PlaywrightCapture 1.31.6__tar.gz → 1.32.0__tar.gz

{playwrightcapture-1.31.6 → playwrightcapture-1.32.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: PlaywrightCapture
-Version: 1.31.6
+Version: 1.32.0
 Summary: A simple library to capture websites using playwright
 License: BSD-3-Clause
 Author: Raphaël Vinot
@@ -20,9 +20,9 @@ Classifier: Topic :: Security
 Provides-Extra: recaptcha
 Requires-Dist: SpeechRecognition (>=3.14.3) ; extra == "recaptcha"
 Requires-Dist: aiohttp-socks (>=0.10.1)
-Requires-Dist: aiohttp[speedups] (>=3.12.14)
+Requires-Dist: aiohttp[speedups] (>=3.12.15)
 Requires-Dist: async-timeout (>=5.0.1) ; python_version < "3.11"
-Requires-Dist: beautifulsoup4[charset-normalizer,lxml] (>=4.13.4)
+Requires-Dist: beautifulsoup4[charset-normalizer,lxml] (>=4.13.5)
 Requires-Dist: dateparser (>=1.2.2)
 Requires-Dist: dnspython (>=2.7.0,<3.0.0)
 Requires-Dist: playwright (>=1.54.0)

{playwrightcapture-1.31.6 → playwrightcapture-1.32.0}/playwrightcapture/__init__.py

@@ -1,4 +1,6 @@
 from .capture import Capture  # noqa
+from .capture import CaptureResponse  # noqa
+from .capture import SetCookieParam, Cookie  # noqa
 from .helpers import get_devices  # noqa
 from .exceptions import (PlaywrightCaptureException, UnknownPlaywrightDeviceType,  # noqa
                          UnknownPlaywrightBrowser, UnknownPlaywrightDevice,
@@ -6,6 +8,8 @@ from .exceptions import (PlaywrightCaptureException, UnknownPlaywrightDeviceType
 
 __all__ = [
     'Capture',
+    'CaptureResponse',
+    'SetCookieParam', 'Cookie',
     'get_devices',
     'PlaywrightCaptureException',
     'UnknownPlaywrightDeviceType',

{playwrightcapture-1.31.6 → playwrightcapture-1.32.0}/playwrightcapture/capture.py

@@ -46,18 +46,6 @@ if sys.version_info < (3, 11):
 else:
     from asyncio import timeout
 
-if sys.version_info < (3, 12):
-    from typing_extensions import TypedDict
-else:
-    from typing import TypedDict
-
-if TYPE_CHECKING:
-    from playwright._impl._api_structures import (SetCookieParam, Geolocation,
-                                                  HttpCredentials, Headers,
-                                                  ViewportSize, Cookie,
-                                                  ProxySettings, StorageState)
-    BROWSER = Literal['chromium', 'firefox', 'webkit']
-
 try:
     if sys.version_info < (3, 10):
         from pydub import AudioSegment  # type: ignore[attr-defined]
@@ -68,6 +56,54 @@ try:
 except ImportError:
     CAN_SOLVE_CAPTCHA = False
 
+# ####
+# This bit is required as long as we need to support python < 3.12, stop removing it, you idiot
+# That's the runtime failure: https://github.com/ail-project/lacus/actions/runs/16447900822/job/46484807036
+# And the MyPy failure: https://github.com/ail-project/LacusCore/actions/runs/16447753492/job/46484287947
+if sys.version_info < (3, 12):
+    from typing_extensions import TypedDict
+else:
+    from typing import TypedDict
+
+# These two classes are copies of te ones in Playwright, we need them until we can discard python 3.11
+# They come from this file:
+#   https://github.com/microsoft/playwright-python/blob/main/playwright/_impl/_api_structures.py
+
+
+class SetCookieParam(TypedDict, total=False):
+    name: str
+    value: str
+    url: str
+    domain: str
+    path: str
+    expires: float
+    httpOnly: bool
+    secure: bool
+    sameSite: Literal["Lax", "None", "Strict"]
+    partitionKey: str
+
+
+class Cookie(TypedDict, total=False):
+    name: str
+    value: str
+    domain: str
+    path: str
+    expires: float
+    httpOnly: bool
+    secure: bool
+    sameSite: Literal["Lax", "None", "Strict"]
+    partitionKey: str
+
+# ####################################
+
+
+if TYPE_CHECKING:
+    from playwright._impl._api_structures import (Geolocation,
+                                                  HttpCredentials, Headers,
+                                                  ViewportSize,
+                                                  ProxySettings, StorageState)
+    BROWSER = Literal['chromium', 'firefox', 'webkit']
+
 
 class CaptureResponse(TypedDict, total=False):
 
@@ -198,8 +234,6 @@ class Capture():
 
     async def __aenter__(self) -> Capture:
         '''Launch the browser'''
-        self._temp_harfile = NamedTemporaryFile(delete=False)
-
         self.playwright = await async_playwright().start()
 
         if self.device_name:
@@ -219,11 +253,12 @@ class Capture():
         # Set of URLs that were captured in that context
         self._already_captured: set[str] = set()
 
+        # Create the temporary file to store the HAR content.
+        self._temp_harfile = NamedTemporaryFile(delete=False, prefix="playwright_capture_har", suffix=".json")
+
         return self
 
     async def __aexit__(self, exc_type: Any, exc: Any, tb: Any) -> bool:
-        if hasattr(self, '_temp_harfile'):
-            os.unlink(self._temp_harfile.name)
 
         try:
             await self.browser.close(reason="Closing browser at the end of the capture.")
@@ -235,6 +270,13 @@ class Capture():
         except Exception as e:
             # this should't happen, but just in case it does...
             self.logger.info(f'Unable to stop playwright: {e}')
+
+        if hasattr(self, '_temp_harfile'):
+            try:
+                os.unlink(self._temp_harfile.name)
+            except Exception as e:
+                self.logger.warning(f'Unable to remove temp HAR file {self._temp_harfile.name}: {e}')
+
         return True
 
     @property
@@ -515,7 +557,9 @@ class Capture():
 
         if self.cookies:
             try:
-                await self.context.add_cookies(self.cookies)
+                # NOTE: Ignore type until we can use python 3.12 + only
+                # playwrightcapture.capture.SetCookieParam == playwright._impl._api_structures.SetCookieParam
+                await self.context.add_cookies(self.cookies)  # type: ignore[arg-type]
             except Exception:
                 self.logger.exception(f'Unable to set cookies: {self.cookies}')
 
@@ -1237,7 +1281,9 @@ class Capture():
 
                 try:
                     async with timeout(15):
-                        to_return['cookies'] = await self.context.cookies()
+                        # NOTE: Ignore type until we can use python 3.12 + only
+                        # playwrightcapture.capture.SetCookieParam == playwright._impl._api_structures.SetCookieParam
+                        to_return['cookies'] = await self.context.cookies()  # type: ignore[typeddict-item]
                 except (TimeoutError, asyncio.TimeoutError):
                     self.logger.warning("Unable to get cookies (timeout).")
                     errors.append("Unable to get the cookies (timeout).")

{playwrightcapture-1.31.6 → playwrightcapture-1.32.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "PlaywrightCapture"
-version = "1.31.6"
+version = "1.32.0"
 description = "A simple library to capture websites using playwright"
 authors = [
     {name="Raphaël Vinot", email= "raphael.vinot@circl.lu"}
@@ -14,14 +14,14 @@ dynamic = [ "classifiers" ]
 dependencies = [
     "playwright (>=1.54.0)",
     "dateparser (>=1.2.2)",
-    "beautifulsoup4[charset-normalizer,lxml] (>=4.13.4)",
+    "beautifulsoup4[charset-normalizer,lxml] (>=4.13.5)",
     "w3lib (>=2.3.1)",
     "tzdata (>=2025.2)",
     "playwright-stealth (>=2)",
     "setuptools (>=80.9.0)",
     "puremagic (>=1.30)",
     "async-timeout (>=5.0.1) ; python_version < \"3.11\"",
-    "aiohttp[speedups] (>=3.12.14)",
+    "aiohttp[speedups] (>=3.12.15)",
     "aiohttp-socks (>=0.10.1)",
     "typing-extensions (>=4.12.2,<5.0.0) ; python_version < \"3.12\"",
     "dnspython (>=2.7.0,<3.0.0)",
@@ -51,9 +51,9 @@ recaptcha = [
 [tool.poetry.group.dev.dependencies]
 types-beautifulsoup4 = "^4.12.0.20250516"
 pytest = "^8.4.1"
-mypy = "^1.17.0"
-types-dateparser = "^1.2.2.20250627"
-types-pytz = "^2025.2.0.20250516"
+mypy = "^1.17.1"
+types-dateparser = "^1.2.2.20250809"
+types-pytz = "^2025.2.0.20250809"
 
 
 [build-system]