From: Tom Felker <tcfelker@...>
Date: 2003-09-17T10:13:14+09:00
Subject: Re: scrambler one-liner

OK, here's the unscrambler code.  Be warned, I chose a rather memory
intensive algorithm, it takes about 110 MB of memory with my 2.4 MB words
file.  I don't know how the Perl I gave works, but it only took about 35
MB.  Anyway, it's very fast after the 5 seconds or so to load the
word list.

It works by reading a word list to make a hash, with the keys being the
word alphabetically instead of randomly scrambled, and the values being
the list of unscrambled words.  I was amazed at how many words are unique.


#!/usr/bin/env ruby

class Unscrambler
	def initialize(wordsFilename = "/usr/share/dict/words")
		@wordsHash = Hash.new{ Array.new }
		File.open(wordsFilename) do |wordsFile|
			wordsFile.each_line do |word|
				word.chomp!
				@wordsHash[Unscrambler::word_to_key(word)] <<= word
			end
		end
	end

	def Unscrambler::word_to_key(word)
		return word unless word.size > 3
		array = word.split(//)
		array[1..-2] = array[1..-2].sort
		return array.join;
	end

	def unscramble(word)
		return @wordsHash[Unscrambler::word_to_key(word)]
	end

end

puts "Loading wordlist..." if $stdin.isatty

u = Unscrambler.new

puts "Ready." if $stdin.isatty

$stdin.each_line do |line|
	line.gsub!(/\w+/) do |match|
		choices = u.unscramble(match.downcase)
		case choices.size
			when 0 then match
			when 1 then choices[0]
			else "[" + choices.join(", ") + "]"
		end
	end
	print line
end

-- 
Tom Felker, <tcfelker@mtco.com>
<http://vlevel.sourceforge.net> - Stop fiddling with the volume knob.

Torvalds, explaining SCO's actions: "They are smoking crack."