A simple parser for pulling structured data from the HTML of an Instagram profile.

instagram-html-parser.js

const jsdom = require('jsdom')
const {
  JSDOM
} = jsdom

function toJSDOM(responseBody) {
  return new JSDOM(responseBody)
}

/**
 * Pull structured data from the HTML of an Instagram page (e.g. www.instagram.com/theshaderoom)
 */
module.exports = function parse(body) {
  const rawDom = toJSDOM(body)
  const dom = rawDom.window.document.body

  function formatNumber(text, labelText) {
    let num = text.replace(labelText, '').replace(/,/g, '')

    // '17.4m'
    if (text.includes('m')) {
      // 17.4
      // expand out millions
      condensedNum = Number(condensedNum.replace('m', '')) * 1000000
    }

    return Number(num.trim())
  }

  const blocks = {
    posts: null,
    followers: null,
    following: null
  }

  // Map description row chunks (mostly lacking good semantic CSS) into structured data
  Array.from(dom.querySelectorAll('ul li'))
    .map(node => node.textContent)
    .forEach(textNode => {
      // '128,123 posts'
      if (textNode.includes(' posts')) {
        blocks.posts = formatNumber(textNode, ' posts')

        // '27.4m followers'
      } else if (textNode.includes(' followers')) {
        blocks.followers = formatNumber(textNode, ' followers', '')

        // '128,123 following'
      } else if (textNode.includes(' following')) {
        blocks.following = formatNumber(textNode, ' following', '')
      }
    })

    return {
      name: dom.querySelector('h1').textContent,
      isVerified: Boolean(dom.querySelector('.coreSpriteVerifiedBadge')),
      ...blocks
    }
}