test_nokogiri_sax_parser.rb

require "rubygems"
require "nokogiri"
require "eventmachine"

##
# This is the XML SAX Parser that accepts "pushed" content used by Babylon Gem.
class XmppParser < Nokogiri::XML::SAX::Document
  
  ##
  # Initialize the parser and adds the callback that will be called upen stanza completion
  def initialize(&callback)
    @callback = callback
    super()
    @parser = Nokogiri::XML::SAX::Parser.new(self)
    @doc = nil
    @elem = nil
  end
  
  ##
  # Parses the received data
  def parse(data)
    @parser.parse data
  end

  ## 
  # Called when the document received in the stream is started
  def start_document
    @doc = Nokogiri::XML::Document.new
  end
  
  ##
  # Adds characters to the current element (being parsed)
  def characters(string)
    @elem.add_child(Nokogiri::XML::Text.new(string, @doc))
  end

  ##
  # Instantiate a new current Element, adds the corresponding attributes and namespaces
  # The new element is eventually added to a parent element (if present).
  # If this element is the first element (the root of the document), then instead of adding it to a parent, we add it to the document itself. In this case, the current element will not be terminated, so we activate the callback immediately.
  def start_element(qname, attributes = [])
    e = Nokogiri::XML::Element.new(qname, @doc)
    add_namespaces_and_attributes_to_node(attributes, e)

    # Adding the newly created element to the @elem that is being parsed, or, if no element is being parsed, then we set the @root and the @elem to be this newly created element.
    @elem = @elem ? @elem.add_child(e) : (@root = e)
    
    if @elem.name == "stream:stream"
      # Should be called only for stream:stream.
      # We re-initialize the document and set its root to be the doc.
      # Also, we activate the callback since this element  will never end.
      @doc = Nokogiri::XML::Document.new
      @doc.root = @root = @elem
      @callback.call(@elem)
    end
  end

  ##
  # Terminates the current element and calls the callback
  #
  # BUG : We might have a problem of missed stanzas : (this was taken from the log, the second iq stanza hasn't been parsed!)
  # DEBUG BABYLON: RECEIVED : <iq from='julien@superfeedr.com/babylon_client_6970' to='firehoser.superfeedr.com' type='set' id='1419'>
  #           <pubsub xmlns='http://jabber.org/protocol/pubsub'>
  #               <subscribe jid='julien@superfeedr.com' node='http://feeds2.feedburner.com/NotifixiousFoundersBlog'/>
  #           </pubsub>
  #       </iq><iq from='julien@superfeedr.com/babylon_client_6970' to='firehoser.superfeedr.com' type='set' id='7858'>
  #           <pubsub xmlns='http://jabber.org/protocol/pubsub'>
  #               <unsubscribe jid='julien@superfeedr.com' node='http://feeds2.feedburner.com/NotifixiousFoundersBlog'/>
  #           </pubsub>
  #       </iq>
  # DEBUG BABYLON: PARSED : <iq from="julien@superfeedr.com/babylon_client_6970" to="firehoser.superfeedr.com" type="set" id="1419">
  #           <pubsub xmlns:xmlns="http://jabber.org/protocol/pubsub" xmlns:http://jabber.org/protocol/pubsub="xmlns">
  #               <subscribe jid="julien@superfeedr.com" node="http://feeds2.feedburner.com/NotifixiousFoundersBlog"/>
  #           </pubsub>
  #       </iq>
  # INFO BABYLON: ROUTING TO SubscriptionsController::subscribe
  def end_element(name)
    if @elem
      if @elem.parent == @root
        @callback.call(@elem) 
        # And we also need to remove @elem from its tree
        @elem.unlink 
        # And the current elem is the next sibling or the root
        @elem = @root 
      else
        @elem = @elem.parent 
      end
    end
  end
  
  private
  
  ##
  # Adds namespaces and attributes. Nokogiri passes them as a array of [name, value, name, value]...
  def add_namespaces_and_attributes_to_node(attrs, node) 
    (attrs.size / 2).times do |i|
      name, value = attrs[2 * i], attrs[2 * i + 1]
      if name =~ /xmlns/
        node.add_namespace(name, value)
      else
        node.set_attribute name, value
      end
    end
  end
  
end


# We're simulating an XMPP session to show that the last element it not parsed correctly... 
content = ["<?xml version='1.0'?><stream:stream xmlns:stream='http://etherx.jabber.org/streams' xmlns='jabber:component:accept' id='314321979' from='superfeedr.com'>", 
  "<handshake/>",
  "<iq from='julien@superfeedr.com/babylon_client_6970' to='firehoser.superfeedr.com' type='set' id='1419'>
          <pubsub xmlns='http://jabber.org/protocol/pubsub'>
              <subscribe jid='julien@superfeedr.com' node='http://feeds2.feedburner.com/NotifixiousFoundersBlog'/>
          </pubsub>
      </iq><iq from='julien@superfeedr.com/babylon_client_6970' to='firehoser.superfeedr.com' type='set' id='7858'>
          <pubsub xmlns='http://jabber.org/protocol/pubsub'>
              <unsubscribe jid='julien@superfeedr.com' node='http://feeds2.feedburner.com/NotifixiousFoundersBlog'/>
          </pubsub>
      </iq>"]

@parser = XmppParser.new do |element|
  puts "========="
  puts "PARSED : "
  puts "#{element} "
  puts ""
end

EM.run do
  index = 0
  EM.add_periodic_timer(1) {
    if index < content.size
      @parser.parse(content[index])
      index = index + 1
    else
      EM.stop_event_loop
    end
  }
end

# Actual output is :

# $ ruby saxparser.rb 
# =========
# PARSED : 
# <stream:stream xmlns:xmlns:stream="http://etherx.jabber.org/streams" xmlns:xmlns="jabber:component:accept" xmlns:stream:http://etherx.jabber.org/streams="xmlns:stream" xmlns:jabber:component:accept="xmlns" id="314321979" from="superfeedr.com"/> 
# 
# =========
# PARSED : 
# <handshake/> 
# 
# =========
# PARSED : 
# <iq from="julien@superfeedr.com/babylon_client_6970" to="firehoser.superfeedr.com" type="set" id="1419">
#           <pubsub xmlns:xmlns="http://jabber.org/protocol/pubsub" xmlns:http://jabber.org/protocol/pubsub="xmlns">
#               <subscribe jid="julien@superfeedr.com" node="http://feeds2.feedburner.com/NotifixiousFoundersBlog"/>
#           </pubsub>
#       </iq> 


# While it should probably be :

# $ ruby saxparser.rb 
# =========
# PARSED : 
# <stream:stream xmlns:xmlns:stream="http://etherx.jabber.org/streams" xmlns:xmlns="jabber:component:accept" xmlns:stream:http://etherx.jabber.org/streams="xmlns:stream" xmlns:jabber:component:accept="xmlns" id="314321979" from="superfeedr.com"/> 
# 
# =========
# PARSED : 
# <handshake/> 
# 
# =========
# PARSED : 
# <iq from="julien@superfeedr.com/babylon_client_6970" to="firehoser.superfeedr.com" type="set" id="1419">
#           <pubsub xmlns:xmlns="http://jabber.org/protocol/pubsub" xmlns:http://jabber.org/protocol/pubsub="xmlns">
#               <subscribe jid="julien@superfeedr.com" node="http://feeds2.feedburner.com/NotifixiousFoundersBlog"/>
#           </pubsub>
#       </iq> 
# 
# =========
# PARSED :
# <iq from='julien@superfeedr.com/babylon_client_6970' to='firehoser.superfeedr.com' type='set' id='7858'>
#     <pubsub xmlns='http://jabber.org/protocol/pubsub'>
#         <unsubscribe jid='julien@superfeedr.com' node='http://feeds2.feedburner.com/NotifixiousFoundersBlog'/>
#     </pubsub>
# </iq>