#------------------------------------------------------------
# TITLE:   example-REXML-problem.rb
# AUTHOR:  MAS
# DATE:    2012-11-04
# VERSION: 1.0
#------------------------------------------------------------
# DESCRIPTION:
#   ...
#  
#   This program illustrates problem with REXML.  Input
#   file contains HTML entities like ’. These are
#   converted on the fly to their literal Unicode
#   character point, which is NOT the same as their
#   original HTML function. Note that these 
#   transliterations seem to only occur when using ANY 
#   form of Xpath expression.
#
#------------------------------------------------------------
# REQUIREMENTS:
#   Ruby 1.9  (for UTF ability)
#   REXML capability (may be installed already with 1.9?)
#------------------------------------------------------------
# LIABILITY:
#   USE AT YOUR OWN RISK!!
#------------------------------------------------------------
# USAGE:
#   Run program as ruby <program name> Input/output file
#   names can be specified in variables below.
#------------------------------------------------------------
# CHANGE THESE VARIABLES AS NEEDED:
input_source_file  = "sample-REXML-input.xml"
output_source_file = "sample-REXML-output.xml"
this_year = 2012
data_version = '1815' # Match with whatever current CN puts out
#------------------------------------------------------------
#  AVOID CHANGING ANYTHING BELOW THIS LINE UNLESS YOU KNOW
#  WHAT YOU'RE DOING 
#------------------------------------------------------------

require 'date'
require "rexml/document"
require 'cgi'
include REXML  

file = File.new(input_source_file)
doc = Document.new( file )
names = XPath.match( doc, "//node" ) 

string = <<EOF
 <notebook version='#{data_version}'>
   <tags>
     <tag autodeleted='1' autoexpanded='0' name='#DUMMY'/>
   </tags>
    <note created='20130508T131005' link='' modified='20130509T131005' plainText='1' source='' tags='' title='DELETE ME'><![CDATA[]]></note>
 </notebook>
EOF

doc_out = Document.new string
doc_out << XMLDecl.new
root = doc_out.root

#tags = doc_out.elements["notebook/tags"]

notes = doc_out.elements["notebook"]

curDate = DateTime.now
strDate = curDate.to_s
strYear,strMo,strDay = strDate.slice(0,4), strDate.slice(5,2), strDate.slice(8,2)
strStamp = "#{strYear}#{strMo}#{strDay}T080000"


names.each do |element|

  title = element.elements["title"].text.strip
  article = element.elements["article"].text
#puts article  
  el = notes.add_element "note" 
  el.attributes["title"] = title
  el.attributes["tags"] = "IMPORT" 
  el.attributes["created"] = strStamp
  el.attributes["modified"] = strStamp
  el.attributes["source"] = ''
  el.attributes["link"] = ''
  el.attributes["plainText"] = '0'
  el.add_text(CData.new(article))

end

temp = doc_out.text
#File.open(output_source_file, 'wt:UTF-16LE') do |f|
# I output here in UTF-8 so that the user can observe and 
# confirm the stated problem with code conversion. The
# The conversion problem has actually already occurred by
# the time we have reached this point.
File.open(output_source_file, 'wt:UTF-8') do |f|
  doc_out.write f
end