Highlight whole words, omit HTML

Question

I am writing some C# code to parse RSS feeds and highlight specific whole words in the content, however, I need to only highlight words that are outside HTML. So far I have:

string contentToReplace = "This is <a href=\"test.aspx\" alt=\"This is test content\">test</a> content";

string pattern = "\b(this|the|test|content)\b";

string output = Regex.Replace(contentToReplace, pattern, "<span style=\"background:yellow;\">$1</span>", RegexOptions.Singleline | RegexOptions.IgnoreCase);

This works fine, except it will highlight the word "test" in the alt tag. I can easily write a function that strips the HTML, then does the replace, but I need the keep the HTML to display the content.

Answer 1

If the input is valid XHTML/XML you could parse it to a tree structure (DOM/XLinq), recursively walk through the tree, replace all keyword occurrences in text nodes and finally serialize the tree structure back to a string.

Untested pseudo-code:

XNode Highlight(XElement element, List<string> keywords)
{
    var result = new XElement(element.Name);
    // copy element attributes to result

    foreach (var node in element)
    {
        if (node.Type == NodeType.Text)
        {
            var value = node.Value;
            // while value contains keyword
            // {
            //      add substring before keyword in value to result
            //      add new XElement with highlighted keyword to result
            //      remove consumed substring from value
            // }
        }
        else if (node.Type == NodeType.Element)
        {
            result.Add(Highlight((XElement)node, keywords));
        }
        else
        {
            result.Add(node);
        }
    }

    return result;
}

var output = Highlight(XElement.Parse(input), new List<string> {...}).ToString();

Answer 2

Another solution if you have valid XML but don't want to parse it: First split the input string into parts such that each part contains only a tag or text but not both. For example:

"This is ",
"<a href=\"test.aspx\" alt=\"This is test content\">",
"test"
"</a>"
" content"

Then iterate through the parts and apply your regex only to strings that don't start with '<'. Finally, join all parts back into a single string.

Answer 3

Here's a basic one.

private void Form1_Load(object sender, EventArgs e)
    {
        string contentToReplace = "This is <a href=\"test.aspx\" alt=\"This is test content\"> hello test world</a> content";

        string pattern = @"(>{1}.*)(test)(.*<{1})";

        string output = Regex.Replace(contentToReplace, pattern, "$1<span>$2</span>$3", RegexOptions.Singleline | RegexOptions.IgnoreCase);

        //output is :
        //This is <a href="test.aspx" alt="This is test content"> hello <span>test</span> world</a> content


        MessageBox.Show(output);
        Close();
    }