C# System.Windows.Automation get element text

Question

I am trying to get text/labels from application controls with Automation in C#.

So far I am able to obtain AutomationElement tree of application (for example Notepad) with this function:

    private void WalkControlElements(AutomationElement rootElement, TreeNode treeNode)
    {
        AutomationElement elementNode = TreeWalker.ContentViewWalker.GetFirstChild(rootElement);;

        while (elementNode != null)
        {
            TreeNode childTreeNode = treeNode.Nodes.Add(elementNode.Current.ControlType.LocalizedControlType);

            // here I want to get text from 'elementNode'

            WalkControlElements(elementNode, childTreeNode);
            elementNode = TreeWalker.ControlViewWalker.GetNextSibling(elementNode);
        }
    }

I tried to follow this article http://msdn.microsoft.com/en-us/library/ms788751(v=vs.110).aspx but it only can get text attributes as font name, font weight and so on.

Could anybody point me to the right procedure how to get element text with Automation?

Answer 1

That sample is showing you how to get text attributes, i.e. information about the display of the text in the UI, not the actual displayed text. Getting all the actual displayed text for a general application is more difficult that it might first appear.

It is made difficult by the fact that there are several ways get text and there is inconsistent support by applications and controls. There are two patterns that are of some use, ValuePattern and TextPattern. By convention the Name property contains text displayed to the user however adherence to this is inconsistent. Below is a helper method that I've used in UI automation for testing. It basically goes through those patterns checking the control for support and falls back to the Name.

public static class AutomationExtensions
{
    public static string GetText(this AutomationElement element)
    {
        object patternObj;
        if (element.TryGetCurrentPattern(ValuePattern.Pattern, out patternObj))
        {
            var valuePattern = (ValuePattern)patternObj;
            return valuePattern.Current.Value;
        }
        else if (element.TryGetCurrentPattern(TextPattern.Pattern, out patternObj))
        {
            var textPattern = (TextPattern)patternObj;
            return textPattern.DocumentRange.GetText(-1).TrimEnd('\r'); // often there is an extra '\r' hanging off the end.
        }
        else
        {
            return element.Current.Name;
        }
    }
}

This takes care of getting the text out of simple controls like labels, textboxes (both vanilla textbox and richtextbox), and buttons. Controls like listboxes and comboboxes (esp. in WPF) can be tricker because their items can be virtualized so they may not exist in the automation tree until the user interacts with them. You may want to filter and call this method only on certain UI Automation control types like Edit, Text, and Document which you know contain text.

Answer 2

Mike Zboray answer works fine. In case you have access to pattern-Matching, here is the same (condensed) code :

public static class AutomationExtensions
{
    public static string GetText(this AutomationElement element)
    => element.TryGetCurrentPattern(ValuePattern.Pattern, out object patternValue) ? ((ValuePattern)patternValue).Current.Value
        : element.TryGetCurrentPattern(TextPattern.Pattern, out object patternText) ? ((TextPattern)patternText).DocumentRange.GetText(-1).TrimEnd('\r') // often there is an extra '\r' hanging off the end.
        : element.Current.Name;
}