num2txt: convert a number into a lot of possible 'telephone words'. pass `-w` to remove all the non-words according to /usr/share/dict/words

gistfile1.rb

#!/usr/bin/env ruby

NUMS = [[], [], ['a', 'b', 'c'], ['d', 'e', 'f'],
        ['g', 'h', 'i'], ['j', 'k', 'l'], ['m', 'n', 'o'],
        ['p', 'q', 'r', 's'], ['t', 'u', 'v'], ['w', 'x', 'y', 'z']]

def num2txt str
  ps = []
  str.each_char do |char|
    if char.match(/\d/)
      ps << NUMS[char.to_i]
    else
      ps << [char]
    end
  end
  words = possibilities ps, ps.length
  if $onlyindict
    words & $dict
  else
    words
  end
end

def possibilities ps, length
  words = []
  ps.each_with_index do |letters, i|
    mps = ps[(i+1)..-1]
    if mps.empty?
      words.push(*letters) if (length == 1)
    else
      letters.each do |letter|
        possibilities(mps, length-1).each do |mp|
          word = "#{letter}#{mp}"
          words << word unless (word.length < length)
        end
      end
    end
  end
  words
end

$onlyindict = ((ARGV[0] == '-w') ? (ARGV.shift; true) : false)

if $onlyindict
  $dict = File.read("/usr/share/dict/words").split("\n").map(&:downcase)
end

if ARGV.empty?
  while gets
    puts num2txt $_
  end
else
  ARGV.each {|word| puts num2txt word }
end

Fun game: when you have this and txt2num, you can search for hidden meanings in your favourite memorable 1-800 numbers by doing something like this:

$ num2txt -w `txt2num <word>`

Honestly: personally, I haven't found anything interesting yet. But I'll report back when it turns out that some well-known customer support line number spells ASSHOLE or BOOBIES or any of those funny calculator words. Or anything else, really.

(I wrote these to help search for Twilio numbers.)