make a diff of two files containing Atom Feed made for JMA XML (not all Atom Feed tags handled)

atomdif.rb

#!/usr/bin/ruby

require 'rexml/parsers/baseparser'
require 'rexml/parsers/streamparser'
require 'rexml/streamlistener'

class AtomParse
  include REXML::StreamListener

  def initialize
    @tag = nil
    @rec = nil
    @callback = proc
  end

  def text(str)
    return unless @tag
    @rec[@tag] = str
    # will not save following text nodes
    @tag = nil
  end

  alias :cdata :text

  TAGS = /^(name|author|id|title|subtitle|updated)$/

  def tag_start(name, attrs)
    case name
    when 'feed' then
      @rec = {:feed => true}
    when 'entry' then
      @callback.call(@rec) if @rec and @rec[:feed]
      @rec = {}
    when TAGS then
      @tag = name
    when /^(content|rights)$/ then
      @tag = "#{name}.#{attrs['type']}"
    when 'link' then
      aname = 'link.' + attrs['rel'].to_s
      @rec[aname] = attrs['href']
    else
      raise Errno::EINVAL, "unsupported tag #{name}"
    end
  end

  def tag_end(name)
    case name
    when 'entry' then 
      @callback.call(@rec)
      @rec = nil
    end
  end

end

class FeedBuild

  class << FeedBuild

    def xmltext str
      str.to_s.encode(Encoding::UTF_8, :xml => :text)
    end

    def xmlattr str
      str.to_s.encode(Encoding::UTF_8, :xml => :attr)
    end

    def open io
      fb = FeedBuild.new(io)
      yield fb
    ensure
      fb.close
    end

  end

  def initialize io
    @io = io
    @io.set_encoding 'UTF-8'
    @state = :init
  end

  def dump hash
    for key in %w(title subtitle updated id)
      next unless hash.include? key
      @io.puts ['<', key, '>', FeedBuild.xmltext(hash[key]), '</', key, '>'].join
    end
    if hash.include? 'name'
      @io.puts ['<author><name>', FeedBuild.xmltext(hash['name']), '</name></author>'].join
    end
    for key in %w(link.related link.self link.hub link.)
      next unless hash.include? key
      @io.write '<link '
      if rel = key.split(/\./)[1] then
        @io.write 'rel='
        @io.write FeedBuild.xmlattr(rel)
        @io.write ' href='
        @io.write FeedBuild.xmlattr(hash[key])
      else
        @io.write 'type="application/xml" href='
        @io.write FeedBuild.xmlattr(hash[key])
      end
      @io.puts ' />'
    end
    for key in %w(rights.html content.text content.html)
      next unless hash.include? key
      tag, type = key.split(/\./, 2)
      @io.write '<'
      @io.write tag
      @io.write ' type='
      @io.write FeedBuild.xmlattr(type)
      @io.write '>'
      @io.write FeedBuild.xmltext(hash[key])
      @io.write '</'
      @io.write tag
      @io.puts '>'
    end
  end

  def top hash
    raise "top() called twice" unless @state == :init
    @state = :top
    @io.puts '<?xml version="1.0" encoding="utf-8">'
    @io.puts '<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="ja">'
    dump(hash)
  end

  def entry hash
    raise "entry() before top()" unless @state == :top
    @io.puts '<entry>'
    dump(hash)
    @io.puts '</entry>'
  end

  def close
    raise "close without top()" if @state == :init
    @io.puts '</feed>'
    @io.close
  end

end

# main logic

db = {}

fnam = ARGV.shift
saver = AtomParse.new {|rec|
  id = rec[:feed] ? :feed : rec[id]
  db[id] = rec
}
File.open(fnam, 'r:UTF-8') {|fp|
  begin
    REXML::Parsers::StreamParser.new(fp.read, saver).parse
  rescue REXML::ParseException => e
    STDERR.puts "feed #{arg} - #{e.message}"
    exit 16
  end
}

fnam = ARGV.shift
FeedBuild.open($stdout){|builder|
  differ = AtomParse.new {|rec|
    if rec[:feed] then
      builder.top(rec)
    else
      builder.entry(rec) unless db[rec['id']]   
    end
  }
  File.open(fnam, 'r:UTF-8') {|fp|
    begin
      REXML::Parsers::StreamParser.new(fp.read, differ).parse
    rescue REXML::ParseException => e
      STDERR.puts "feed #{arg} - #{e.message}"
      exit 16
    end
  }
}