Java regex to enable link inside text except already enclosed tag

Question

Suppose a user input text that contains HTML and maybe links , I want to enable links , and make already_tag_closed urls intact .

(I know there are a lot of regex url pattern questions asked , but I cannot find this solution)

for example :

String urlRegex = "((https?|ftp|gopher|telnet|file):((//)|(\\\\))+[\\w\\d:#@%/;$()~_?\\+-=\\\\\\.&]*)";
Pattern urlPattern = Pattern.compile(urlRegex, Pattern.CASE_INSENSITIVE);

String s = ...
urlPattern.matcher(s).replaceAll("<a href='$0' target='_blank'>$0</a>")

It can translate "google https://google.com" to google <a href='https://google.com' target='_blank'>https://google.com</a> , good.

But if string is

"<a href=\"http://www.google.com/\">google</a> " +
" http://www.google.com/  " +
" <a href=\"https://facebook.com/\">facebook</a> " +
" https://facebook.com ";

It will become

<a href="<a href='http://www.google.com/' target='_blank'>http://www.google.com/</a>">google</a>  <a href='http://www.google.com/' target='_blank'>http://www.google.com/</a>   <a href="<a href='https://facebook.to/' target='_blank'>https://facebook.to/</a>">facebook</a>  <a href='https://facebook.com' target='_blank'>https://facebook.com</a> 

It should not touch values in href , so I change the urlRegex to :

urlRegexExceptAnchor = "(?<!\\<a\\ href=\")(http|https):\\/\\/[^ ]*"; 

Well , it can handle text mixed with anchor tags.

But , if the text includes iframe , it will fail again :

<iframe src="https://www.facebook.com/plugins/post.php?href=https%3A%2F%2Fwww.facebook.com%2Fnytimes%2Fposts%2F10151112309519999&width=500" width="500" height="525" style="border:none;overflow:hidden" scrolling="no" frameborder="0" allowTransparency="true"></iframe>

becomes

<iframe src="<a href='https://www.facebook.com/plugins/post.php?href=https%3A%2F%2Fwww.facebook.com%2Fnytimes%2Fposts%2F10151112309519999&width=500"' target='_blank'>https://www.facebook.com/plugins/post.php?href=https%3A%2F%2Fwww.facebook.com%2Fnytimes%2Fposts%2F10151112309519999&width=500"</a> width="500" height="525" style="border:none;overflow:hidden" scrolling="no" frameborder="0" allowTransparency="true"></iframe>

It's invalid again.

I think I'll face more and more situation , because There are a lot of tags that accepts URLs . I cannot just escape a or iframe tags...

The text is input by user , sure I can filter out some invalid tags , such as form , head , input ... , but there are still a lot of tags to process... ( or even inlined css background url )

What I can think of now is to use something like JSoup to transfer the whole text to html doc , and process the textNode one by one. But I think that's too overkill. (Each page display will invoke JSoup ... )

Is there any easier ways to achieve this ?

Answer 1

To anyone facing similar problems , this is my JSoup solution :

  private static void processNode(Node node) {
    if (node instanceof TextNode) {

      Node parent = node.parent();
      if (parent != null && (StringUtils.equalsAnyIgnoreCase(parent.nodeName(),
        "a", "iframe", "embed" , "img" , "object" , "script" , "video" , "applet"))) {
        logger.debug("parent = {} , skipped", parent.nodeName());
      }
      else {
        TextNode textNode = (TextNode) node;

        String text = textNode.text();
        text = urlPattern.matcher(text).replaceAll("<a href='$0' target='_blank'>$0</a>");

        TextNode r = new TextNode(text , null);
        node.replaceWith(r);
      }
    } else if (node instanceof Element) {
      Element ele = (Element) node;
      for (Node childNode : ele.childNodes()) {
        processNode(childNode);
      }
    }
  }

It works fine ... (for now)