Question

I’m trying to parse this website using Ruby and Nokogiri:

Here’s my code:

require 'nokogiri'
require 'open-uri'

class StreamsController < ApplicationController
  def index
  end

  def updateall
    doc = Nokogiri::HTML(open('http://www.own3d.tv/game/League+of+Legends'))

    # Grab all the Live streams from the front page.
    doc.css('div#top_live .VIDEOS-1grid-box').each do |stream|
      s = Stream.new

      # Parse the URL.
      s.url = stream.css('a.small-tn')['href']
    end
  end
end

On the # Parse the URL bit, I get the error Cannot convert String to Integer.

I’m kind of confused on how to use Nokogiri for this simple use case.

How can I get the href attribute of each link inside each |stream| object?

Answer 1

stream.css('a.small-tn') will return a collection of nodes. So calling ['href'] on the collection isn’t going to work since the collection acts as an array, and it thinks you’re trying to access an element at a certain index (hence the error). Rather, you need to decide if you want to iterate through them, or just grab the first:

s.url = stream.css('a.small-tn').first['href']

If you want to make this a bit more safe, you can check for nils:

node = stream.css('a.small-tn').first
s.url = node['href'] if node

Or you can use the at_css helper (as @AJcodez) pointed out, which does the same thing.