Bible.com - Strips all HTML that comes back from devotional content and adds spacing and escaping in its proper place.

html-to-text.js

/**
 * Converts HTML from a Bible plan to String
 * - REQUIRES: entities
 */
const entities = require('entities')

/* source: https://stackoverflow.com/a/1144788 */
function replaceAll(str, find, replace) {
  return str.replace(new RegExp(find, 'g'), replace)
}

function replaceBlockTagsWithNewLines(content) {
  // puts a new line at the end of all block content tags
  return replaceAll(content, '</div>|</p>|</h[1-6]+>| *<br ?/?>+|</li>', '\n')
}

function replaceListTagsWithBullets(content) {
  return replaceAll(content, '<li>', '• ')
}

function stripHTMLTags(content) {
  return replaceAll(content, '<[^>]+>', '')
}

function cleanupSpacing(content) {
  let updatedContent = content

  /* replace new lines
   * \r = CR (Carriage Return) → Used as a new line character in Mac OS before X
   * \n = LF(Line Feed) → Used as a new line character in Unix / Mac OS X
   * \r\n = CR + LF → Used as a new line character in Windows
   * source: https://stackoverflow.com/a/15433225
   */
  updatedContent = replaceAll(updatedContent, '\r\n|\r', '\n')

  // replace multiple new lines with only two
  updatedContent = replaceAll(updatedContent, '\n+', '\n\n')

  // replace non-breaking spaces or multiple spaces with a regular space
  updatedContent = replaceAll(updatedContent, '\u00a0| {2,}', ' ')

  // trim content
  updatedContent = updatedContent.trim()

  return updatedContent
}

/**
 * Strips all HTML that comes back from devotional content
 * and adds spacing and escaping in its proper place.
 *
 * This is used to convert Bible devotional content to be
 * ready for the Stories devotional module, so that the
 * text can closely mimmic the preview on iOS.
 *
 * Code from Bryan Montz and converted to JS by Cam Pak.
 *
 * @param {string} html
 */
function htmlToText(html) {
  let content = html

  content = entities.decodeHTML(content)
  content = replaceBlockTagsWithNewLines(content)
  content = replaceListTagsWithBullets(content)
  content = stripHTMLTags(content)
  content = cleanupSpacing(content)

  return content
}

export default htmlToText