add support for custom URL

[ajadach]

Hi, sometime we do not have access to the google pages from inside corporate network, so we keep zip's on our custom host. I would like to provide my own link to download the dedicated .zip file.

we just need add new argument custom_url_to_zip, add change variable for folder_name and chromedriver_version

`def download_chromedriver(path: Optional[AnyStr] = None, no_ssl: bool = False, custom_url_to_zip: str = ""):

if not custom_url_to_zip:
    chrome_version = get_chrome_version()
    if not chrome_version:
        logging.debug("Chrome is not installed.")
        return
    chromedriver_version, download_options = get_matched_chromedriver_version(chrome_version, no_ssl)

    major_version = get_major_version(chromedriver_version)
    folder_name = major_version

    if not chromedriver_version or (major_version >= "115" and not download_options):
        logging.warning(
            "Can not find chromedriver for currently installed chrome version."
        )
        return
else:
    folder_name = 'chromedriver'
    chromedriver_version = 'custom'

if path:
    if not os.path.isdir(path):
        raise ValueError(f"Invalid path: {path}")
    chromedriver_dir = os.path.join(os.path.abspath(path), folder_name)
else:
    chromedriver_dir = os.path.join(
        os.path.abspath(os.path.dirname(__file__)), folder_name
    )
chromedriver_filename = get_chromedriver_filename()
chromedriver_filepath = os.path.join(chromedriver_dir, chromedriver_filename)
if not os.path.isfile(chromedriver_filepath) or not check_version(
    chromedriver_filepath, chromedriver_version
):
    logging.info(f"Downloading chromedriver ({chromedriver_version})...")
    if not os.path.isdir(chromedriver_dir):
        os.makedirs(chromedriver_dir)

    if not custom_url_to_zip:
        url = get_chromedriver_url(chromedriver_version=chromedriver_version, download_options=download_options, no_ssl=no_ssl)
    else:
        url = custom_url_to_zip
    try:
        response = urllib.request.urlopen(url)
        if response.getcode() != 200:
            raise urllib.error.URLError("Not Found")
    except urllib.error.URLError:
        raise RuntimeError(f"Failed to download chromedriver archive: {url}")
    archive = BytesIO(response.read())
    with zipfile.ZipFile(archive) as zip_file:
        # v115+ have files in subdirectories- need to adjust filename before extracting
        for zip_info in zip_file.infolist():
            if os.path.basename(zip_info.filename) == chromedriver_filename:
                zip_info.filename = chromedriver_filename
                zip_file.extract(zip_info, chromedriver_dir)
                break
else:
    logging.info("Chromedriver is already installed.")
if not os.access(chromedriver_filepath, os.X_OK):
    os.chmod(chromedriver_filepath, 0o744)
return chromedriver_filepath`

[shawnCaza]

Thanks for the suggestion @ajadach.

As much of what this package does is identify the need for a driver, and match the need to the appropriate system and browser version, I'm not sure it makes sense to incorporate this change for a cudstom use case that appears to bypass most of that.

I do think the download_chromedriver function is doing too much work.

if the logic below was moved into a function of its own, that you could call directly with a 'url' and 'folder' variable would that cover most of what you need to do?

try:
        response = urllib.request.urlopen(url)
        if response.getcode() != 200:
            raise urllib.error.URLError("Not Found")
    except urllib.error.URLError:
        raise RuntimeError(f"Failed to download chromedriver archive: {url}")
    archive = BytesIO(response.read())
    with zipfile.ZipFile(archive) as zip_file:
        # v115+ have files in subdirectories- need to adjust filename before extracting
        for zip_info in zip_file.infolist():
            if os.path.basename(zip_info.filename) == chromedriver_filename:
                zip_info.filename = chromedriver_filename
                zip_file.extract(zip_info, chromedriver_dir)
                break

Since you're setting chromedriver_version = 'custom', does the following always return false?

check_version( chromedriver_filepath, chromedriver_version ):