@antora/assembler/lib/util/unconvert-inline-asciidoc.js

A JavaScript library that merges AsciiDoc content from multiple pages in an Antora site into assembly files and delegates to an exporter to convert those files to another format, such as PDF.

'use strict'

const ATTRIBUTE_REFERENCE_RX = /\{[a-z0-9_][a-z0-9_-]*\}/g
const STRICT_WORD_CHAR_RX = /[\p{L}\d_]/u
const WORD_CHAR_RX = /[\p{L}\d_;}:<>]/u

const MARK_FOR_TAG = { code: '`', em: '_', mark: '#', span: '#', strong: '*' }
const SKIP_SPAN = { icon: '<i ', image: '<img ' }

module.exports = (str) => {
  if (!str) return str
  let matchIndex = str.indexOf('<')
  if (!~matchIndex) return ~str.indexOf('{') ? str.replace(ATTRIBUTE_REFERENCE_RX, '\\$&') : str
  let current = { contents: '' }
  const stack = [current]
  let lastIndex = 0
  do {
    if (matchIndex > lastIndex) {
      const matched = str.slice(lastIndex, matchIndex)
      current.contents += ~matched.indexOf('{') ? matched.replace(ATTRIBUTE_REFERENCE_RX, '\\$&') : matched
    }
    const isCloseTag = str[++matchIndex] === '/' ? ++matchIndex : false
    let tagName = str.slice(matchIndex, (lastIndex = str.indexOf('>', matchIndex) + 1) - 1)
    if (isCloseTag) {
      const parent = current // TODO expect tagName to equal current.tagName
      stack.pop()
      current = stack[stack.length - 1]
      if (parent.mark) {
        let { contents, mark, id, role } = parent
        const attrlist = (id ? '#' + id : '') + (role ? '.' + role.replace(/ /g, '.') : '')
        if (
          current.mark === mark ||
          current.mark === '_' ||
          isWordChar(str.charAt(lastIndex), true) ||
          isWordChar(current.contents[current.contents.length - 1] || current.mark)
        ) {
          mark = mark.repeat(2)
        }
        current.contents += (attrlist ? '[' + attrlist + ']' : '') + mark + contents + mark
      } else {
        current.contents += parent.contents
      }
    } else {
      let attrs, attrlistIndex, role
      if (~(attrlistIndex = tagName.indexOf(' '))) {
        role = (attrs = parseAttrlist(tagName.slice(attrlistIndex))).class
        tagName = tagName.slice(0, attrlistIndex)
      }
      if (tagName === 'img') {
        current.contents += 'image:' + attrs.src + '[' + attrs.alt + ']'
      } else if (tagName === 'i' && current.tagName === 'span' && current.role === 'icon') {
        current.contents += 'icon:' + role.slice(6) + '[]'
        lastIndex += 4
      } else {
        let check, mark
        const id = attrs?.id
        if (tagName !== 'span' || id || (role && !((check = SKIP_SPAN[role]) && str.startsWith(check, lastIndex)))) {
          mark = MARK_FOR_TAG[tagName]
        }
        stack.push((current = { tagName, role, id, mark, contents: '' }))
      }
    }
  } while (~(matchIndex = str.indexOf('<', lastIndex)))
  const rest = str.slice(lastIndex)
  if (rest) current.contents += ~rest.indexOf('{') ? rest.replace(ATTRIBUTE_REFERENCE_RX, '\\$&') : rest
  return current.contents
}

function parseAttrlist (str) {
  let lastIndex = 0
  const attrs = {}
  while (str.charAt(lastIndex++) === ' ') {
    const spaceIndex = str.indexOf(' ', lastIndex)
    const equalsIndex = str.indexOf('=', lastIndex)
    if (~spaceIndex && spaceIndex < equalsIndex) {
      attrs[str.slice(lastIndex, (lastIndex = spaceIndex))] = true
    } else if (~equalsIndex) {
      const name = str.slice(lastIndex, equalsIndex)
      const valueIndex = equalsIndex + 1
      attrs[name] =
        str.charAt(valueIndex) === '"'
          ? str.slice(valueIndex + 1, (lastIndex = str.indexOf('"', valueIndex + 1) + 1) - 1)
          : str.slice(valueIndex, (lastIndex = ~spaceIndex ? spaceIndex : str.length))
    } else {
      attrs[str.slice(lastIndex)] = true
      break
    }
  }
  return attrs
}

function isWordChar (str, strict) {
  if (!str) return false
  return (strict ? STRICT_WORD_CHAR_RX : WORD_CHAR_RX).test(str)
}