Find tweets that are anagrams of each other.

tweet-anagrams.rb

require 'set'
require 'pp'
require 'date'
require 'pry'

class Index

  include Enumerable

  def initialize
    @tweets = Hash.new { |h,k| h[k] = Set.new }
  end

  def add(tweet)
    @tweets[tweet.anagram_hash].add tweet
  end

  def each(&block)
    @tweets.each(&block)
  end

  def size
    @tweets.size
  end

end

class Tweet < Struct.new(:date, :user, :msg)

  include Comparable

  def <=>(other)
    self.msg <=> other.msg
  end

  def stripped
    @stripped ||= msg.downcase.gsub(/[^\w]/, '')
  end

  def anagram_hash
    @anagram_hash ||= stripped.chars.sort.join('')
  end

  def hash
    @hash ||= stripped.hash
  end

  def eql?(other)
    stripped.eql? other.stripped
  end

  def initialize(line)
    if line =~ /\[(\d\d\/\d\d\/\d\d \d\d:\d\d:\d\d)\] <([^>]+)> ?(.*)/
      self.user = $2
      self.msg = $3
      self.date = DateTime.strptime($1, "%m/%d/%y %H:%M:%S")
    else
      raise "bad line: #{line}"
    end
  end

end

if $0 == __FILE__
  index = Index.new

  open("tweets.txt").each do |line|
    begin
      tweet = Tweet.new(line)
      index.add tweet
    rescue => e
    #   p e
    end
  end

  p index.size
  matches = index.select{|k,v| v.size > 1 }

  binding.pry

end