Check what tag is used for `main`

check-main-tag.js

'use strict';

var async   = require('async');
var _       = require('lodash');
var cheerio = require('cheerio');
var phantom = require('phantom');

var userAgent = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_5) AppleWebKit/537.1 (KHTML, like Gecko) Chrome/21.0.1180.89 Safari/537.1';

var urls = [
  'https://www.google.com',
  'https://www.apple.com',
  'http://www.amazon.com',
  'https://twitter.com',
  'https://github.com',
  'https://www.facebook.com',
  'https://www.yahoo.com',
  'https://www.pinterest.com'
];

var mainSelectors = [
  'main',
  '#main',
  '.main',
  '[role="main"]'
];

function checkMainUsage(url, body) {
  var $ = cheerio.load(body);
  var mainFound = false;
  _.forEach(mainSelectors, function (selector) {
    if ($(selector).length > 0) {
      console.log(url + ' is using ' + selector);
      mainFound = true;
    }
  });
  if (!mainFound) {
    console.log(url + ' does not seem to use any main like tag');
  }
}

function runPhantom(ph, url, done) {
  async.waterfall([
    function (cb) {
      ph.createPage(function (page) { cb(null, page); });
    },
    function (page, cb) {
      page.set('settings.userAgent', userAgent);
      page.open(url, function () { cb(null, page); });
    },
    function (page, cb) {
      page.evaluate(function () {
        return document.querySelector('body').innerHTML;
      }, function (body) {
        checkMainUsage(url, body);
        cb();
      });
    }
  ], done);
}


phantom.create(function (ph) {
  async.each(urls, function (url, cb) {
    runPhantom(ph, url, cb);
  }, function () {
    ph.exit();
  });
});


// result
// https://www.google.com is using #main
// https://github.com is using [role="main"]
// https://twitter.com does not seem to use any main like tag
// https://www.yahoo.com is using [role="main"]
// https://www.facebook.com is using [role="main"]
// https://www.apple.com does not seem to use any main like tag
// https://www.pinterest.com does not seem to use any main like tag
// http://www.amazon.com does not seem to use any main like tag