Remove emojis from text

Removes emoji and pictographs, including skin-tone modifiers and the joined sequences that look like one character but are made of several.

A family emoji is genuinely five codepoints joined by zero-width joiners, and a skin-toned thumbs-up is two. Matching a fixed codepoint range — which is what most emoji-removing snippets do — deletes part of a sequence and leaves the remains behind as stray characters. Matching on the Extended_Pictographic property plus the joiners handles the whole cluster, and keeps working as Unicode adds new emoji rather than needing the range updated. Worth knowing why you might want this: emoji in a filename, a CSV or a database column with the wrong collation cause real problems, and stripping them is often easier than fixing the encoding.