crap thông tin sản phẩm của maihoang.com.vn

crap.js

const got = require('got')
const cheerio = require('cheerio')

let parseProductFromLink = (url) => {
  url = encodeURI(url)
  got(url)
    .then(response => cheerio.load(response.body))
    .then($ => {
      $('.pro-item .pro-img a').map((i, a) => 
        parseProductInfo($(a).attr('href'))
      )
      return $
    })
    .then($ => {
      if ($('.pagination').find('span').next('a').text() != 'Trang cuối') {
        let urlNext = $('.pagination').find('span').next('a').attr('href')
        parseProductFromLink(urlNext)
      }
    })
    .catch(error => console.log(error))

}

let parseProductInfo = (url) => {
  url = encodeURI(url)
  console.log(url)
  got(url)
    .then(response => cheerio.load(response.body))
    .then($ => {
      let ten = $('.product-home h1').first().text()
      let gia = $('.product-home .price').text().replace(/\D/g, '')
      let hinhanh = $('.product-home .tiny-img img').map((i, img) => $(img).attr('src')).get()
      let loaisanpham = $('.thanhdieuhuong li').last().prev().text()
      
      // bug. 
      // work well on js browser
      // nodejs not work
      // you can try here https://regex101.com/r/WIcJ8S/1
      let hangsanxuat = (/Hãng sản xuất: (\w+)<br>/g).exec($('.thongtinsanpham .desc-info').html())[1]
      console.log(hangsanxuat)
    
      // console.log({
      //   ten,
      //   gia,
      //   hinhanh,
      //   loaisanpham,
      //   hangsanxuat
      // })
    })
    .catch(error => console.log(error))
}


parseProductFromLink('http://maihoang.com.vn/main-bo-mach-chu')