selenium webdriver (chromedriver) and accessing sh

2020-02-07 06:49发布

问题:

I'm testing out a new application that uses shadow dom like so:

 #shadow-root (open)
    <div class="th_filePicker">
        <div class="th_fp_header">
            <div class="th_fp_title" role="heading" aria-level="1" data-l10n-id="th_fp_title">Select Image</div>
                <div class="th_fp_Close"><button class="close-popup" data-l10n-id="close_popup" title="Close"></button></div>
        </div>
    </div>

Does anyone have any idea on how I can access the elements in the file picker control - specifically, the close icon?

回答1:

One way would be to use a piercing CSS selector (/deep/ or >>>). Though it's not supported by all the browsers and it may be removed in a future release.

This one should give you the close button with Chrome 62:

driver.findElement(By.css("* /deep/ button[title='Close']"))


回答2:

It's possible, but it will take a couple steps. As a preliminary, check out this page about accessing shadow dom. I found it really informative.

Start with two methods to get the shadow dom element:

private WebElement shadowDom;

private WebElement expandRootElement(WebElement element) {
    WebElement ele = (WebElement) ((JavascriptExecutor) driver).executeScript("return arguments[0].shadowRoot",element);
    return ele;
}
private void findByShadowRoot(WebDriver driver) {
    shadowDom = expandRootElement(driver.findElement(By.id("whatEverTheShadowDomIdIs")));
}

From there, you create methods as a pseudo POM

private WebElement findByShadowButton() {
    findByShadowRoot(driver);
    return shadowDom.findElement(By.cssSelector("div.th_fp_Close"));
}

Basically the first two methods are for creating a starting point, and then all the other methods call those methods and say, "from this starting point, find the element beneath it".

Then you can statements like:

findByShadowButton().click();


回答3:

You can try this "heavy" approach (C# but depending on your language it can be something like that):

public IWebElement DeepFind(By search)
{
    try
    {
        // search a result in the main dom
        return Driver.FindElement(search);
    }
    catch (NoSuchElementException)
    {
        // if nothing we will take a look to the shadow dom(s)
        var shadowRoots = new List<IWebElement>();
        try
        {
            // will use the recursive method that search for all shadow roots
            ListShadowRoots(search, Driver.FindElements(By.XPath("//*")), shadowRoots);
        }
        catch (NoSuchElementException)
        {
            //
        }
        // return the first element that match the By search
        return shadowRoots.FirstOrDefault(s => s.FindElement(search) != null);
    }
}

private void ListShadowRoots(By search, ReadOnlyCollection<IWebElement> elements, List<IWebElement> shadowRoots)
{
    elements.ToList().ForEach(e =>
    {
        var jsResult = (IWebElement)ExecuteJavascript("return arguments[0].shadowRoot", new object[] { e });
        if (jsResult != null)
        {
            shadowRoots.Add(jsResult);
            try
            {
                ListShadowRoots(search, jsResult.FindElements(By.XPath("//*")), shadowRoots);
            }
            catch (NoSuchElementException)
            {
                //
            }
        }
    });
}

private object ExecuteJavascript(string code, object[] args)
{
    IJavaScriptExecutor js = (IJavaScriptExecutor)Driver;
    js.ExecuteScript(code, args);
}

Driver is the web driver (IWebDriver)

Performances are not so bad and it does the job ;) Hope that it can help



回答4:

I took the information from MivaScott's answer and created a recursive method for my solution, I thought it could be of use to other people so here it is. I was using this to click a play button in a video player.

All you need to provide is a string array of the CSS selectors of the shadow-root. The method will return the final shadow root element, so you can add another selector onto the end (in my case, an svg). Please see my example:

The Player's Shadow-Root Structure

public IWebElement PlayButton {

        get {

            string[] shadowRootSelectors = { "apc-controls", "apc-control-footer", "apc-toggle-play", "apc-icon-play" };

            return FindShadowRootElementRecursive(shadowRootSelectors).FindElement(By.CssSelector("svg"));

        }
        set {
        }

}

And the recursive method itself:

public IWebElement FindShadowRootElementRecursive(string[] selectors = null, IWebElement element = null) {

            IWebElement root = null;
            IWebElement selectorElement = null;
            bool baseCase = false;

            //Get the first selector from the array
            string selector = selectors[0];

            if (selectors.Length == 1)
            {

                baseCase = true;

            }
            else {

                //If there are more selectors, then remove this selector and recurse with the rest
                selectors = selectors.Where(w => w != selectors[0]).ToArray();

            }

            //If this is the first call...
            if (element == null)
            {

                //Use the driver to select the element
                selectorElement = Driver.FindElement(By.CssSelector(selector));

            }
            else {

                //Otherwise, use the previously found element
                selectorElement = element.FindElement(By.CssSelector(selector));

            }

            //Get the shadow root
            root = (IWebElement)((IJavaScriptExecutor)Driver).ExecuteScript("return arguments[0].shadowRoot", selectorElement);

            if (baseCase)
            {

                return root;

            }
            else {

                //Recurse
                root = FindShadowRootElementRecursive(selectors, root);

            }

            return root;

        }

I then clicked the button like so:

PlayButton.Click();