +678,7 @@ def _patch_raw_browser_type(browser_type: Any, chromium: Any, name: str, is_asyn
         async def launch(**kwargs: Any) -> Any:
             browser_configuration = _take_browser_configuration(kwargs)
             if _browser_patch_disabled():
-                return await orig_launch(**_with_gvisor_args(kwargs))
+                return await orig_launch(**_with_sandbox_args(kwargs))
             cdp_url = _raw_launch_cdp_url(kwargs, browser_configuration)
             if not cdp_url:
                 if not _is_at_home():
@@ -691,7 +691,7 @@ def _patch_raw_browser_type(browser_type: Any, chromium: Any, name: str, is_asyn
         async def launch_persistent_context(*args: Any, **kwargs: Any) -> Any:
             browser_configuration = _take_browser_configuration(kwargs)
             if _browser_patch_disabled():
-                return await orig_lpc(*args, **_with_gvisor_args(kwargs))
+                return await orig_lpc(*args, **_with_sandbox_args(kwargs))
             cdp_url = _raw_launch_cdp_url(kwargs, browser_configuration)
             if not cdp_url:
                 if not _is_at_home():
@@ -710,7 +710,7 @@ def _patch_raw_browser_type(browser_type: Any, chromium: Any, name: str, is_asyn
         def launch(**kwargs: Any) -> Any:
             browser_configuration = _take_browser_configuration(kwargs)
             if _browser_patch_disabled():
-                return orig_launch(**_with_gvisor_args(kwargs))
+                return orig_launch(**_with_sandbox_args(kwargs))
             cdp_url = _raw_launch_cdp_url(kwargs, browser_configuration)
             if not cdp_url:
                 if not _is_at_home():
@@ -722,7 +722,7 @@ def _patch_raw_browser_type(browser_type: Any, chromium: Any, name: str, is_asyn
         def launch_persistent_context(*args: Any, **kwargs: Any) -> Any:
             browser_configuration = _take_browser_configuration(kwargs)
             if _browser_patch_disabled():
-                return orig_lpc(*args, **_with_gvisor_args(kwargs))
+                return orig_lpc(*args, **_with_sandbox_args(kwargs))
             cdp_url = _raw_launch_cdp_url(kwargs, browser_configuration)
             if not cdp_url:
                 if not _is_at_home():
@@ -847,7 +847,7 @@ def _apply_pyppeteer_patch() -> None:
             options = dict(options)
             browser_configuration = _take_browser_configuration(options) or browser_configuration
         if _browser_patch_disabled():
-            return await _orig_launch(_with_gvisor_args({**(options or {}), **kwargs}))
+            return await _orig_launch(_with_sandbox_args({**(options or {}), **kwargs}))
         cdp_url = get_cdp_url()
         if not cdp_url:
             if not _is_at_home():
@@ -876,7 +876,7 @@ def _apply_pyppeteer_patch() -> None:
 
 # ─── Crawlee patches without importing Crawlee ───────────────────────────────
 #
-# Importing Crawlee costs ~1 s per run under gVisor, so this module never
+# Importing Crawlee costs ~1 s per run on the platform, so this module never
 # imports it. The Crawlee patches mutate classes in place, so they only have to
 # run before actor code uses those classes: they are applied immediately when
 # the defining modules are already loaded, otherwise by a post-import hook the

scrapely/solver.py

@@ -88,7 +88,7 @@ DEGRADED_STATUS_MESSAGE = "Site protection temporarily unavailable, results may
 # Codes the platform answers with when a vendor cannot be solved right now.
 UNAVAILABLE_CODES = ("vendor_degraded", "vendor_down", "provider_unavailable", "vendor_unsupported", "solver_not_enabled")
 
-_COMPAT_SUFFIXES = ("/compat/v1", "/compat/xhrdev")
+_COMPAT_SUFFIXES = ("/compat/v1",)
 _STATUS_TTL_SECS = 15.0
 _MAX_BODY = 4 * 1024 * 1024
 

{scrapely_python-0.1.26.dist-info → scrapely_python-0.1.28.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: scrapely-python
-Version: 0.1.26
+Version: 0.1.28
 Summary: Scrapely Python SDK - A toolkit for building Actors on the Scrapely platform
 Author-email: Scrapely <support@scrape.ly>
 License-Expression: MIT AND Apache-2.0

{scrapely_python-0.1.26.dist-info → scrapely_python-0.1.28.dist-info}/RECORD

@@ -1,6 +1,6 @@
-scrapely/__init__.py,sha256=z-CyEPAlCd07V32Wnv8bNeSubMhXAuoejvSW7-XG4co,9661
+scrapely/__init__.py,sha256=xBnYBJZQ88Yr59AVLs7QO2rAUt7cDT0AzK8VgnyQ-Vs,9664
 scrapely/_browser_locations.py,sha256=8QVVnkms2bDn3_LDYgYDR59MvFx8757l_JYnZ7xb3oI,145269
-scrapely/actor.py,sha256=1SBHZYBRLmU4k_erjKL7yRfP0EgR7ZKXefAK8wWKRz0,88004
+scrapely/actor.py,sha256=CBdphCMl6zf72br8if8zcFIXsiXgDa0-VYrWuOTWgs4,88007
 scrapely/browser_configuration.py,sha256=jAZapNSHLFW05B96df4wNBtcyiiU9DA3ml0yTL1FwfA,5770
 scrapely/cdp_retry.py,sha256=hdTJEAIKhApM7IOk1DrQEl8NoflSJSmqd4RbNJ2IsgI,8540
 scrapely/charging.py,sha256=ns-kuUJtl8hBI4-3iBisC62BggJG67yYOa1yVVItoUI,83800
@@ -8,11 +8,11 @@ scrapely/charging_client.py,sha256=-4q-Ww7gu9EZXpM9QIn1Vd4vwAXmgL_WEKqaQIx3ViA,3
 scrapely/configuration.py,sha256=Of-lKm7hoPGUQO8d08kV7fAy2co9QTfb8Of_8ds5PSA,17028
 scrapely/errors.py,sha256=DCAlVvE3jToZuPEG2HVDL7i8D8hLboomQ5JpSOcsXoQ,4107
 scrapely/log.py,sha256=iVKoQdMLQXEWz1STzbSXrNUxsS_FNwiIdqoEg7CzeqU,1440
-scrapely/patch_browser_pool.py,sha256=R-1QDKa-ZvfdQ9wgyA8JCtmQqD9v7dOsaCZnSUoyf7U,42214
+scrapely/patch_browser_pool.py,sha256=eZck-JSOvCuXdoGQaX6xUsKVGzolDlB4uKiBSBrFyr8,42190
 scrapely/pricing_types.py,sha256=J6wCddZd-U1MYeOtaaEdeHSsp1HBgiQLfvSInBOvhIo,5305
 scrapely/proxy_configuration.py,sha256=mQhUjXnqM1PcUrnRmXxVWdLnlaRlb8Je4K2j54lC6YQ,38129
 scrapely/py.typed,sha256=XtWlywqUjc6x_M2gAXlDtYXbAsYkCjKZEjFyJvaHvm4,105
-scrapely/solver.py,sha256=QO7sl_M5AbXRqmCzUowNWG2kBHoXoH7CgtYwAdp36VM,35880
+scrapely/solver.py,sha256=RVoLNJTxniTI5SzDRpUdxe0ajzy4qqjuEmToTL1GRzA,35863
 scrapely/types.py,sha256=ceVuHvhYbPai2CL3vFmXLTMoGaMR27IefJFgWhzd7ws,8900
 scrapely/webhook.py,sha256=4WOM42ypOSOs4lKtP5ofxAh6lX-7c8p5U9lOfuFzx-g,2830
 scrapely/crypto/__init__.py,sha256=MbMPs7Eq_fGSVjOZxNts_JckDCUtxCVqdkHU0VCfN4M,481
@@ -39,10 +39,10 @@ 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-0.1.26.dist-info/licenses/LICENSE,sha256=pQGY0XwjIYh-a2HfawxHq1en53bXkFkDj2Q70jKWHr4,1088
-scrapely_python-0.1.26.dist-info/licenses/LICENSE-APACHE,sha256=N8m0FzMq8i4UbTgXZvKEZJSD-27OF8pWZYS_atz4kTU,11355
-scrapely_python-0.1.26.dist-info/licenses/NOTICE,sha256=DEsNrw6UdSIqdYYqMWVFgJnyt6jUivTD7iFU6YM_e1A,1122
-scrapely_python-0.1.26.dist-info/METADATA,sha256=mag9UA_ijOHkoBHTTvSWq2Nai46Okg-M-wkW0aAO-Hk,3487
-scrapely_python-0.1.26.dist-info/WHEEL,sha256=YVMoNqKzERt-wjUZwJ33xBGAwnFl-4cqbYkTtWa4itE,91
-scrapely_python-0.1.26.dist-info/top_level.txt,sha256=T4B9UMrs4q99QSAFyD1zdXhl40KnLhBKqrDkPhOTsmY,9
-scrapely_python-0.1.26.dist-info/RECORD,,
+scrapely_python-0.1.28.dist-info/licenses/LICENSE,sha256=pQGY0XwjIYh-a2HfawxHq1en53bXkFkDj2Q70jKWHr4,1088
+scrapely_python-0.1.28.dist-info/licenses/LICENSE-APACHE,sha256=N8m0FzMq8i4UbTgXZvKEZJSD-27OF8pWZYS_atz4kTU,11355
+scrapely_python-0.1.28.dist-info/licenses/NOTICE,sha256=DEsNrw6UdSIqdYYqMWVFgJnyt6jUivTD7iFU6YM_e1A,1122
+scrapely_python-0.1.28.dist-info/METADATA,sha256=J4RQA1RzD7fF1MbOd1NnJCPPAP79ZPJJuBkdKIUZ6E8,3487
+scrapely_python-0.1.28.dist-info/WHEEL,sha256=YVMoNqKzERt-wjUZwJ33xBGAwnFl-4cqbYkTtWa4itE,91
+scrapely_python-0.1.28.dist-info/top_level.txt,sha256=T4B9UMrs4q99QSAFyD1zdXhl40KnLhBKqrDkPhOTsmY,9
+scrapely_python-0.1.28.dist-info/RECORD,,