uc_gui_click_cf does not click exactly CF checkbox position

[ngocdangrby]

Hi, I am trying to run simple script to get traffic information.

And I got some problems:

1. uc_gui_click_cf click wrong position for checkbox
2. I can not do something else on my computer when this script is because the mouse is moving to click checkbox. Is any option to run headless or "slient mode". I am using Ubuntu.

lib version:

selenium             4.33.0
seleniumbase         4.39.0

here is my code:

from seleniumbase import SB

with SB(uc=True, xvfb=True,test=True, incognito=True, locale="en") as sb:
    domain = "ahrefs.com"
    url = f"https://ahrefs.com/traffic-checker/?input={domain}&mode=subdomains"
    sb.uc_open_with_reconnect(url, reconnect_time=4)
    sb.uc_gui_click_cf()
    print(sb.get_page_title())
    sb.wait_for_text_not_visible("Checking", timeout=12)
    sb.highlight(f"p:contains('{domain}')", timeout=12)
    sb.set_messenger_theme(location="bottom_center")
    sb.post_message("SeleniumBase wasn't detected!")

Video:

screen-capture.24.webm

[mdmintz]

Make sure that the scale-factor is set to 100%, and that fractional scaling is disabled. (You may also need to maximize the window.)

---

For your 2nd issue, PyAutoGUI can only control one mouse at a time, and it can only click on headed browsers.
(Xvfb gets you around that limitation on headless Linux systems.)

---

There's also sb.cdp.gui_click_element(selector), which should be used on the element directly above the shadow-root of the Turnstile. Example here: SeleniumBase/examples/cdp_mode/raw_planetmc.py.

[ngocdangrby]

thank for quick reply.

I updated the code and passed CF in my laptop.

but when I deploy in docker,

sb.cdp.maximize() does not maximize the window, even it reduces the window size.

import time
import os
from seleniumbase import SB

import pyautogui

print(pyautogui.size())

with SB(uc=True, incognito=True,xvfb=True, xvfb_metrics="1920,1080", locale="en") as sb:
    domain = "ahrefs.com"
    url = f"https://ahrefs.com/traffic-checker/?input={domain}&mode=subdomains"
    sb.activate_cdp_mode(url)
    print("SeleniumBase is ready!")
    print(sb.cdp.get_window_size())
    print("Maximizing window...")
    sb.cdp.maximize()
    time.sleep(1)
    print(sb.cdp.get_window_size())
    sb.cdp.save_screenshot(f"images/screenshot_{domain}_{time.time()}.png")
    sb.cdp.gui_click_x_y(900,295, timeframe=1)
    sb.cdp.sleep(1)
    print(sb.cdp.get_window_size())
    print("Clicking on 'Check' button...")

here is my log got

Size(width=1920, height=1080) <-------- correct size of screen xvfb
SeleniumBase is ready!
{'width': 1280, 'height': 840}
Maximizing window...
{'width': 1280, 'height': 720}  <-------- not correct full size window
 <DEBUG> pyautogui.click(900, 295)
{'width': 1280, 'height': 720}
Clicking on 'Check' button...

Thank you.

[mdmintz]

The bigger issue with Docker is that it leaves a detectable fingerprint, so it's not advisable to use UC Mode / CDP Mode with Docker. Regular Ubuntu without Docker is best for Linux.

[ngocdangrby]

thanks.
i will try to run it in a VM or VPS.

[dean985]

@mdmintz
Hey, thanks for your work.
Can you please elaborate on what fingerprint does docker leave? Compared to a regular Ubuntu machine, for instance.