scrapely-python 0.1.12__tar.gz → 0.1.14__tar.gz

{scrapely_python-0.1.12/src/scrapely_python.egg-info → scrapely_python-0.1.14}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: scrapely-python
-Version: 0.1.12
+Version: 0.1.14
 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-0.1.12 → scrapely_python-0.1.14}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "scrapely-python"
-version = "0.1.12"
+version = "0.1.14"
 description = "Scrapely Python SDK - A toolkit for building Actors on the Scrapely platform"
 readme = "README.md"
 license = "MIT AND Apache-2.0"

{scrapely_python-0.1.12 → scrapely_python-0.1.14}/src/scrapely/__init__.py

@@ -132,7 +132,7 @@ from .errors import (
     raise_for_status,
 )
 
-__version__ = '0.1.12'
+__version__ = '0.1.14'
 __all__ = [
     # Core
     'Actor',

{scrapely_python-0.1.12 → scrapely_python-0.1.14}/src/scrapely/actor.py

@@ -213,6 +213,14 @@ class _ClassProperty:
         return self.func(owner)
 
 
+def _system_exit_code(exc: Any) -> int:
+    """Process exit code carried by a SystemExit (sys.exit semantics)."""
+    code = getattr(exc, 'code', None)
+    if code is None:
+        return 0
+    return code if isinstance(code, int) else 1
+
+
 class _ActorMeta(type):
     """Metaclass enabling async with Actor: by delegating to classmethods."""
 
@@ -226,7 +234,11 @@ class _ActorMeta(type):
         return cls
 
     async def __aexit__(cls, exc_type: Any, exc_val: Any, exc_tb: Any) -> None:
-        if exc_type is not None:
+        if exc_type is not None and issubclass(exc_type, SystemExit):
+            # Actor.exit()/fail()/sys.exit() inside `async with Actor:` is a deliberate
+            # exit, not a failure (Apify parity): keep its exit code, log nothing.
+            cls._exit_code = _system_exit_code(exc_val)
+        elif exc_type is not None:
             cls._exit_code = 1
             logger.error(f'Actor failed: {exc_val}')
         await cls._shutdown()
@@ -286,7 +298,10 @@ class Actor(metaclass=_ActorMeta):
 
     async def __aexit__(self, exc_type: Any, exc_val: Any, exc_tb: Any) -> None:
         cls = type(self)
-        if exc_type is not None:
+        if exc_type is not None and issubclass(exc_type, SystemExit):
+            # Deliberate exit via Actor.exit()/fail()/sys.exit(), not a failure.
+            cls._exit_code = _system_exit_code(exc_val)
+        elif exc_type is not None:
             cls._exit_code = 1
             logger.error(f'Actor failed: {exc_val}')
         else:
@@ -652,12 +667,12 @@ class Actor(metaclass=_ActorMeta):
 
     @classmethod
     def is_at_home(cls) -> bool:
-        return cls._config.get('is_at_home', False) if cls._config else False
+        return bool(cls._env_config().get('is_at_home', False))
 
     @classmethod
     def is_standby(cls) -> bool:
         """Return True when the actor run was started in standby mode (Node SDK ``isStandby`` parity)."""
-        return (cls._config.get('meta_origin') == 'STANDBY') if cls._config else False
+        return cls._env_config().get('meta_origin') == 'STANDBY'
 
     @_ClassProperty
     def is_paying(cls) -> bool:
@@ -721,7 +736,7 @@ class Actor(metaclass=_ActorMeta):
             return cast(Dict[str, Any], cls._cached_input)
 
         kvs = await cls.open_key_value_store()
-        input_key = cls._config.get('input_key', 'INPUT') if cls._config else 'INPUT'
+        input_key = cls._env_config().get('input_key', 'INPUT')
         result = await kvs.get_value(input_key)
         if isinstance(result, dict):
             result = decrypt_secret_inputs(result)
@@ -848,7 +863,7 @@ class Actor(metaclass=_ActorMeta):
                 name = name_or_id
 
         if alias is not None:
-            storages_json = cls._config.get('actor_storages_json') if cls._config else None
+            storages_json = cls._env_config().get('actor_storages_json')
             resolved = resolve_alias_from_storages_json(alias, storage_type, storages_json)
             if resolved is not None:
                 id = resolved
@@ -1129,6 +1144,12 @@ class Actor(metaclass=_ActorMeta):
             cls._config = Configuration.get_global_config()
         return cls._config
 
+    @classmethod
+    def _env_config(cls) -> Configuration:
+        # Apify parity: 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()
+
     @classmethod
     def new_client(cls, token: Optional[str] = None, base_url: Optional[str] = None) -> ApiClient:
         config = cls.get_config()

{scrapely_python-0.1.12 → scrapely_python-0.1.14}/src/scrapely/configuration.py

@@ -266,6 +266,22 @@ class Configuration:
         value = self.get('standby_port')
         return value if isinstance(value, int) else 4321
 
+    @property
+    def web_server_port(self) -> int:
+        """Apify name for the container port (ACTOR_WEB_SERVER_PORT, default 4321)."""
+        value = self.get('container_port')
+        return value if isinstance(value, int) else 4321
+
+    @property
+    def web_server_url(self) -> str:
+        """Apify name for the container URL (ACTOR_WEB_SERVER_URL)."""
+        return cast(str, self.get('container_url'))
+
+    @property
+    def standby_url(self) -> str:
+        """Public standby URL (ACTOR_STANDBY_URL), falling back to the container URL."""
+        return cast(str, self.get('standby_url') or self.get('container_url'))
+
     @property
     def token(self) -> Optional[str]:
         """Get the authentication token."""

{scrapely_python-0.1.12 → scrapely_python-0.1.14}/src/scrapely/storage/api_client.py

@@ -5,6 +5,7 @@ Base API client for Scrapely storage operations.
 from __future__ import annotations
 
 import asyncio
+import json
 import logging
 import os
 import random
@@ -311,6 +312,10 @@ class ApiClient:
         request_headers = dict(headers or {})
         if json_data is not None and 'Content-Type' not in request_headers:
             request_headers['Content-Type'] = 'application/json'
+        if json_data is not None:
+            # Apify/crawlee parity: unknown objects (datetimes, LLM result
+            # objects, ...) are stored as strings instead of raising.
+            content = json.dumps(json_data, ensure_ascii=False, default=str).encode('utf-8')
 
         if idempotent is None:
             idempotent = method.upper() in IDEMPOTENT_METHODS
@@ -343,7 +348,6 @@ class ApiClient:
                     method,
                     url,
                     params=params,
-                    json=json_data,
                     content=content,
                     headers=request_headers or None,
                     **request_kwargs,

{scrapely_python-0.1.12 → scrapely_python-0.1.14}/src/scrapely/storage/client.py

@@ -6,6 +6,7 @@ Ports the patterns from the TypeScript ApiStorageClient.
 
 from __future__ import annotations
 
+import os
 import logging
 import re
 import uuid
@@ -358,7 +359,14 @@ class ApiStorageClient:
         return cast(Dict[str, Any], await self.api_client.get_json(f'/v2/key-value-stores/{store_id}/keys', params=params))
 
     def kvs_public_url(self, store_id: str, key: str) -> str:
-        return f'{self.api_client.api_url}{kvs_record_path(store_id, key)}'
+        # api_url is the container-internal API address; links handed to users
+        # must use the public base URL the platform injects.
+        public_base = (
+            os.environ.get('APIFY_API_PUBLIC_BASE_URL')
+            or os.environ.get('ACTOR_API_PUBLIC_BASE_URL')
+            or self.api_client.api_url
+        ).rstrip('/')
+        return f'{public_base}{kvs_record_path(store_id, key)}'
 
     # --- Request queue resource ops ---
 

{scrapely_python-0.1.12 → scrapely_python-0.1.14/src/scrapely_python.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: scrapely-python
-Version: 0.1.12
+Version: 0.1.14
 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-0.1.12 → scrapely_python-0.1.14}/tests/test_apify_parity.py

@@ -9,6 +9,7 @@ update them when targeting a newer Apify version.
 from __future__ import annotations
 
 import importlib
+import json
 
 import pytest
 
@@ -83,3 +84,52 @@ def test_configuration_has_apify_global_accessor() -> None:
     from scrapely import Configuration
 
     assert Configuration.get_global_configuration() is Configuration.get_global_config()
+
+
+def test_configuration_has_apify_web_server_names(monkeypatch: pytest.MonkeyPatch) -> None:
+    from scrapely import Configuration
+
+    monkeypatch.setenv('ACTOR_WEB_SERVER_PORT', '5555')
+    monkeypatch.setenv('ACTOR_WEB_SERVER_URL', 'http://10.0.0.1:5555')
+    config = Configuration()
+    assert config.web_server_port == 5555
+    assert config.web_server_url == 'http://10.0.0.1:5555'
+    assert config.standby_url == 'http://10.0.0.1:5555'
+
+
+def test_is_at_home_reads_env_before_init(monkeypatch: pytest.MonkeyPatch) -> None:
+    # Apify templates pick their port at import time, before Actor.init().
+    from scrapely import Actor, Configuration
+
+    monkeypatch.setenv('APIFY_IS_AT_HOME', '1')
+    monkeypatch.setenv('APIFY_META_ORIGIN', 'STANDBY')
+    Configuration.reset_global_state()
+    monkeypatch.setattr(Actor, '_config', None)
+    try:
+        assert Actor.is_at_home() is True
+        assert Actor.is_standby() is True
+        assert Actor._config is None
+    finally:
+        Configuration.reset_global_state()
+
+
+async def test_json_body_stringifies_unknown_objects() -> None:
+    import httpx
+
+    from scrapely.storage.api_client import ApiClient
+
+    class AgentOutput:
+        def __str__(self) -> str:
+            return 'the answer'
+
+    bodies: list = []
+
+    def handler(request: httpx.Request) -> httpx.Response:
+        bodies.append((request.headers['content-type'], request.content))
+        return httpx.Response(201, json={})
+
+    client = ApiClient('http://api.test', 'st_x', 'run1', transport=httpx.MockTransport(handler))
+    await client.post_json('/v2/datasets/d1/items', [{'answer': AgentOutput(), 'q': 'café'}])
+    content_type, body = bodies[0]
+    assert content_type == 'application/json'
+    assert json.loads(body) == [{'answer': 'the answer', 'q': 'café'}]

{scrapely_python-0.1.12 → scrapely_python-0.1.14}/tests/test_parity.py

@@ -64,13 +64,18 @@ def test_actor_is_standby_reflects_meta_origin() -> None:
         Actor._config = original
 
 
-def test_actor_is_standby_without_config_is_false() -> None:
+def test_actor_is_standby_without_config_is_false(monkeypatch: pytest.MonkeyPatch) -> None:
+    # Before init() is_standby() reads the environment (Apify parity), so clear it.
+    monkeypatch.delenv('APIFY_META_ORIGIN', raising=False)
+    monkeypatch.delenv('ACTOR_META_ORIGIN', raising=False)
     original = Actor._config
     try:
         Actor._config = None
+        scrapely.Configuration.reset_global_state()
         assert Actor.is_standby() is False
     finally:
         Actor._config = original
+        scrapely.Configuration.reset_global_state()
 
 
 def test_parity_helpers_exported_from_package_root() -> None: