Hi ejplugge, are you a user of phono-semantic composition? Do you implement a features similar to:
not exactly, a phono-semantic tool, but visually similar kanji often share the same radicals (and phono-semantic as a consequence) :