Find and clean duplicate files (Ruby)

cleandups.rb

#!/opt/local/bin/ruby1.9
require 'digest/md5'
require 'shellwords'
# argument processing -- goes through Dir[], so can use '**/*' etc.
ARGV[0] ||= '.'
filenames = ARGV.map do |dirn|
  Dir[if File.directory?(dirn)
        "#{dirn}/*"
      else
        dirn
      end]
end.flatten(1)
# find sets of same-content files
equivsets = filenames.map do |fn|
  [fn, File.lstat(fn)]
end.reject do |fn, stat|
  stat.symlink? || stat.directory? || stat.size == 0
end.group_by { |fn, stat| stat.size }.values.reject do |list|
  list.size == 1
end.map do |sublist|
  sublist.map do |fn, stat|
    [fn, stat, Digest::MD5.hexdigest(File.read(fn))]
  end.group_by { |fn, stat, digest| digest }.values.reject do |list|
    list.size == 1
  end
end.flatten(1).map do |list|
  list.map(&:first).sort_by(&:size)
end
# now remove the duplicate files (alternative: create links!)
commands = equivsets.map do |list|
  keep = list.shift             # the file we keep
  list.map do |candidate|
    %Q{
if cmp #{[keep, candidate].shelljoin}
then rm #{[candidate].shelljoin}
fi
}
  end.join
end.join
puts commands