From: MonkeeSage <MonkeeSage@...>
Date: 2007-12-26T18:55:00+09:00
Subject: Re: Sorting Titles in Ruby

On Dec 24, 12:01 pm, Thomas Wieczorek <wieczo...@googlemail.com>
wrote:
> On Dec 24, 2007 6:11 PM, Rick DeNatale <rick.denat...@gmail.com> wrote:
>
> > In other words, leading articles should be moved to the end, and
> > numbers should be spelled out before sorting.
>
> You can use the Linguistics
> gem(http://www.deveiate.org/projects/Linguistics/) to convert numbers
> to words. I never used it, so you need probably a simple regex to
> extract the numbers.
> Leading articles can be easily removed with a regex.

Simple implementation of that...

# Glenn Parker's solution to Ruby Quiz #25 [1]
require "num2eng"

class String
  def title_case
    # from somewhere...can't recall
    gsub(/\b\w/) { $&.upcase }
  end
  def canonical_form
    nums = scan(/[\d,_]+/)
    unless nums.empty?
      nums.each { | num |
        sub = num.gsub(",","").to_i.to_english.title_case
        gsub!(num, sub)
      }
    end
    gsub(/\A(An?|The)(.*)/, '\2, \1').strip
  end
end

titles = [
  "20,000 Leagues Under the Sea",
  "A Fine Madness",
  "Fanny",
  "Marooned",
  "The Man Who Came to Dinner",
  "5 Days and 6 Nights of Fictitious Cinema"
]

puts titles.map { | title |
  title.canonical_form
}.sort

# =>
Fanny
Fine Madness, A
Five Days and Six Nights of Fictitious Cinema
Man Who Came to Dinner, The
Marooned
Twenty Thousand Leagues Under the Sea

Ps. Corner cases like "20, 000 Leagues ..." will break, but it's a
start. :)

[1] http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-talk/135449

Regards,
Jordan