twl-linker/usfm-alignment-remover.js

Biblical Semantic Linker - Uses the biblical context database to create semantic links between USFM Bible text and biblical articles with confidence scoring.

/* eslint-disable no-async-promise-executor, no-throw-literal */

// Note: This version doesn't use usfm-js to avoid external dependencies
// It implements a simple USFM alignment remover for the specific case

/**
 * Simple USFM alignment remover that strips alignment markers while preserving text
 * @param {string} usfmContent - The USFM content with potential alignment data
 * @return {string} - Clean USFM content without alignment markers
 */
export const removeAlignments = (usfmContent) => {
  if (!usfmContent) return '';

  let cleanContent = usfmContent;

  // Remove word-level alignment markers like \w word|lemma="lemma" strong="H1234"\w*
  cleanContent = cleanContent.replace(/\\w\s+([^|\\]+)\|[^\\]*\\w\*/g, '$1');

  // Remove milestone markers like \zaln-s | \zaln-e\*
  cleanContent = cleanContent.replace(/\\zaln-[se][^\\]*\\?\*?/g, '');

  // Remove other alignment-related markers
  cleanContent = cleanContent.replace(/\\k-[se][^\\]*\\?\*?/g, '');

  // Remove empty lines that might result from marker removal
  cleanContent = cleanContent.replace(/\n\s*\n\s*\n/g, '\n\n');

  // Clean up any remaining alignment syntax patterns
  cleanContent = cleanContent.replace(/\|[^\\]*(?=\\)/g, '');

  return cleanContent.trim();
};