yes - the node uris are converted via the Fusion object UrisConverter or smt. In its php implementation, you fill find a funny regex which preg_replace_callback’s everything that looks like a node id. You would need to adjust it to find out if you are in an attribute (href) or in a text. “Parsing” html with regex is quite dangerous, but might work mostly as ckeditor only generates valid and predictable html. Otherwise you could also use the XML Dom parser from PHP.