Function cleanHtml

lib/html.js:40–74

(html)

const DROP_ATTRS = new Set(['style'])

function cleanHtml(html) {
  const document = parse(html)

  function walk(node) {
    if (!node) return false

    if (DROP_TAGS.has(node.nodeName) || (node.nodeName === 'script' && !(node.attrs || []).some(a => a.name === 'src'))) {
      const parent = node.parentNode
      const idx = parent.childNodes.indexOf(node)
      if (idx >= 0) parent.childNodes.splice(idx, 1)
      return true
    }

    if (node.attrs) {
      node.attrs = node.attrs.filter(attr => {
        if (DROP_ATTRS.has(attr.name)) return false
        if (attr.name === 'class') {
          attr.value = filterClassValue(attr.value)
          if (!attr.value) return false
        }
        return true
      })
    }

    if (node.childNodes) {
      for (let i = node.childNodes.length - 1; i >= 0; i--) {
        walk(node.childNodes[i])
      }
    }
    return false
  }

  walk(document)
  return serialize(document)
}

const defaultHtmlOpts = {
  interactiveElements: ['a', 'input', 'button', 'select', 'textarea', 'option'],