scrapely-python 0.1.27__tar.gz → 0.1.30__tar.gz

scrapely_python-0.1.30/CHANGELOG.md

@@ -0,0 +1,12 @@
+# Changelog
+
+All notable changes to this project will be documented in this file.
+
+## [0.1.30] - 2026-10-20
+
+- Removed `apify_margin_percentage` from the pricing info models; the API no longer returns it.
+- Added Apache-2.0 change notices to the storage clients, storage utils and log module derived from apify-sdk-python, and listed them in NOTICE.
+
+## [0.1.29] - 2026-10-19
+
+- Initial public release.

scrapely_python-0.1.30/MANIFEST.in

@@ -0,0 +1,2 @@
+include CHANGELOG.md
+prune tests

scrapely_python-0.1.30/NOTICE

@@ -0,0 +1,16 @@
+scrapely-python
+Copyright (c) 2025-2026 TeraShift GmbH (Scrape.ly)
+
+Licensed under the MIT License (see LICENSE), except for the files below,
+which are derived from apify-sdk-python (https://github.com/apify/apify-sdk-python),
+Copyright Apify Technologies s.r.o., and remain under the Apache License,
+Version 2.0 (see LICENSE-APACHE). They have been modified by TeraShift GmbH:
+
+  - src/scrapely/events/_types.py
+  - src/scrapely/events/event_manager.py
+  - src/scrapely/log.py
+  - src/scrapely/request_loaders/_scrapely_request_list.py
+  - src/scrapely/storage/_file_system/_key_value_store_client.py
+  - src/scrapely/storage/_file_system/_storage_client.py
+  - src/scrapely/storage/smart_scrapely_storage_client.py
+  - src/scrapely/storage/utils.py

{scrapely_python-0.1.27/src/scrapely_python.egg-info → scrapely_python-0.1.30}/PKG-INFO

@@ -1,13 +1,15 @@
 Metadata-Version: 2.4
 Name: scrapely-python
-Version: 0.1.27
+Version: 0.1.30
 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
 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
+Project-URL: Documentation, https://scrape.ly/docs/sdk-python
+Project-URL: Repository, https://github.com/scrapely-app/scrapely-python
+Project-URL: Issues, https://github.com/scrapely-app/scrapely-python/issues
+Project-URL: Changelog, https://github.com/scrapely-app/scrapely-python/blob/main/CHANGELOG.md
+Keywords: scraping,actor,serverless,scrapely
 Classifier: Development Status :: 3 - Alpha
 Classifier: Intended Audience :: Developers
 Classifier: Operating System :: OS Independent
@@ -43,7 +45,7 @@ Build Python Actors for the [Scrapely](https://scrape.ly) platform with lifecycl
 
 ## Install
 
-Scrapely requires Python 3.9 or newer.
+Scrapely requires Python 3.11 or newer.
 
 ```bash
 python -m pip install scrapely-python
@@ -81,12 +83,11 @@ if __name__ == "__main__":
 - Automatic remote Chromium CDP wiring through `CDP_URL`
 - Actor start, call, abort, reboot, and metamorph operations
 - Ad-hoc webhooks, proxy configuration, and charging/pricing types
-- Apify-compatible environment aliases and asynchronous client access
+- Legacy environment variable aliases and asynchronous client access
 
-See the [complete Python SDK guide](../docs/sdk-python.md) for lifecycle details, option objects, storage selection, events, CDP behavior, run control, environment variables, errors, and the full public export list.
+See the [complete Python SDK guide](https://scrape.ly/docs/sdk-python) for lifecycle details, option objects, storage selection, events, CDP behavior, run control, environment variables, errors, and the full public export list.
 
 ## License
 
-MIT (see [LICENSE](./LICENSE)), except for three files derived from the Apache-2.0 licensed
-[apify-sdk-python](https://github.com/apify/apify-sdk-python), which stay under the Apache License 2.0
+MIT (see [LICENSE](./LICENSE)), except for three files that stay under the Apache License 2.0
 (see [LICENSE-APACHE](./LICENSE-APACHE) and [NOTICE](./NOTICE)).

{scrapely_python-0.1.27 → scrapely_python-0.1.30}/README.md

@@ -4,7 +4,7 @@ Build Python Actors for the [Scrapely](https://scrape.ly) platform with lifecycl
 
 ## Install
 
-Scrapely requires Python 3.9 or newer.
+Scrapely requires Python 3.11 or newer.
 
 ```bash
 python -m pip install scrapely-python
@@ -42,12 +42,11 @@ if __name__ == "__main__":
 - Automatic remote Chromium CDP wiring through `CDP_URL`
 - Actor start, call, abort, reboot, and metamorph operations
 - Ad-hoc webhooks, proxy configuration, and charging/pricing types
-- Apify-compatible environment aliases and asynchronous client access
+- Legacy environment variable aliases and asynchronous client access
 
-See the [complete Python SDK guide](../docs/sdk-python.md) for lifecycle details, option objects, storage selection, events, CDP behavior, run control, environment variables, errors, and the full public export list.
+See the [complete Python SDK guide](https://scrape.ly/docs/sdk-python) for lifecycle details, option objects, storage selection, events, CDP behavior, run control, environment variables, errors, and the full public export list.
 
 ## License
 
-MIT (see [LICENSE](./LICENSE)), except for three files derived from the Apache-2.0 licensed
-[apify-sdk-python](https://github.com/apify/apify-sdk-python), which stay under the Apache License 2.0
+MIT (see [LICENSE](./LICENSE)), except for three files that stay under the Apache License 2.0
 (see [LICENSE-APACHE](./LICENSE-APACHE) and [NOTICE](./NOTICE)).

{scrapely_python-0.1.27 → scrapely_python-0.1.30}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "scrapely-python"
-version = "0.1.27"
+version = "0.1.30"
 description = "Scrapely Python SDK - A toolkit for building Actors on the Scrapely platform"
 readme = "README.md"
 license = "MIT AND Apache-2.0"
@@ -22,7 +22,7 @@ classifiers = [
     "Topic :: Software Development :: Libraries :: Python Modules",
     "Topic :: Internet :: WWW/HTTP :: Dynamic Content",
 ]
-keywords = ["scraping", "actor", "serverless", "apify", "scrapely"]
+keywords = ["scraping", "actor", "serverless", "scrapely"]
 requires-python = ">=3.11"
 dependencies = [
     "httpx>=0.25.0",
@@ -47,8 +47,10 @@ dev = [
 
 [project.urls]
 Homepage = "https://scrape.ly"
-Documentation = "https://docs.scrape.ly"
-Repository = "https://github.com/SandeepDev1/scrapely"
+Documentation = "https://scrape.ly/docs/sdk-python"
+Repository = "https://github.com/scrapely-app/scrapely-python"
+Issues = "https://github.com/scrapely-app/scrapely-python/issues"
+Changelog = "https://github.com/scrapely-app/scrapely-python/blob/main/CHANGELOG.md"
 
 [tool.setuptools.packages.find]
 where = ["src"]

{scrapely_python-0.1.27 → scrapely_python-0.1.30}/src/scrapely/__init__.py

@@ -227,13 +227,13 @@ if TYPE_CHECKING:  # Static analysers and IDEs see the lazy names as plain impor
         set_api_storage,
     )
 
-__version__ = '0.1.27'
+__version__ = '0.1.30'
 __all__ = [
     # Core
     'Actor',
     'Configuration',
     'set_cdp_url',
-    # ApifyClient
+    # Charging client
     'create_charging_client',
     'USES_PUSH_DATA_INTERCEPTION',
     'push_data_charging_context',

{scrapely_python-0.1.27 → scrapely_python-0.1.30}/src/scrapely/actor.py

@@ -204,7 +204,7 @@ def _run_options_params(
     force_permission_level: Optional[Any] = None,
     webhooks: Optional[List[Any]] = None,
 ) -> Dict[str, Any]:
-    """Build the Apify-style run-option query parameters for run-start requests.
+    """Build the run-option query parameters for run-start requests.
 
     Server wire contract: the POST body of a run-start request is the run
     input itself; run options (build/timeout/memory/maxTotalChargeUsd/
@@ -319,9 +319,9 @@ class Actor(metaclass=_ActorMeta):
         exit_process: Optional[bool] = None,
         event_listeners_timeout: Optional[timedelta] = None,
     ):
-        """Apify parity: ``async with Actor(configure_logging=False):`` as well as ``async with Actor:``.
+        """Support ``async with Actor(configure_logging=False):`` as well as ``async with Actor:``.
 
-        ``options`` is a configuration dict (Apify passes a Configuration here). The context
+        ``options`` is a configuration dict. The context
         manager never calls ``sys.exit`` (``exit_process`` is accepted for signature parity),
         and ``event_listeners_timeout`` is accepted but not applied.
         """
@@ -422,7 +422,7 @@ class Actor(metaclass=_ActorMeta):
                         'run_id': run_id,
                     })
                 elif not cls._config.get('is_at_home'):
-                    # Local development without credentials: Smart → FS (Apify docs shape)
+                    # Local development without credentials: Smart → FS
                     logger.info('[Scrapely] Initializing local file-system storage')
                     smart = _lazy('set_api_storage')()
                 else:
@@ -635,7 +635,7 @@ class Actor(metaclass=_ActorMeta):
     ) -> None:
         """Exit the Actor with an optional terminal status message.
 
-        Apify-Python parity: supports keyword-only ``status_message`` and
+        Supports keyword-only ``status_message`` and
         ``exit_code`` arguments (``await Actor.exit(status_message="Done")``).
         The pre-existing positional forms (``Actor.exit("message")`` /
         ``Actor.exit(ExitOptions(...))``) remain supported; explicit keyword
@@ -678,7 +678,7 @@ class Actor(metaclass=_ActorMeta):
     ) -> None:
         """Fail the Actor run (exit code defaults to ``1``).
 
-        Apify-Python parity: supports keyword-only ``status_message``,
+        Supports keyword-only ``status_message``,
         ``exit_code`` and ``exception`` arguments
         (``await Actor.fail(status_message="Boom", exception=e)``). When
         ``exception`` is given its traceback is logged.
@@ -724,7 +724,7 @@ class Actor(metaclass=_ActorMeta):
     def is_paying(cls) -> bool:
         """Whether the user who started this Actor run is a paying user.
 
-        Mirrors the Node SDK's ``Actor.isPaying`` (``APIFY_USER_IS_PAYING`` /
+        Reads ``APIFY_USER_IS_PAYING`` (config key
         ``user_is_paying``). Note the platform currently sets this from the
         subscription plan only — PAYG balance alone does not flip it.
         """
@@ -833,8 +833,8 @@ class Actor(metaclass=_ActorMeta):
     ) -> ChargeResult:
         """Push data to the default dataset, charging pay-per-event items.
 
-        Matches the Apify Python SDK's ``Actor.push_data(data, charged_event_name=...)``:
-        when the Actor uses pay-per-event pricing, each pushed item is charged for the
+        Usage: ``Actor.push_data(data, charged_event_name=...)``.
+        When the Actor uses pay-per-event pricing, each pushed item is charged for the
         default-dataset-item event (if priced) and, when ``charged_event_name`` is
         provided, for that explicit event as well. Actors without pay-per-event
         pricing push without charging and receive an empty result.
@@ -842,7 +842,7 @@ class Actor(metaclass=_ActorMeta):
         Args:
             data: Item or list of items to push to the default dataset.
             charged_event_name: Optional pricing event to charge for each pushed item
-                (Apify Python SDK name; also accepted positionally).
+                (also accepted positionally).
             event_name: Deprecated alias of ``charged_event_name`` (JS ``eventName``),
                 kept so existing callers keep working.
             charges: Several events per row, charged atomically: ``{'result': 1, 'page': 3}``
@@ -872,7 +872,7 @@ class Actor(metaclass=_ActorMeta):
             raise ValueError('Pass only one of charged_event_name / event_name')
         event_name = charged_event_name if charged_event_name is not None else event_name
         # Exact synthetic names only (JS parity) — a prefix heuristic would block
-        # real priced events like 'apify-foo'.
+        # real priced events.
         if event_name and is_default_dataset_item_event(event_name):
             raise ValueError(f"Cannot charge for synthetic event '{event_name}' manually")
 
@@ -1168,7 +1168,7 @@ class Actor(metaclass=_ActorMeta):
         Args:
             message: The status message to set.
             options: Optional settings (``is_status_message_terminal``, ``level``).
-            is_terminal: Apify Python SDK-style terminal flag. When given, it
+            is_terminal: Terminal flag. When given, it
                 overrides ``options.is_status_message_terminal``.
 
         Returns:
@@ -1276,7 +1276,7 @@ class Actor(metaclass=_ActorMeta):
 
     @classmethod
     async def _persist_terminal_status_message(cls, message: Optional[str], timeout_secs: float = 10.0) -> None:
-        """Store the exit/fail status message on the run as terminal (Apify parity).
+        """Store the exit/fail status message on the run as terminal.
 
         Runs before teardown so it lands while the run still exists and the
         process is alive, on success and failure alike. Best effort: a failed
@@ -1322,7 +1322,7 @@ class Actor(metaclass=_ActorMeta):
 
     @classmethod
     def _env_config(cls) -> Configuration:
-        # Apify parity: is_at_home() and friends read the environment even
+        # is_at_home() and friends read the environment even
         # before init(). Doesn't set cls._config, so init() state is unchanged.
         return cls._config if cls._config is not None else Configuration.get_global_config()
 
@@ -1335,14 +1335,14 @@ class Actor(metaclass=_ActorMeta):
         return ApiClient(api_url=api_url, token=api_token, run_id=run_id)
 
     # -------------------------------------------------------------------------
-    # Properties for Apify SDK compatibility
+    # Client and configuration properties
     # -------------------------------------------------------------------------
 
     @classmethod
     def _get_apify_client(cls) -> 'ApifyClientAsync':
         """Get or create a cached charging-configured ``ApifyClientAsync`` instance.
         
-        This client provides full ApifyClient compatibility including
+        This client provides the full ApifyClient API including
         actor(), dataset(), key_value_store(), etc. methods.
         """
         if cls._apify_client_instance is None:
@@ -1372,9 +1372,9 @@ class Actor(metaclass=_ActorMeta):
 
     @_ClassProperty
     def apify_client(cls) -> 'ApifyClientAsync':
-        """Get a charging-configured ``ApifyClientAsync`` for Apify SDK compatibility.
+        """Get a charging-configured ``ApifyClientAsync``.
         
-        This provides full ApifyClient compatibility including:
+        This provides the full ApifyClient API including:
         - client.actor('id').call()
         - client.dataset('id').list_items()
         - client.key_value_store('id').get_record('key')
@@ -1390,7 +1390,7 @@ class Actor(metaclass=_ActorMeta):
 
     @_ClassProperty
     def config(cls) -> Configuration:
-        """Return the Configuration instance (Apify SDK ``Actor.config`` parity)."""
+        """Return the Configuration instance."""
         return cls.get_config()
 
     @_ClassProperty
@@ -1530,7 +1530,7 @@ class Actor(metaclass=_ActorMeta):
         client = cls.new_client(token=opts.token) if opts.token else cls.new_client()
 
         # Wire format (server contract): the POST body is the run input itself;
-        # run options are Apify-style query parameters. Inline webhooks ride
+        # run options are query parameters. Inline webhooks ride
         # along base64-encoded in the `webhooks` query param.
         params = _run_options_params(
             build=opts.build,
@@ -1615,7 +1615,7 @@ class Actor(metaclass=_ActorMeta):
         """Poll ``GET /v2/actor-runs/:runId`` until the run reaches a terminal status.
 
         When ``timeout_seconds`` elapses first, the run is aborted (best-effort)
-        and ``TimeoutError`` is raised — Node SDK (apify-client) parity.
+        and ``TimeoutError`` is raised.
         """
         deadline = time.monotonic() + timeout_seconds if timeout_seconds is not None else None
 
@@ -1693,7 +1693,7 @@ class Actor(metaclass=_ActorMeta):
 
         # Wire format (server contract): the POST body is the task input
         # override (merged over the task's saved input by the server); run
-        # options are Apify-style query parameters. The task-run endpoint is
+        # options are query parameters. The task-run endpoint is
         # /v2/actor-tasks/:taskId/runs (there is no /v2/tasks/:taskId/runs).
         params = _run_options_params(
             build=opts.build,