I have two scripts:

1. other.py looks like this:

# some stuff is done here and a list of urls is created such as:

urls = ['https://www.walmart.com/ip/Sabrina-Carpenter-Cherry-Pop-EDP-30ml-1oz/5492571361?classType=REGULAR&athbdg=L1600', 'https://www.walmart.com/ip/Hoey-5-1-Painless-Hair-Remover-Women-Facial-Removal-Electric-Cordless-Shaver-Set-Wet-Dry-Lady-Razor-Women-Bikini-Line-Nose-Hair-Eyebrow-Arm-Leg-USB-R/647670434?classType=REGULAR']

# Then, the script runs another script called get_url.py and passes the urls to it to be processed: 
subprocess.Popen(['python', 'get_url.py', str(urls)])

#it is important that this does not block the code and the rest of the code in this script can run without waiting for get_url.py to  complete.

2. get_url.py called above looks like this and downloads each url passed to it:

import pandas as pd
import os
import time
from datetime import datetime
import pyautogui
from selenium.webdriver.chrome.options import Options
from selenium import webdriver
from concurrent.futures import ProcessPoolExecutor

def get_page(url):
    file_name = f"{url[:20]}_{pd.to_datetime(datetime.now()).strftime('%Y-%m-%d %H-%M-%S')}.html"
    file_path = os.path.join(os.getcwd(), 'data', 'htmls')
    path_and_name = os.path.join(file_path, file_name) 

    driver = webdriver.Chrome(options=options)
    driver.get(url)

    time.sleep(1)
    pyautogui.hotkey('ctrl', 's') # open the save as window
    time.sleep(1)
    pyautogui.typewrite(path_and_name ) # enter the path and file name so the webpage is downloaded in the desired directory
    time.sleep(.5)
    pyautogui.hotkey('enter')   
    time.sleep(.2)

    while True: # wait until the download is complete, then close the driver
        files = os.listdir(file_path)
        if file_name in files: 
            driver.close()
            break
        time.sleep(.1)

urls = sys.argv[1] # getting urls from other.py
#converting the string urls to an actual list:
urls = ast.literal_eval(page_urls.replace('[', '').replace(']', '').replace('n', ', '))

if __name__ =='__main__': # multi-processing the urls to speed up things(necessary)
    with ProcessPoolExecutor(max_workers=10) as executer:
        executer.map(get_page, urls, chunksize = 1)

The function works fine as long as I open one browser. However, as soon as multiple windows open by the ProcessPoolExecutor, it appears that the pyautogui.typewrite part of the function loses track of the windows, which may lead to path_and_name being typed multiple times in the “save as” window or typed incomplete leading to the page not downloading or being downloaded with a bad name/directory. Even worse, if I click somewhere like inside my code editor when the function is running, pyautogui may type the path_and_name value in the editor where the cursor is active. Running the browser in “headless” mode so that I don’t accidentally mess with the windows does not help.

So, basically, how do I fix the above code?