Parser regards a half of surrogate pair as “emoji”

[yk1979]

if target text has a half of surrogate pair (like 0xFE0F), parser regards its type as "emoji"
twemoji-parser v12.1.3

Actual behavior

import { parse } from 'twemoji-parser';

const str = 'Twemoji✨️Parser';
// this text has invisible character(0xFE0F) after sparkles emoji
escape(str); // 'Twemoji%u2728%uFE0FParser'

const entities = parse(str);

/*
entities = [
  {
    url: 'https://twemoji.maxcdn.com/v/latest/svg/2728.svg',
    indices: [ 7, 8 ],
    text: '✨',
    type: 'emoji'
  },
  // url is empty!!!! but type: emoji
  { url: '', indices: [ 8, 9 ], text: '️', type: 'emoji' }
]
*/

Expected behavior

I'd be grateful if the result is like this！

/*
entities = [
  {
    url: 'https://twemoji.maxcdn.com/v/latest/svg/2728.svg',
    indices: [ 7, 8 ],
    text: '✨',
    type: 'emoji'
  }
]
*/

The cause(I think)

when the text which has an invisible character is checked by emojiRegex, the result includes an invisible character.
https://github.com/twitter/twemoji-parser/blob/master/src/index.js#L34

const str = 'Twemoji✨️Parser';
const res1 = emojiRegex.exec(str);
res1[0] // "✨"

const res2 = emojiRegex.exec(str);
res2[0] // "️"
escape(res2[0]) // "%uFE0F"