scrapely-python 0.1.23__tar.gz → 0.1.24__tar.gz

{scrapely_python-0.1.23/src/scrapely_python.egg-info → scrapely_python-0.1.24}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: scrapely-python
-Version: 0.1.23
+Version: 0.1.24
 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-0.1.23 → scrapely_python-0.1.24}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "scrapely-python"
-version = "0.1.23"
+version = "0.1.24"
 description = "Scrapely Python SDK - A toolkit for building Actors on the Scrapely platform"
 readme = "README.md"
 license = "MIT AND Apache-2.0"

{scrapely_python-0.1.23 → scrapely_python-0.1.24}/src/scrapely/__init__.py

@@ -227,7 +227,7 @@ if TYPE_CHECKING:  # Static analysers and IDEs see the lazy names as plain impor
         set_api_storage,
     )
 
-__version__ = '0.1.23'
+__version__ = '0.1.24'
 __all__ = [
     # Core
     'Actor',

{scrapely_python-0.1.23 → scrapely_python-0.1.24}/src/scrapely/actor.py

@@ -922,6 +922,36 @@ class Actor(metaclass=_ActorMeta):
             cls._default_kvs = await cls.open_key_value_store()
         await cls._default_kvs.set_value(key, value, content_type=content_type)
 
+    # Debug capture: raw source payloads for platform diagnostics. Only active when the platform
+    # sets SCRAPELY_DEBUG_CAPTURE=1 on its own diagnostic runs; a no-op in every other run.
+    _DEBUG_CAPTURE_MAX_BYTES = 256 * 1024
+    _DEBUG_CAPTURE_MAX_RECORDS = 500
+    _debug_capture_count = 0
+
+    @staticmethod
+    def debug_capture_record_key(key: str) -> str:
+        """The key-value store record name for a capture key (item URL or id)."""
+        import hashlib
+        return 'DEBUG_CAPTURE_' + hashlib.sha1(str(key).strip().encode('utf-8')).hexdigest()[:24]
+
+    @classmethod
+    async def debug_capture(cls, key: str, payload: Any) -> None:
+        """Save the raw source payload (API JSON or HTML) an item was extracted from, keyed by
+        the item's URL or id. Does nothing unless the platform enabled capture for this run."""
+        if os.environ.get('SCRAPELY_DEBUG_CAPTURE') != '1' or not key:
+            return
+        if cls._debug_capture_count >= cls._DEBUG_CAPTURE_MAX_RECORDS:
+            return
+        try:
+            text = payload if isinstance(payload, str) else json.dumps(payload, default=str, ensure_ascii=False)
+            if isinstance(text, bytes):
+                text = text.decode('utf-8', 'replace')
+            text = text[: cls._DEBUG_CAPTURE_MAX_BYTES]
+            cls._debug_capture_count += 1
+            await cls.set_value(cls.debug_capture_record_key(key), {'key': str(key), 'payload': text})
+        except Exception as err:  # never break the run
+            logging.getLogger('scrapely').debug('debug_capture failed: %s', err)
+
     @classmethod
     def _resolve_open_args(
         cls,

{scrapely_python-0.1.23 → scrapely_python-0.1.24}/src/scrapely/proxy_configuration.py

@@ -203,7 +203,7 @@ def _platform_missing_error() -> str:
         'proxy credentials (SCRAPELY_PROXY_*/APIFY_PROXY_*) were injected into this '
         'environment. Run the actor on Scrapely with proxy access enabled, or remove the '
         'platform proxy entries. To use the platform proxy locally or from outside '
-        'Scrapely, copy your proxy credentials from the Scrapely console (Proxy page) '
+        'Scrapely, copy your proxy credentials from the Scrapely dashboard (Proxy page) '
         'into the SCRAPELY_PROXY_HOSTNAME, SCRAPELY_PROXY_PORT, SCRAPELY_PROXY_USERNAME '
         'and SCRAPELY_PROXY_PASSWORD environment variables (or pass password=... '
         'explicitly). External platform proxy access requires an active plan or a '
@@ -431,7 +431,7 @@ class ProxyConfiguration:
                 logger.warning(
                     'The Scrapely platform proxy is only available in on-platform runs when your '
                     'account has proxy access. Locally, either add proxy_urls to your actor input, '
-                    'or copy your proxy credentials from the Scrapely console (Proxy page) into the '
+                    'or copy your proxy credentials from the Scrapely dashboard (Proxy page) into the '
                     'SCRAPELY_PROXY_HOSTNAME, SCRAPELY_PROXY_PORT, SCRAPELY_PROXY_USERNAME and '
                     'SCRAPELY_PROXY_PASSWORD environment variables. Using the platform proxy from '
                     'outside Scrapely requires an active plan or a prepaid (PAYG) balance of at '

{scrapely_python-0.1.23 → scrapely_python-0.1.24/src/scrapely_python.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: scrapely-python
-Version: 0.1.23
+Version: 0.1.24
 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-0.1.23 → scrapely_python-0.1.24}/src/scrapely_python.egg-info/SOURCES.txt

@@ -57,6 +57,7 @@ tests/test_charge_limit_guard.py
 tests/test_charge_retries.py
 tests/test_charging_parity.py
 tests/test_charging_pending.py
+tests/test_debug_capture.py
 tests/test_input_proxy_aliases.py
 tests/test_lazy_crawlee.py
 tests/test_parity.py

scrapely_python-0.1.24/tests/test_debug_capture.py

@@ -0,0 +1,31 @@
+import asyncio
+
+from scrapely.actor import Actor
+
+
+def test_debug_capture_noop_without_flag(monkeypatch):
+    monkeypatch.delenv('SCRAPELY_DEBUG_CAPTURE', raising=False)
+    calls = []
+
+    async def fake_set_value(key, value, content_type=None):
+        calls.append(key)
+
+    monkeypatch.setattr(Actor, 'set_value', fake_set_value)
+    asyncio.run(Actor.debug_capture('https://x/item/1', {'a': 1}))
+    assert calls == []
+
+
+def test_debug_capture_writes_keyed_capped_record(monkeypatch):
+    monkeypatch.setenv('SCRAPELY_DEBUG_CAPTURE', '1')
+    saved = {}
+
+    async def fake_set_value(key, value, content_type=None):
+        saved[key] = value
+
+    monkeypatch.setattr(Actor, 'set_value', fake_set_value)
+    monkeypatch.setattr(Actor, '_debug_capture_count', 0)
+    asyncio.run(Actor.debug_capture('https://x/item/1', 'x' * (Actor._DEBUG_CAPTURE_MAX_BYTES + 10)))
+    key = Actor.debug_capture_record_key('https://x/item/1')
+    assert key.startswith('DEBUG_CAPTURE_') and len(key) == len('DEBUG_CAPTURE_') + 24
+    assert saved[key]['key'] == 'https://x/item/1'
+    assert len(saved[key]['payload']) == Actor._DEBUG_CAPTURE_MAX_BYTES