convert_xml.rb

#
# convert xml data to hash with hpricot
# 
# by xianhua.zhou@gmail.com
require 'rubygems'
require 'hpricot'
require 'ap'

class ConvertXml
  def initialize(filename)
    @filename = filename
  end 

  def get_valid_nodes(nodes)
    nodes.find_all{|it|it.is_a? Hpricot::Elem}
  end 

  def convert(xml_nodes = nil, parent_node = nil)

    if xml_nodes.nil?
      xml_nodes = Hpricot(File.read(@filename)).search('/')
      key = :root
    else
      key = parent_node.name
    end 

    # results
    nodes = {}

    get_valid_nodes(xml_nodes).each do |node|
      children = get_valid_nodes(node.children)

      nodes[key] = [] if nodes[key].nil?
      nodes[key] <<   \
        if children.size > 0
          convert(children, node)
        else
          node
        end
    end

    nodes
  end
end

cx = ConvertXml.new('users.xml')
ap cx.convert[:root]

users.xml

<?xml version="1.0" encoding="utf-8" ?>
<people>
  <person>
    <hi>
      <name>zhou</name>
      <info>test</info>
      <age>12</age>
    </hi>
    <hello>world</hello>
  </person>
  <person>
    <hi>
      <name>li</name>
      <info>hello</info>
      <age>10</age>
    </hi>
  </person>
</people>

users2.xml

<?xml version="1.0" encoding="utf-8" ?>
<people>
  nothing
</people>
<people>
  <user>Me</user>
</people>