PlaywrightCapture 1.27.0__tar.gz → 1.27.2__tar.gz

{playwrightcapture-1.27.0 → playwrightcapture-1.27.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: PlaywrightCapture
-Version: 1.27.0
+Version: 1.27.2
 Summary: A simple library to capture websites using playwright
 Home-page: https://github.com/Lookyloo/PlaywrightCapture
 License: BSD-3-Clause
@@ -22,15 +22,15 @@ Classifier: Topic :: Security
 Provides-Extra: recaptcha
 Requires-Dist: SpeechRecognition (>=3.11.0) ; extra == "recaptcha"
 Requires-Dist: aiohttp-socks (>=0.9,<0.10)
-Requires-Dist: aiohttp[speedups] (>=3.10.10,<4.0.0)
+Requires-Dist: aiohttp[speedups] (>=3.11.7,<4.0.0)
 Requires-Dist: async-timeout (>=4.0.3,<5.0.0) ; python_version < "3.11"
 Requires-Dist: beautifulsoup4[charset-normalizer,lxml] (>=4.12.3,<5.0.0)
 Requires-Dist: dateparser (>=1.2.0,<2.0.0)
-Requires-Dist: playwright (>=1.48.0,<2.0.0)
+Requires-Dist: playwright (>=1.49.0,<2.0.0)
 Requires-Dist: playwright-stealth (>=1.0.6,<2.0.0)
 Requires-Dist: puremagic (>=1.28,<2.0)
 Requires-Dist: pydub (>=0.25.1,<0.26.0) ; extra == "recaptcha"
-Requires-Dist: setuptools (>=75.3.0,<76.0.0)
+Requires-Dist: setuptools (>=75.6.0,<76.0.0)
 Requires-Dist: tzdata (>=2024.2,<2025.0)
 Requires-Dist: w3lib (>=2.2.1,<3.0.0)
 Project-URL: Repository, https://github.com/Lookyloo/PlaywrightCapture

{playwrightcapture-1.27.0 → playwrightcapture-1.27.2}/playwrightcapture/capture.py

@@ -732,7 +732,7 @@ class Capture():
                             await frame.get_by_label(label).click(timeout=2000)
                             break
                 except (TimeoutError, asyncio.TimeoutError) as e:
-                    self.logger.warning(f'Frame consent timeout: {e}')
+                    self.logger.warning(f'Consent timeout (label {label}) : {e}')
 
                 try:
                     async with timeout(5):
@@ -742,9 +742,9 @@ class Capture():
                             await frame.get_by_role("button", name=label).click(timeout=2000)
                             break
                 except (TimeoutError, asyncio.TimeoutError) as e:
-                    self.logger.warning(f'Frame consent timeout: {e}')
+                    self.logger.warning(f'Frame consent timeout (button {label}): {e}')
         except Exception as e:
-            self.logger.info(f'Issue with frame consent: {e}')
+            self.logger.info(f'Issue with consent validation: {e}')
         return got_button
 
     async def _move_time_forward(self, page: Page, time: int) -> None:
@@ -825,12 +825,18 @@ class Capture():
             capturing_sub = False
             try:
                 page = await self.context.new_page()
-                await page.clock.install()
             except Error as e:
-                self.logger.warning(f'The context is in a broken state: {e}')
+                self.logger.warning(f'Unable to create new page, the context is in a broken state: {e}')
                 self.should_retry = True
                 return to_return
 
+            try:
+                await page.clock.install()
+                clock_set = True
+            except Error as e:
+                self.logger.warning(f'Unable to install the clock: {e}')
+                clock_set = False
+
             if allow_tracking:
                 # Add authorization clickthroughs
                 await self.__dialog_didomi_clickthrough(page)
@@ -898,8 +904,7 @@ class Capture():
                     await page.bring_to_front()
                     self.logger.debug('Page moved to front.')
                 except Error as e:
-                    self.logger.warning('Page in a broken state.')
-                    raise e
+                    self.logger.warning(f'Unable to bring the page to the front: {e}.')
 
                 # page instrumentation
                 await self._wait_for_random_timeout(page, 5)  # Wait 5 sec after document loaded
@@ -969,7 +974,8 @@ class Capture():
                             self.logger.debug('Got button on main frame')
                             await self._wait_for_random_timeout(page, 10)  # Wait 10 sec after click
 
-                    await self._move_time_forward(page, 10)
+                    if clock_set:
+                        await self._move_time_forward(page, 10)
 
                     if parsed_url.fragment:
                         # We got a fragment, make sure we go to it and scroll only a little bit.
@@ -1037,8 +1043,9 @@ class Capture():
                                 z.writestr(f'{i}_{filename}', file_content)
                         to_return["downloaded_file"] = mem_zip.getvalue()
 
-                # fast forward ~30s
-                await self._move_time_forward(page, 30)
+                if clock_set:
+                    # fast forward ~30s
+                    await self._move_time_forward(page, 30)
 
                 self.logger.debug('Done with instrumentation, waiting for network idle.')
                 await self._wait_for_random_timeout(page, 5)  # Wait 5 sec after instrumentation
@@ -1460,6 +1467,7 @@ class Capture():
                 'net::ERR_CONNECTION_TIMED_OUT',
                 'net::ERR_HTTP_RESPONSE_CODE_FAILURE',
                 'net::ERR_HTTP2_PROTOCOL_ERROR',
+                'net::ERR_INVALID_HTTP_RESPONSE',
                 'net::ERR_INVALID_REDIRECT',
                 'net::ERR_NAME_NOT_RESOLVED',
                 'net::ERR_NETWORK_ACCESS_DENIED',

{playwrightcapture-1.27.0 → playwrightcapture-1.27.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "PlaywrightCapture"
-version = "1.27.0"
+version = "1.27.2"
 description = "A simple library to capture websites using playwright"
 authors = ["Raphaël Vinot <raphael.vinot@circl.lu>"]
 license = "BSD-3-Clause"
@@ -19,7 +19,7 @@ classifiers=[
 
 [tool.poetry.dependencies]
 python = "^3.9"
-playwright = "^1.48.0"
+playwright = "^1.49.0"
 dateparser = "^1.2.0"
 beautifulsoup4 = {version= "^4.12.3", extras = ["lxml", "charset_normalizer"]}
 w3lib = "^2.2.1"
@@ -27,10 +27,10 @@ pydub = {version = "^0.25.1", optional = true}
 SpeechRecognition =  {version = ">=3.11.0", optional = true}
 tzdata = "^2024.2"
 playwright-stealth = "^1.0.6"
-setuptools = "^75.3.0"
+setuptools = "^75.6.0"
 puremagic = "^1.28"
 async-timeout = {version = "^4.0.3", python = "<3.11"}
-aiohttp = {version = "^3.10.10", extras = ["speedups"]}
+aiohttp = {version = "^3.11.7", extras = ["speedups"]}
 aiohttp-socks = "^0.9"
 
 [tool.poetry.extras]