Searching for tags while parsing Wordpress XML with Nokogiri

[admin]

I have an XML file of a Wordpress blog that consists of quotes:

<item>
  <title>Brothers Karamazov</title>
  <content:encoded><![CDATA["I think that if the Devil doesn't exist and, consequently, man has created him, he has created him in his own image and likeness."]]></content:encoded>
  <category domain="post_tag" nicename="dostoyevsky"><![CDATA[Dostoyevsky]]></category>
  <category domain="post_tag" nicename="humanity"><![CDATA[humanity]]></category>
  <category domain="category" nicename="quotes"><![CDATA[quotes]]></category>
  <category domain="post_tag" nicename="the-devil"><![CDATA[the Devil]]></category>
</item>

The things I'm trying to extract are title, author, content and tags. Here's my code so far:

require "rubygems"
require "nokogiri"

doc = Nokogiri::XML(File.open("/Users/charliekim/Downloads/quotesfromtheunderground.wordpress.2013-04-14.xml"))

doc.css("item").each do |item|
  title   = item.at_css("title").text
  tag     = item.at_xpath("category").text
  content = item.at_xpath("content:encoded").text

  #each post will later be pushed to an array, but I'm not worried about that yet, so for now....
  puts "#{title} #{tag}"
end

I'm struggling to get all the tags from each

item

. I'm getting returns of something like

Brothers Karamazov Dostoyevsky

. I'm not worried about how it's formatted as it's only a test to see that it's picking things up correctly. Anyone know how I can go about this?

I also want to make tags that are capitalized = Author, so if you know how to do that it would help, too, although I haven't even tried it yet.

<hr>

EDIT: I changed the code to this:

doc.css("item").each do |item|
  title   = item.at_css("title").text
  content = item.at_xpath("content:encoded").text
  tag     = item.at_xpath("category").each do |category|
        category
  end

  puts "#{title}: #{tag}"
end

which returns:

Brothers Karamazov: [#<Nokogiri::XML::Attr:0x80878518 name="domain" value="post_tag">,     #<Nokogiri::XML::Attr:0x80878504 name="nicename" value="dostoyevsky">]

and which seems a bit more manageable. It screws up my plans for taking the Author from a capitalized tag, but, well, it's not so big of a deal. How could I pull just the second

value

?