fix: actually use the updated user agent

docs: some clarity about the state of the project

README.md

@@ -2,13 +2,13 @@
 
 Transplanted from [puppeteer-extra-plugin-stealth](https://github.com/berstend/puppeteer-extra/tree/master/packages/puppeteer-extra-plugin-stealth), with some improvements. Don't expect this to bypass anything but the simplest of bot detection methods. Consider this a proof-of-concept starting point.
 
-This is a WIP fork, with the goal of replacing the out-of-date upstream. See the [changelog](./README.md).
+This is a WIP fork. I've merged some of the outstanding PRs and added some features with the goal of replacing the out-of-date upstream. See the [changelog](./README.md).
 
 ## Install
 
 Install the latest stable release:
 ```
-$ pip install git+https://github.com/Mattwmaster58/playwright_stealth@rc2
+$ pip install git+https://github.com/Mattwmaster58/playwright_stealth@rc3
 ```
 or straight from `main`
 ```
@@ -27,6 +27,7 @@ from playwright_stealth import Stealth, ALL_EVASIONS_DISABLED_KWARGS
 async def main():
     # This is the recommended usage. All pages created will have stealth applied:
     async with Stealth().use_async(async_playwright()) as p:
+        # or, to hook every browser launched from this context: stealth.hook_playwright_context(p)
         browser = await p.chromium.launch()
         page = await browser.new_page()
         print("from new_page: ", await page.evaluate("navigator.webdriver"))
@@ -41,8 +42,8 @@ async def main():
         init_scripts_only=True
     )
     async with async_playwright() as p:
-        # or, to hook every browser launched from this context: stealth.hook_playwright_context(p)
         browser = await p.chromium.launch()
+        # this isn't, certain evasions will work worse, and some won't be able to be applied at all
         context = await browser.new_context()
         await stealth.apply_stealth_async(context)
         page_1 = await context.new_page()

playwright_stealth/stealth.py

@@ -301,10 +301,12 @@ def _generate_hooked_new_page(self, new_page_method: Callable) -> Callable:
         *args and **kwargs even though these methods may not take any number of arguments,
         we want to preserve accurate stack traces when caller passes args improperly
         """
-        browser_instance = new_page_method.__self__
         USER_AGENT_OVERRIDE_PIGGYBACK_KEY = "_stealth_user_agent"
         SEC_CH_UA_OVERRIDE_PIGGYBACK_KEY = "_stealth_sec_ch_ua"
 
+        browser_instance = new_page_method.__self__
+        is_chromium = browser_instance.browser_type.name == "chromium"
+
         async def get_user_agent_and_sec_ch_ua_async(page_method: Callable) -> Tuple[str, str]:
             """
             If there's no override, it's Chrome, and we haven't cached a UA value prior, we need to come up
@@ -316,8 +318,8 @@ async def get_user_agent_and_sec_ch_ua_async(page_method: Callable) -> Tuple[str
                 user_agent, sec_ch_ua
             """
             temp_page: Optional[async_api.Page]
-            stealth_user_agent = getattr(browser_instance, USER_AGENT_OVERRIDE_PIGGYBACK_KEY)
-            sec_ch_ua = getattr(browser_instance, SEC_CH_UA_OVERRIDE_PIGGYBACK_KEY)
+            stealth_user_agent = getattr(browser_instance, USER_AGENT_OVERRIDE_PIGGYBACK_KEY, None)
+            sec_ch_ua = getattr(browser_instance, SEC_CH_UA_OVERRIDE_PIGGYBACK_KEY, None)
             if stealth_user_agent is None or sec_ch_ua is None:
                 temp_page = await page_method()
                 stealth_user_agent = (await temp_page.evaluate("navigator.userAgent")).replace(
@@ -345,43 +347,43 @@ def get_user_agent_and_sec_ch_ua_sync(page_method: Callable) -> Tuple[str, str]:
         async def hooked_browser_method_async(*args, **kwargs):
             # respect any override the user passes themselves
             if self.navigator_user_agent and kwargs.get("user_agent") is None:
-                user_agent_override = self.navigator_user_agent_override
-                if user_agent_override is None and browser_instance.browser_type == "chromium":
-                    user_agent_override, _ = await get_user_agent_and_sec_ch_ua_async(new_page_method)
-                kwargs["user_agent"] = self.navigator_user_agent_override
+                resolved_user_agent_override = self.navigator_user_agent_override
+                if resolved_user_agent_override is None and is_chromium:
+                    resolved_user_agent_override, _ = await get_user_agent_and_sec_ch_ua_async(new_page_method)
+                kwargs["user_agent"] = resolved_user_agent_override
 
             extra_http_headers = kwargs.get("extra_http_headers", {})
             # respect any override the user passes themselves
             if self.sec_ch_ua and CaseInsensitiveDict(extra_http_headers).get("sec-ch-ua") is None:
-                sec_ch_ua_override = self.sec_ch_ua_override
-                if sec_ch_ua_override is None and browser_instance.browser_type == "chromium":
-                    _, sec_ch_ua_override = await get_user_agent_and_sec_ch_ua_async(new_page_method)
-                if sec_ch_ua_override is not None:
+                resolved_sec_ch_ua_override = self.sec_ch_ua_override
+                if resolved_sec_ch_ua_override is None and is_chromium:
+                    _, resolved_sec_ch_ua_override = await get_user_agent_and_sec_ch_ua_async(new_page_method)
+                if resolved_sec_ch_ua_override is not None:
                     # this could be tricky is a differently cased key of the same thing exists,
                     # but we have done a case-insensitive check above that precludes this
-                    extra_http_headers["sec-ch-ua"] = sec_ch_ua_override
+                    extra_http_headers["sec-ch-ua"] = resolved_sec_ch_ua_override
                     kwargs["extra_http_headers"] = extra_http_headers
             page = await new_page_method(*args, **kwargs)
             await self.apply_stealth_async(page)
             return page
 
         def hooked_browser_method_sync(*args, **kwargs):
             if self.navigator_user_agent and kwargs.get("user_agent") is None:
-                user_agent_override = self.navigator_user_agent_override
-                if user_agent_override is None and browser_instance.browser_type == "chromium":
-                    user_agent_override, _ = get_user_agent_and_sec_ch_ua_sync(new_page_method)
+                resolved_user_agent_override = self.navigator_user_agent_override
+                if resolved_user_agent_override is None and is_chromium:
+                    resolved_user_agent_override, _ = get_user_agent_and_sec_ch_ua_sync(new_page_method)
                 kwargs["user_agent"] = self.navigator_user_agent_override
 
             extra_http_headers = kwargs.get("extra_http_headers", {})
             if self.sec_ch_ua and CaseInsensitiveDict(extra_http_headers).get("sec-ch-ua") is None:
-                sec_ch_ua_override = self.sec_ch_ua_override
+                resolved_sec_ch_ua_override = self.sec_ch_ua_override
                 # respect any override the user has already made
-                if sec_ch_ua_override is None and browser_instance.browser_type == "chromium":
-                    _, sec_ch_ua_override = get_user_agent_and_sec_ch_ua_sync(new_page_method)
-                if sec_ch_ua_override is not None:
+                if resolved_sec_ch_ua_override is None and is_chromium:
+                    _, resolved_sec_ch_ua_override = get_user_agent_and_sec_ch_ua_sync(new_page_method)
+                if resolved_sec_ch_ua_override is not None:
                     # this could be tricky is a differently cased key of the same thing exists,
                     # but we have done a case-insensitive check above that precludes this
-                    extra_http_headers["sec-ch-ua"] = sec_ch_ua_override
+                    extra_http_headers["sec-ch-ua"] = resolved_sec_ch_ua_override
                     kwargs["extra_http_headers"] = extra_http_headers
             page = new_page_method(*args, **kwargs)
             self.apply_stealth_sync(page)
@@ -428,7 +430,7 @@ def _get_greased_chrome_sec_ua_ch(user_agent: str):
         greased_versions = [8, 99, 24]
         greasy_chars = " ():-./;=?_"
         greasy_brand = f"Not{random.choice(greasy_chars)}A{random.choice(greasy_chars)}Brand"
-        version = re.search(r"Chrome/([\d.]+)", user_agent, re.IGNORECASE)
+        version = re.search(r"Chrome/(\d+)[\d.]+", user_agent, re.IGNORECASE)
         major_version = version.group(1)
         brands = [
             ("Chromium", major_version),

setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = playwright-stealth
-version = 2.0.0rc1
+version = 2.0.0rc3
 author = AtuboDad
 author_email = lcjasas@sina.com
 maintainer = Mattwmaster58