crawlee 1.0.1b5__py3-none-any.whl → 1.0.1b6__py3-none-any.whl

crawlee/browsers/_playwright_browser_controller.py

@@ -2,6 +2,7 @@
 
 from __future__ import annotations
 
+from asyncio import Lock
 from datetime import datetime, timedelta, timezone
 from typing import TYPE_CHECKING, Any, cast
 
@@ -77,6 +78,19 @@ class PlaywrightBrowserController(BrowserController):
 
         self._total_opened_pages = 0
 
+        self._context_creation_lock: Lock | None = None
+
+    async def _get_context_creation_lock(self) -> Lock:
+        """Get context checking and creation lock.
+
+        It should be done with lock to prevent multiple concurrent attempts to create context, which could lead to
+        memory leak as one of the two concurrently created contexts will become orphaned and not properly closed.
+        """
+        if self._context_creation_lock:
+            return self._context_creation_lock
+        self._context_creation_lock = Lock()
+        return self._context_creation_lock
+
     @property
     @override
     def pages(self) -> list[Page]:
@@ -137,12 +151,6 @@ class PlaywrightBrowserController(BrowserController):
         Raises:
             ValueError: If the browser has reached the maximum number of open pages.
         """
-        if not self._browser_context:
-            self._browser_context = await self._create_browser_context(
-                browser_new_context_options=browser_new_context_options,
-                proxy_info=proxy_info,
-            )
-
         if not self.has_free_capacity:
             raise ValueError('Cannot open more pages in this browser.')
 
@@ -154,11 +162,12 @@ class PlaywrightBrowserController(BrowserController):
             )
             page = await new_context.new_page()
         else:
-            if not self._browser_context:
-                self._browser_context = await self._create_browser_context(
-                    browser_new_context_options=browser_new_context_options,
-                    proxy_info=proxy_info,
-                )
+            async with await self._get_context_creation_lock():
+                if not self._browser_context:
+                    self._browser_context = await self._create_browser_context(
+                        browser_new_context_options=browser_new_context_options,
+                        proxy_info=proxy_info,
+                    )
             page = await self._browser_context.new_page()
 
         # Handle page close event
@@ -169,7 +178,6 @@ class PlaywrightBrowserController(BrowserController):
         self._last_page_opened_at = datetime.now(timezone.utc)
 
         self._total_opened_pages += 1
-
         return page
 
     @override
@@ -206,7 +214,6 @@ class PlaywrightBrowserController(BrowserController):
         `self._fingerprint_generator` is available.
         """
         browser_new_context_options = dict(browser_new_context_options) if browser_new_context_options else {}
-
         if proxy_info:
             if browser_new_context_options.get('proxy'):
                 logger.warning("browser_new_context_options['proxy'] overriden by explicit `proxy_info` argument.")
@@ -244,5 +251,4 @@ class PlaywrightBrowserController(BrowserController):
         browser_new_context_options['extra_http_headers'] = browser_new_context_options.get(
             'extra_http_headers', extra_http_headers
         )
-
         return await self._browser.new_context(**browser_new_context_options)

{crawlee-1.0.1b5.dist-info → crawlee-1.0.1b6.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: crawlee
-Version: 1.0.1b5
+Version: 1.0.1b6
 Summary: Crawlee for Python
 Project-URL: Apify Homepage, https://apify.com
 Project-URL: Changelog, https://crawlee.dev/python/docs/changelog

{crawlee-1.0.1b5.dist-info → crawlee-1.0.1b6.dist-info}/RECORD

@@ -45,7 +45,7 @@ crawlee/browsers/_browser_controller.py,sha256=-g0pB5Nx5q67eMZVka49x-HMfQqJYoI6k
 crawlee/browsers/_browser_plugin.py,sha256=Wuojop___8ZO9eDoMs4JFmwMAFe5mZaTl0-Vz1PjkD8,3057
 crawlee/browsers/_browser_pool.py,sha256=2pT4m_g0DfopjTHYXb-piN6GqxvkayOeb4gmOtn1QNM,15634
 crawlee/browsers/_playwright_browser.py,sha256=1yXD6cXuLefZZGUG1m0CT38xXYSwIC7n95bJBdMOxbo,3820
-crawlee/browsers/_playwright_browser_controller.py,sha256=IJ6UGD7n99WmyvDCc4RbjURfdAay_R9ymqPbcI2f4Q8,9764
+crawlee/browsers/_playwright_browser_controller.py,sha256=YaY19slRj8gIKrZy0M8rzF_zy2Z1Ym6d0S_vXcMX108,10215
 crawlee/browsers/_playwright_browser_plugin.py,sha256=axZa_yZNCPHyM3Ijx9jW4CzzRXQTVzYAswcGAZHP3Hk,8106
 crawlee/browsers/_types.py,sha256=eWgpoLMWu103hMQQTObkA01sVc_7hdPESl-TCyDMMV0,426
 crawlee/browsers/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -186,8 +186,8 @@ crawlee/storages/_key_value_store.py,sha256=3oI5hVoM_NpTQVKXCbQCmb0sZhW7vN2oXQo-
 crawlee/storages/_request_queue.py,sha256=jt-d-NkI9lAorLssoI2r_lZjeEipe-5Cn6z9bfQqY3k,13154
 crawlee/storages/_storage_instance_manager.py,sha256=iFX3ymsIXyTg8tMHtx5Wn9XyaC77dIf15GpuggsJPDM,7821
 crawlee/storages/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-crawlee-1.0.1b5.dist-info/METADATA,sha256=HbXaga93niEC3vxL0zIv9dvMafyGFSqawpT2lOLoTr4,29314
-crawlee-1.0.1b5.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-crawlee-1.0.1b5.dist-info/entry_points.txt,sha256=1p65X3dA-cYvzjtlxLL6Kn1wpY-3uEDVqJLp53uNPeo,45
-crawlee-1.0.1b5.dist-info/licenses/LICENSE,sha256=AsFjHssKjj4LGd2ZCqXn6FBzMqcWdjQre1byPPSypVw,11355
-crawlee-1.0.1b5.dist-info/RECORD,,
+crawlee-1.0.1b6.dist-info/METADATA,sha256=cjGXdw9ooazFDIzQ_Yj5wJon0rBiEqUtGJ-_AecYKjs,29314
+crawlee-1.0.1b6.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+crawlee-1.0.1b6.dist-info/entry_points.txt,sha256=1p65X3dA-cYvzjtlxLL6Kn1wpY-3uEDVqJLp53uNPeo,45
+crawlee-1.0.1b6.dist-info/licenses/LICENSE,sha256=AsFjHssKjj4LGd2ZCqXn6FBzMqcWdjQre1byPPSypVw,11355
+crawlee-1.0.1b6.dist-info/RECORD,,