Automatically insert ZWSP into text.

khmer-zwsp-insert.js

import MagicString from "magic-string";

/**
 * Automatically insert ZWSP into text.
 * @param {string} value 
 * @param {string | undefined} sep 
 * @returns {string}
 */
export function insertZeroWidthSpace(value, sep = '\u200b') {

  const regex = /[\u1780-\u17ff]+/gm
  const str = new MagicString(value);
  const segmenter = new Intl.Segmenter('km', { granularity: "word" });

  let result;
  while (result = regex.exec(value)) {

    const { index: start } = result;
    const end = regex.lastIndex;
    const text = result[0];
    const iterator = segmenter.segment(text);
    const segments = Array.from(iterator);

    // skipped because cannot segment
    if (segments.length <= 1) {
      continue;
    };

    const parts = new MagicString('');
    for (let i = 0; i < segments.length; i++) {
      const { segment, isWordLike } = segments[i];
      parts.append(segment);
      const nextSegment = segments[i + 1]
      if (isWordLike && nextSegment && nextSegment.isWordLike) {
        if (nextSegment.segment.length > 1) {
          parts.append(sep);
        }
      }
    }
    str.overwrite(start, end, parts.toString());
  }
  return str.toString();
}