fix: Remove tmp folder for PlaywrightCrawler in non-headless mode

src/crawlee/browsers/_playwright_browser.py

@@ -4,14 +4,13 @@
 import shutil
 import tempfile
 from logging import getLogger
+from pathlib import Path
 from typing import TYPE_CHECKING, Any
 
 from playwright.async_api import Browser
 from typing_extensions import override
 
 if TYPE_CHECKING:
-    from pathlib import Path
-
     from playwright.async_api import BrowserContext, BrowserType, CDPSession, Page
 
 logger = getLogger(__name__)
@@ -36,7 +35,7 @@ def __init__(
         self._browser_type = browser_type
         self._browser_launch_options = browser_launch_options
         self._user_data_dir = user_data_dir
-        self._temp_dir: str | None = None
+        self._temp_dir: Path | None = None
 
         self._context: BrowserContext | None = None
         self._is_connected = True
@@ -63,7 +62,7 @@ async def new_context(self, **context_options: Any) -> BrowserContext:
             user_data_dir = self._user_data_dir
         else:
             user_data_dir = tempfile.mkdtemp(prefix=self._TMP_DIR_PREFIX)
-            self._temp_dir = user_data_dir
+            self._temp_dir = Path(user_data_dir)
 
         self._context = await self._browser_type.launch_persistent_context(
             user_data_dir=user_data_dir, **launch_options
@@ -74,9 +73,9 @@ async def new_context(self, **context_options: Any) -> BrowserContext:
 
         return self._context
 
-    async def _delete_temp_dir(self, _: BrowserContext) -> None:
-        if self._temp_dir:
-            await asyncio.to_thread(shutil.rmtree, self._temp_dir)
+    async def _delete_temp_dir(self, _: BrowserContext | None) -> None:
+        if self._temp_dir and self._temp_dir.exists():
+            await asyncio.to_thread(shutil.rmtree, self._temp_dir, ignore_errors=True)
 
     @override
     async def close(self, **kwargs: Any) -> None:
@@ -85,6 +84,8 @@ async def close(self, **kwargs: Any) -> None:
             await self._context.close()
             self._context = None
         self._is_connected = False
+        await asyncio.sleep(0.1)
+        await self._delete_temp_dir(self._context)
 
     @property
     @override

tests/unit/browsers/test_playwright_browser.py

@@ -0,0 +1,44 @@
+from __future__ import annotations
+
+from pathlib import Path
+from typing import TYPE_CHECKING
+
+import pytest
+from playwright.async_api import async_playwright
+
+from crawlee.browsers._playwright_browser import PlaywrightPersistentBrowser
+
+if TYPE_CHECKING:
+    from collections.abc import AsyncGenerator
+
+    from playwright.async_api import Playwright
+
+
+@pytest.fixture
+async def playwright() -> AsyncGenerator[Playwright, None]:
+    async with async_playwright() as playwright:
+        yield playwright
+
+
+async def test_init(playwright: Playwright) -> None:
+    browser_type = playwright.chromium
+    persist_browser = PlaywrightPersistentBrowser(browser_type, user_data_dir=None, browser_launch_options={})
+    assert persist_browser._browser_type == browser_type
+    assert persist_browser.browser_type == browser_type
+    assert persist_browser._browser_launch_options == {}
+    assert persist_browser._temp_dir is None
+    assert persist_browser._user_data_dir is None
+    assert persist_browser._is_connected is True
+    assert persist_browser.is_connected() is True
+
+
+async def test_delete_temp_folder_with_close_browser(playwright: Playwright) -> None:
+    persist_browser = PlaywrightPersistentBrowser(
+        playwright.chromium, user_data_dir=None, browser_launch_options={'headless': True}
+    )
+    await persist_browser.new_context()
+    assert isinstance(persist_browser._temp_dir, Path)
+    current_temp_dir = persist_browser._temp_dir
+    assert current_temp_dir.exists()
+    await persist_browser.close()
+    assert not current_temp_dir.exists()