scrapely-python 0.1.33__tar.gz → 0.1.34__tar.gz

{scrapely_python-0.1.33 → scrapely_python-0.1.34}/CHANGELOG.md

@@ -2,6 +2,10 @@
 
 All notable changes to this project will be documented in this file.
 
+## [0.1.34] - 2026-10-23
+
+- Debug captures (`Actor.debug_capture`, `Actor.debug_capture_sync`) store parsed XML/HTML elements as their markup instead of a repr such as `<Element 'change' at 0x...>`: `xml.etree.ElementTree` elements and trees, lxml elements and trees, BeautifulSoup tags, selectolax nodes and parsel / Scrapy selectors, passed alone or inside a dict or list. The text after an element (its tail) is not included. Other objects are still stored as their `str()`. Captures are only active on the platform's own diagnostic runs (`SCRAPELY_DEBUG_CAPTURE=1`).
+
 ## [0.1.33] - 2026-10-22
 
 - Debug captures (`Actor.debug_capture`, `Actor.debug_capture_sync`) keep payloads over 256 KiB whole. They are stored gzip-compressed and base64-encoded as `{key, payload, encoding: 'gzip-base64', length, truncated}` instead of being cut at 262,144 characters; up to 8 Mi characters are kept. A payload that does not compress enough keeps its first 256 KiB, now marked `truncated: True` with its full `length`. Dict and list payloads are dumped as compact JSON, and `bytes` payloads are stored as UTF-8 text (they used to be stored as their `repr`). Captures are only active on the platform's own diagnostic runs (`SCRAPELY_DEBUG_CAPTURE=1`).

{scrapely_python-0.1.33 → scrapely_python-0.1.34}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: scrapely-python
-Version: 0.1.33
+Version: 0.1.34
 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-0.1.33 → scrapely_python-0.1.34}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "scrapely-python"
-version = "0.1.33"
+version = "0.1.34"
 description = "Scrapely Python SDK - A toolkit for building Actors on the Scrapely platform"
 readme = "README.md"
 license = "MIT AND Apache-2.0"

{scrapely_python-0.1.33 → scrapely_python-0.1.34}/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__ = '0.1.33'
+__version__ = '0.1.34'
 __all__ = [
     # Core
     'Actor',

{scrapely_python-0.1.33 → scrapely_python-0.1.34}/src/scrapely/actor.py

@@ -285,6 +285,53 @@ class _ActorMeta(type):
         await cls._shutdown()
 
 
+
+def _element_markup(obj: Any) -> Optional[str]:
+    """The markup of a parsed XML/HTML element, or None when ``obj`` is not one.
+
+    Debug captures store the source an item was read from. Feed and page parsers hand over elements
+    (xml.etree / lxml elements and trees, BeautifulSoup tags, selectolax nodes, parsel / Scrapy
+    selectors); their ``str()`` is only a repr such as ``<Element 'change' at 0x...>``, which shows
+    nothing of the source. Optional parsers are recognised by module, never imported here.
+    """
+    import xml.etree.ElementTree as ET
+
+    if isinstance(obj, ET.ElementTree):
+        root = obj.getroot()
+        return ET.tostring(root, encoding='unicode') if root is not None else ''
+    if isinstance(obj, ET.Element):
+        if obj.tail:  # the text after the element belongs to its parent, as with lxml's with_tail=False
+            import copy
+
+            obj = copy.copy(obj)
+            obj.tail = None
+        return ET.tostring(obj, encoding='unicode')
+    modules = [getattr(c, '__module__', '') or '' for c in type(obj).__mro__]
+    if any(m == 'lxml' or m.startswith('lxml.') for m in modules):
+        from lxml import etree  # type: ignore[import-not-found]
+
+        out = etree.tostring(obj, encoding='unicode', with_tail=False)
+        return out if isinstance(out, str) else None
+    if any(m == 'bs4' or m.startswith('bs4.') for m in modules):
+        return str(obj)
+    if any(m == 'selectolax' or m.startswith('selectolax.') for m in modules):
+        html = getattr(obj, 'html', None)
+        return html if isinstance(html, str) else None
+    if any(m == 'parsel' or m.startswith('parsel.') for m in modules) and callable(getattr(obj, 'get', None)):
+        out = obj.get()
+        return out if isinstance(out, str) else None
+    return None
+
+
+def _capture_json_default(obj: Any) -> Any:
+    """json.dumps fallback for debug captures: elements as their markup, anything else as str()."""
+    try:
+        markup = _element_markup(obj)
+    except Exception:  # an unusual node type must not lose the rest of the payload
+        markup = None
+    return markup if markup is not None else str(obj)
+
+
 class Actor(metaclass=_ActorMeta):
     """Core class for building Actors on the Scrapely platform."""
 
@@ -974,12 +1021,15 @@ class Actor(metaclass=_ActorMeta):
         if cls._debug_capture_count >= cls._DEBUG_CAPTURE_MAX_RECORDS:
             return
         try:
-            if isinstance(payload, (bytes, bytearray)):
+            markup = _element_markup(payload)
+            if markup is not None:
+                text = markup
+            elif isinstance(payload, (bytes, bytearray)):
                 text = bytes(payload).decode('utf-8', 'replace')
             elif isinstance(payload, str):
                 text = payload
             else:
-                text = json.dumps(payload, default=str, ensure_ascii=False, separators=(',', ':'))
+                text = json.dumps(payload, default=_capture_json_default, ensure_ascii=False, separators=(',', ':'))
             record = cls._debug_capture_record(str(key), text)
             cls._debug_capture_count += 1
             await cls.set_value(cls.debug_capture_record_key(key), record)

{scrapely_python-0.1.33 → scrapely_python-0.1.34}/src/scrapely_python.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: scrapely-python
-Version: 0.1.33
+Version: 0.1.34
 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