From: Rick DeNatale <rick.denatale@...>
Date: 2008-01-06T12:06:41+09:00
Subject: Re: Chain, Chain, Chain (was Re: zed shaw zed shaw zed shaw)

On Jan 5, 2008 5:05 PM, Joel VanderWerf <vjoel@path.berkeley.edu> wrote:
> Rick DeNatale wrote:
> > Okay,  I've been playing with doing this programatically.  I'm not yet
> > convinced that it's bug free, but I've found some rather long movie
> > title chains using the non-imdb list posted earlier:
> >
> > 10 ITEMS OR<LESS> THAN<ZERO><DAY> FOR<NIGHT> AND<DAY> OF
> <snip>
> > WILL<TALK><RADIO><DAYS> OF THUNDER
>
> Why didn't it find Thunder Road[1]? (And then maybe Road Trip to
> Bountiful...)

Well first of all, because thunder road isn't on that list.

Second, because it's the longest title the program found in the time I
gave to run it.  As I said the program isn't of the quality that I'd
prefer to share, but what the heck.  It's a pretty dumb search, and I
have it print a chain only the first time it finds one longer (in
terms of the number of movies) than any it had found before.  It's
full of "the simplest thing that could possibly work" decisions with
the idea of making it work before trying to make it fast.

require 'net/http'
class Title

  @first_words = Hash.new {|h,k| h[k] = []}

  def self.register(title)
    @first_words[title.first_word] << title if title.first_word
  end

  def self.process_lines(string)
    string.each_line do |line|
      self.new(line)
    end
  end

  def self.generate_titles
    result = []
    @first_words.keys.sort.each do |fw|
      all_starting_with(fw).each do |title|
        title.chains do | chain |
          result << chain
        end
      end
    end
  end

  def self.all_starting_with(word)
    @first_words[word]
  end

  def print_chain(chain)
    puts "#{chain.length}: #{chain.inject(chain.first.first_word) {
|chained_title, title| title.chain_to(chained_title)}}"
  end

  def chain_to(string)
    "#{string.sub(/\s(\S+)$/,'<\1>')} #{@words[1...@words.length].join(' ')}"
   end

  def self.max_chain?(chain)
    @max ||= 0
    if chain.length > @max
      @max = chain.length
      true
    else
      false
    end
  end

  def chains(chain = [])
    root_chain = chain << self
    raise "Duplicate title #{print_chain(root_chain)}" unless
root_chain.length == root_chain.uniq.length
    print_chain(root_chain) if self.class.max_chain?(root_chain)
    result = []
    self.class.all_starting_with(self.last_word).each do |title|
      unless root_chain.include?(title)
        title.chains(root_chain.dup).each do |chain|
          result << chain
        end
      end
    end
    result
  end

  def initialize(title_string)
    @words = title_string.strip.split
    self.class.register(self)
  end

  def first_word
    @words.first
  end

  def last_word
    @words.last
  end

  def to_s
    @words.join(' ')
  end

  def inspect
    to_s
  end
end
_,x = Net::HTTP.new("itafullsite.dev.neptuneweb.com").get("/careers/puzzles/MOVIES.LST")
Title.process_lines(x)
p Title.generate_titles


-- 
Rick DeNatale

My blog on Ruby
http://talklikeaduck.denhaven2.com/