Hangman (#130)

The three rules of Ruby Q.:

  1. Please do not post any solutions or spoiler discussion for this quiz
    until
    48 hours have passed from the time on this message.

  2. Support Ruby Q. by submitting ideas as often as you can:

http://www.rubyquiz.com/

  1. Enjoy!

Suggestion: A [QUIZ] in the subject of emails about the problem helps
everyone
on Ruby T. follow the discussion. Please reply to the original quiz
message,
if you can.

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

by Brian C.

Most people are probably familiar with the game of Hangman. The first
player
picks a word or phrase, and the second player has to guess it a letter
at a
time. If they make six wrong guesses (i.e. the target word does not
contain the
guessed letter), they lose. If they guess the entire word before then,
they win.

This quiz is to make a Hangman guessing player in Ruby. Play should
proceed as
follows:

  1. The program requests a word or phrase pattern, e.g. “-------”.

  2. The program suggests a letter, or may guess the entire word or
    phrase.

  3. The user indicates which letter positions, if any, match that
    letter.
    If none match, a life is lost. If six (or configurable) lives are
    lost,
    the program loses.

The specification is otherwise open-ended to allow you to focus on
whatever part
of the problem interests you. For example:

  • You may choose the form of user interface (e.g. terminal, GUI
    toolkit,
    web).

  • You can just show the number of wrong guesses made, or you can
    actually
    draw the hangman.

  • You may concentrate on improving the play, for example by using a
    dictionary to improve the guesses made at each stage. A suitable
    file
    is /usr/share/dict/words on many Linux systems.

  • A dynamic solution could start with an empty dictionary, and guess
    the
    answer by chance. If it fails, it would prompt the user for the word
    or phrase they were thinking of. It would add new words or phrases
    to
    its dictionary so as to become a better player over time.

  • You could investigate ways of precomputing a hangman decision tree,
    optimizing it for the minimum number of wrong guesses along each
    branch.
    The aim is to produce an unbeatable guesser for a given dictionary.

  • You may wish to consider how best to decouple the UI from the
    guessing
    logic, to enable different UI’s to work with the same guessing
    engine,
    or vice versa.

Here’s my solution. I tried to tackle many of the suggested ideas:

  • Extensible interface and AI. Just create a new file for the
    interface/ai,
    give it the right filename, and implement the appropriate methods.
    Then
    call hangman.rb with the correct --interface or --ai option.

  • I’ve implemented a simple text interface and one based on Ncurses,
    just to
    try it out. My ncurses code is ugly, but it works ok. I almost tried
    out
    some animation & color a la the rain.rb ncurses example, but have no
    time.

  • Implemented a random AI and one that tries to match items in a
    dictionary,
    though it does not add new words to it. It greps through the
    dictionary
    file on each guess, but is still pretty quick with the 52848 word one
    I’m
    testing with.

Downsides:

  • Not much error checking. Inputing illegal positions and such is an
    easy
    crash.

  • Very little documentation.

hangman.rb is the executable, which creates an interface (subclass of
Interface::Core), and an AI (subclass of AI::Core), and passes them to a
Game object which controls the basic game flow. (Heh, 9 source files for
a quiz submission, a personal record… :slight_smile:

Usage: hangman.rb [–interface | -i INTERFACE]
[–interface-arg | -j INTERFACE_ARGUMENT]
[–ai | -a AI]
[–ai-arg | -b AI_ARGUMENT]

Simple UI / Dictionary AI full example:

$ cat dict.txt
CAR
CAT
DOCK
DOOR
RUBY
ACORN
HINGE
ZEBRA
$ ./hangman.rb -a dictionary
Enter a phrase pattern: —
— | Computer lives: 6
I guess A. What position(s) is it in? 2
-A- | Computer lives: 6
I guess C. What position(s) is it in? 1
CA- | Computer lives: 6
I guess R. What position(s) is it in?
CA- | Computer lives: 5
I guess T. What position(s) is it in? 3
CAT | Computer lives: 5

Woot! I win!

Ncurses UI / Random AI example end screen:

┌──────────────────────────────────────────────────────────────────────────────┐
│ Hangman |
│
│---------+
│
│
│
│ Phrase: V–
│
│
│
│ Computer guess: M
│
│ Positions?
│
│
│
│
│
│ . I lost!
│
│ 0 ±-------±
│
│ | |
│
│ _ |
│
│ | | |
│
│ + |
│
│ -|- |
│
│ / | \ |
│
│ ^ |
│
│ / \ |
│
│ / \ |
│
│ |
│
│ ====================
│
│
│
└──────────────────────────────────────────────────────────────────────────────┘

Not many people solved this one, so I wanted to see if it was too
tough. Didn’t seem so:

#!/usr/bin/env ruby -wKU

puts “One moment…”
puts
words = File.open(ARGV.shift || “/usr/share/dict/words”) do |dict|
dict.inject(Hash.new) do |all, word|
all.update(word.delete("^A-Za-z").downcase => true)
end.keys
end

guesses = Array.new

loop do
puts guesses.empty? ?
“Please enter a word pattern (_ _ _ _ for example):” :
“Please update your pattern according to my guess (_ i _ _
for example):”
pattern = $stdin.gets.to_s.delete("^A-Za-z_")
if pattern.include? “"
if (guesses - pattern.delete("
”).split("")).size > 6
puts “I’m out of guesses. You win.”
puts
exit
end
else
puts “I guessed your word. Pretty smart, huh?”
puts
exit
end

choices = words.grep(/\A#{pattern.tr("", “.”)}\Z/i)
odds = Hash.new(0)
choices.each do |word|
word.split("").each do |l|
next if guesses.include? l
odds[l] += word.count(l)
end
end
guess = odds.max { |(
, c1), (_, c2)| c1 <=> c2 }.first rescue nil

guesses << guess
puts “I guess the letter ‘#{guess}’.”
puts
end

END

James Edward G. II

Ruby Q. wrote:

This quiz is to make a Hangman guessing player in Ruby. Play should proceed as
follows:

I focused on building a program that makes good guesses.

== Algorithm overview

The guesser reads a dictionary and then builds a database (which is
reused) with the following information about each word:

  • size
  • positions of each character
  • number of occurrences of each character

The basic algorithm is as follows:

  • Remove all words that do not have a matching length.
  • While the game has not been solved:
    ** Pick the character included in the most words still remaining.
    ** If the character is not in the word: remove all words with the
    character.
    ** If the character is in the word: remove all words that do not contain
    the character at exactly the revealed positions.

=== Weaknesses

The algorithm is not optimal. The character that’s included in the most
words is not necessarily the character which will give the largest
reduction in number of potential words since position is not considered.
Consider a dictionary with the following:

bdc
ebc
fcb

b and c are tied for number of occurrences, but b would be the better
choice. If we pick b we will in all cases be left with one potential
word. If we pick c and the word is one of the first two we get two
potential words.

My guess is that it’s a good enough heuristic in most cases though.

Also note that this program is built on the assumption that the word is
picked randomly from the dictionary. More refined solutions could weigh
in the relative frequency of different words in normal English text.

== Speed

It takes about 40 minutes to create the database for a dictionary with
4*10^5 words, but it only has to be created once.

Computing all guesses for a word (i.e. from being given the length to
having the correct word) takes about 30 to 40 seconds for a dictionary
with 4*10^5 words. That time includes about 10 seconds to reset the
database from previous uses, another 10 seconds for pruning based on
word length and the rest for the remaining search.

=== Possible improvements

Much of the initial sorting could be precomputed (e.g. split words into
different table based on length and then only work against the table
with the specified length) to cut down on the time needed reset and do
the initial pruning. The first (and possibly some additional steps)
could also be precomputed.

== Dependencies

Requires a mysql database and the mysql-gem. You need to enter your
username, passwords and database name in HangmanGuesser#db_connection
below.

== The code

#!/usr/bin/env ruby

== Synopsis

automated_hangman: plays a game of hangman with the word of your

choice

== Usage

automated_hangman [OPTION] … WORD

-h, --help:

show help

-d, --dictionary [dictionary location]:

sets up the database to use the specified dictionary (defaults to

/usr/share/dict/words), can take some time

WORD: The word that the program should try to guess.

require ‘getoptlong’
require ‘rdoc/usage’
require ‘mysql’

Describes a game of hangman.

class Hangman
LIVES = 6

Creates a new game of hangman where word is the target word.

def initialize(word)
@guesses = []
@word_characters = word.chomp.downcase.split(//)
end

Returns an array containing the incorrect guessed characters.

def incorrect_guesses
@guesses - @word_characters
end

Guesses a specified character. Returns an array of indices (possibly

empty) where the character was found.

def guess(char_guess)
@guesses << char_guess
indices = []
@word_characters.each_with_index do |character, index|
indices << index if character == char_guess
end
return indices
end

Returns a string representation of the current progress.

def to_s
hidden_characters = @word_characters - @guesses
return @word_characters.join(’ ‘) if hidden_characters.empty?
@word_characters.join(’ ').gsub(
/[#{hidden_characters.uniq.join}]/, ‘_’)
end

Checks whether the player has won.

def won?
(@word_characters - @guesses).empty?
end

Checks whether the player has lost.

def lost?
incorrect_guesses.size > LIVES
end

Gets the number of characters in the word.

def character_count
@word_characters.size
end
end

The guessing machine which picks the guesses.

class HangmanGuesser

The location of the default dictionary to use.

DICTIONARY_FILE = ‘/usr/share/dict/words’

An array of the characters that should be considered.

CHARACTERS = (‘a’…‘z’).to_a

Set this to true to see how the search progresses.

VERBOSE = true

The maximum word length accepted.

MAX_WORD_LENGTH = 50

The dictionary given should be the location of a file containing one

word per line. The characters should be an array of all characters

that should be considered (i.e. no words with other characters are

included).

def initialize(hangman_game, characters = CHARACTERS)
@con = self.class.db_connection
@characters = characters
@hangman_game = hangman_game

reset_tables
prune_by_word_length @hangman_game.character_count

end

Returns the guesses that the guesser would make.

def guesses
@guesses = []
log{ “There are #{word_count} potential words left.” }
while not @hangman_game.won?
guess = next_guess
raise ‘The word is not in the dictionary.’ if guess.nil?
@guesses << guess
log{ “Guessing #{guess}” }
add_information(guess, @hangman_game.guess(guess))
log_state
log{ “\n” }
end
return @guesses
end

class << self
# Creates the database and populates it with the dictionary file
# located at the specified location. Only considers the specified
# characters (array).
def create_database(dictionary = DICTIONARY_FILE,
characters = CHARACTERS)
@con = db_connection
@characters = characters
@tables = [‘words’] + @characters +
@characters.map{ |c| c + ‘_occurrences’}
create_tables
populate_tables File.open(dictionary)
end

# Connects to the database that should store the tables.
def db_connection
  # Replace <username> and <password> with the database username and
  # password.
  Mysql.real_connect("localhost", <username>, <password>, "hangman")
end

private

# Creates the tables used to store words.
def create_tables
  # Drop old tables.
  @tables.each do |table|
    @con.query "DROP TABLE IF EXISTS `#{table}`"
  end

  # Words table.
  @con.query <<-"end_sql"
    CREATE TABLE `words` (
      `word_id` mediumint(8) unsigned NOT NULL AUTO_INCREMENT,
      `word` varchar(#{MAX_WORD_LENGTH}) NOT NULL,
      `length` tinyint(3) unsigned NOT NULL,
      `removed` tinyint(1) unsigned NOT NULL DEFAULT '0',
      PRIMARY KEY (`word_id`),
      INDEX (`removed`),
      INDEX (`length`)
    ) ENGINE=MyISAM
  end_sql

  # Tables for the number of occurrences of each character.
  character_occurrences_table_template =<<-'end_template'
    CREATE TABLE `%s_occurrences` (
      `word_id` mediumint(8) unsigned NOT NULL,
      `occurrences` tinyint(3) unsigned NOT NULL,
      PRIMARY KEY (`occurrences`, `word_id`),
      INDEX (`word_id`)
    ) ENGINE=MyISAM
  end_template

  # Tables for the positions of each character.
  character_table_template =<<-'end_template'
    CREATE TABLE `%s` (
      `word_id` mediumint(8) unsigned NOT NULL,
      `position` tinyint(3) unsigned NOT NULL,
      PRIMARY KEY (`position`, `word_id`),
      INDEX (`word_id`)
    ) ENGINE=MyISAM
  end_template

  @characters.each do |character|
    @con.query character_occurrences_table_template % character
    @con.query character_table_template % character
  end
end

# Loads a dictionary into the database.
def populate_tables(dictionary_file)
  # Disable the keys so that we don't update the indices while
  # adding.
  @tables.each do |table|
    @con.query("ALTER TABLE #{table} DISABLE KEYS")
  end

  # Prepare statements.
  add_word = @con.prepare(
    "INSERT INTO words (word, length) VALUES (?, ?)")
  add_character = {}
  add_character_occurrences = {}
  @characters.each do |character|
    add_character[character] = @con.prepare(
      "INSERT INTO #{character} (word_id, position) VALUES (?, ?)")
    add_character_occurrences[character] = @con.prepare(
      "INSERT INTO #{character}_occurrences " +
      "(word_id, occurrences) VALUES (?, ?)")
  end

  # Populate the database.
  previous_word = nil
  dictionary_file.each_line do |line|
    # Only consider words that only contain characters a-z. Make
    # sure we don't get duplicates.
    word = line.chomp.downcase
    next if word == previous_word or word =~ /[^a-z]/ or
      word.size > MAX_WORD_LENGTH

    # Add the word, its character positions and number of
    # occurrences.
    add_word.execute(word, word.size)
    word_id = @con.insert_id
    characters = word.split(//)
    characters.each_with_index do |character, position|
      add_character[character].execute(word_id, position)
    end
    @characters.each do |character|
      occurrences = characters.select{ |c| c == character }.size
      add_character_occurrences[character].execute(
        word_id, occurrences)
    end

    previous_word = word
  end

  # Generate the indices.
  @tables.each do |table|
    @con.query("ALTER TABLE #{table} ENABLE KEYS")
  end
end

end

private

Logs the current state of the guessing process.

def log_state
log do
messages = []
messages << @hangman_game.to_s
count = word_count
messages << “There are #{count} potential words left.”
if count <= 10
res = @con.query(‘SELECT word FROM words WHERE removed = 0’)
res.each{ |row| messages << row[0] }
res.free
end
messages.join("\n")
end
end

Logs the string produced by the block (may not be executed at all).

def log(&block)
puts yield() if VERBOSE
end

Gets the number of potential words left.

def word_count
res = @con.query(‘SELECT COUNT(*) FROM words WHERE removed = 0’)
count = res.fetch_row[0].to_i
res.free
return count
end

Computes the next character that should be guessed. The next guess

is the character (that has not yet been tried) that occurrs in the

most words remaining.

def next_guess
next_character = nil
max_count = 0
(@characters - @guesses).each do |character|
res = @con.query(
"SELECT COUNT(DISTINCT word_id) FROM #{character} " +
“NATURAL JOIN words WHERE removed = 0”)
count = res.fetch_row[0].to_i
res.free
if count > max_count
next_character = character
max_count = count
end
end
return next_character
end

Adds the information about at what indices in the word the specified

character can be found to the guesser.

def add_information(character, indices)
if indices.empty?
# The character isn’t in the word.
sql =<<-“end_sql”
UPDATE words SET removed = 1 WHERE removed = 0 AND word_id IN (
SELECT word_id FROM #{character}
)
end_sql
else
# Remove all words where the character isn’t at the specified
# places.
sql =<<-“end_sql”
UPDATE words NATURAL JOIN #{character}_occurrences
SET removed = 1
WHERE removed = 0
AND (occurrences != #{indices.size}
OR word_id IN (
SELECT word_id FROM #{character}
WHERE position NOT IN (#{indices.join(’, ')})
)
)
end_sql
end
@con.query(sql)
end

Resets the table to start a new round of guesses.

def reset_tables
@con.query(‘UPDATE words SET removed = 0’)
end

Prunes all words that do not have the specified length.

def prune_by_word_length(expected_length)
@con.query(
“UPDATE words SET removed = 1 WHERE length != #{expected_length}”)
end
end

opts = GetoptLong.new(
[ ‘–help’, ‘-h’, GetoptLong::NO_ARGUMENT],
[’–dictionary’, ‘-d’, GetoptLong::OPTIONAL_ARGUMENT])
opts.each do |opt, arg|
case opt
when ‘–help’
RDoc::usage
when ‘–dictionary’
if arg != ‘’
HangmanGuesser.create_database(arg)
else
HangmanGuesser.create_database
end
end
end

if ARGV.size != 1
abort “Incorrect usage, see --help”
end

game = Hangman.new(ARGV[0])
guesses = HangmanGuesser.new(game).guesses
if game.won?
puts ‘Successfully guessed the word.’
else game.lost?
puts ‘Failed guessing the word.’
end
puts “Made the following guesses: #{guesses.join(’, ')}”
puts “Expended a total of #{game.incorrect_guesses.size} lives.”

A second version of my code. Minor changes mainly to support me
testing it. I tried different guessing algorithms, but nothing beat
the original strategy with my dictionary.

First, this is a tiny words.rb lib used by both the guesser and test
scripts:

#!/usr/bin/env ruby -wKU

WORDS_CASH_FILE = “words.cache”

if File.exist? WORDS_CASH_FILE
WORDS = File.open(WORDS_CASH_FILE) { |file| Marshal.load(file) }
else
WORDS = File.open( ARGV.find { |arg| arg =~ /\A[^-]/ } ||
“/usr/share/dict/words” ) do |dict|
dict.inject(Hash.new) do |all, word|
all.update(word.delete("^A-Za-z").downcase => true)
end.keys
end
File.open(WORDS_CASH_FILE, “w”) { |file| Marshal.dump(WORDS, file) }
end

END

Next, my guesser script which works the same as yesterday’s version
but is optimized here and there for testing:

#!/usr/bin/env ruby -wKU

puts “One moment…”
puts
require “words”

choices = WORDS
guesses = Array.new

loop do
puts guesses.empty? ?
“Please enter a word pattern (_ _ _ _ for example):” :
“Please update your pattern according to my guess (_ i _ _
for example):”
$stdout.flush
pattern = $stdin.gets.to_s.delete("^A-Za-z_")

if (guesses - pattern.delete("").split("")).size > 5 and
pattern.include? "
"
puts “I’m out of guesses. You win.”
elsif not pattern.include? “"
puts “I guessed your word. Pretty smart, huh?”
else
choices = choices.grep(/\A#{pattern.tr("
”, “.”)}\Z/i)
odds = Hash.new(0)
choices.each do |word|
word.split("").each do |l|
next if guesses.include? l
odds[l] += word.count(l)
end
end
guess = odds.max { |(, c1), (, c2)| c1 <=> c2 }.first rescue nil

 guesses << guess
 puts "I guess the letter '#{guess}'."
 puts
 next

end

puts
if ARGV.include? “–loop”
choices = WORDS
guesses = Array.new
else
break
end
end

END

Finally, this is my test script:

#!/usr/bin/env ruby -wKU

require “words”

results = Hash.new(0)
at_exit do
results[:total] = results[:right] + results[:wrong]
puts
puts “Words: #{results[:total]}”
puts “Guessed: #{results[:right]}”
puts “Missed: #{results[:wrong]}”
printf “Accuracy: %.2f%%\n”, results[:right] / results
[:total].to_f * 100
puts
end
trap(“INT”) { exit }

IO.popen( File.join(File.dirname(FILE), “hangman.rb --loop”),
“r+” ) do |hangman|
WORDS.each do |word|
pattern = word.tr(“a-z”, “_”)
loop do
input = String.new
hangman.each do |line|
input << line
break if input =~ /^(?:I’m out|I guessed)|:\Z/
end

   if input =~ /^I'm out/
     puts "It missed '#{word}'."
     results[:wrong] += 1
     break
   elsif input =~ /^I guessed/
     puts "It guessed '#{word}'."
     results[:right] += 1
     break
   elsif input =~ /^I guess the letter '(.)'/
     guess = $1
     word.split("").each_with_index do |letter, i|
       pattern[i, 1] = letter if letter == guess
     end
   end

   hangman.puts pattern
 end

end
end

END

James Edward G. II

Maybe people aren’t trying this quiz because they’re busy. I hope
it’s not because they find it uninteresting. I banged out something
fairly quickly during the weekend and was waiting to find some time to
clean it up. It seems I won’t get that, so I’m going to post what I
have now.

It’s a fairly simple algorithm that uses a combination of the
“established” frequency order of letters in the English language, as
seen on Linotype machines, and a dictionary (/usr/dict/share/words, of
course) to provide the next guess. What I find interesting about this
is how many common words it cannot get before exhausting its allotment
of wrong guesses. For instance, it won’t get “book”. Sometimes I’m
surprised with the guesses it gives. Maybe the dictionary I’m using
is too large.


#!/usr/bin/env ruby

ALLOWED_WRONG_GUESSES = 6

WORD_LIST_FILENAME = ‘/usr/share/dict/words’

WORD_LIST = {}

File.open(WORD_LIST_FILENAME) do |f|
f.each { |word| WORD_LIST[word.strip.downcase] = true }
end

FREQUENCY_ORDER = %w{etaoin shrdlu cmfwyp vbgkqj xz}.collect { |elem|
elem.split(’’) }.flatten

GUESSES_MADE = {}

puts ‘Enter a word pattern’
old_pattern = pattern = gets.chomp

loop do
if pattern.match(/^[a-zA-Z-]+$/)
if pattern.match(/-/)
if GUESSES_MADE.values.select { |val| val == false }.length >
ALLOWED_WRONG_GUESSES
puts ‘crap I lost’
exit
end

  regex = Regexp.new("^#{pattern.downcase.gsub(/-/, '.')}$")
  possible_words = WORD_LIST.keys.select { |word|

word.match(regex) }
possible_letters = possible_words.collect { |word|
word.split(’’) }.flatten.uniq
guess = ((FREQUENCY_ORDER - GUESSES_MADE.keys) &
possible_letters).first
puts guess.upcase
pattern = gets.chomp

  GUESSES_MADE[guess] = (pattern != old_pattern)
  puts "wrong guesses made: #{GUESSES_MADE.values.select { |val|

val == false }.length}"
old_pattern = pattern
else
puts ‘Yay I won’
exit
end
else
puts ‘This pattern makes no sense.’
exit
end
end

-yossef

A few words to my solution:

  • It saves the dictionary in a yaml file. If the AI doesn’t know a
    word, it will add it to the dictionary
  • it guesses letters which have the highest possibility

It is a bit rough. I am happy about every suggestion to improve it.

2007/7/11, Thomas W. [email protected]:

2007/7/11, James Edward G. II [email protected]:

Not many people solved this one, so I wanted to see if it was too
tough. Didn’t seem so:

I was busy until today.

Here’s my solution and my first quiz solution since I learn Ruby. It
wasn’t that hard. I learned a lot about Ruby with it.

ruby hangman.rb [-w=word] [-l=lifes]
If you don’t pass a word it will choose a random word from the
dictionary. Lifes is set to 6 if you don’t provide it.

#Ruby Q. 130 #solution by Thomas W. #11.07.2007 require 'yaml' $debug = true

class Player

public

def initialize(word)
@word = word
@dictionary = YAML.load_file(DICTIONARYFILE)
@letters = (‘a’…‘z’).to_a
@guessed = []
scan_dictionary(word)
end

def guess()
return @dictionary[0] if @dictionary.length == 1
while (true)
letter = @probabilities.pop
next if @guessed.include?(letter[0])
@guessed << letter[0]
break
end
return letter[0]
end

def word=(value)

if not value.include?(".") then
  #lost
  #unknown word
  if not @dictionary.include?(value) then
    @dictionary = load_dictionary()
    @dictionary << value
    File.open("dictionary.yaml", "w") { |f| YAML.dump(@dictionary, 

f) }
end
else
if @word.eql?(value) then
@word = value
scan_dictionary(value)
end
end
end

private

DICTIONARYFILE = “dictionary.yaml”

def scan_dictionary(masked)
@dictionary = @dictionary or load_dictionary()
@dictionary = @dictionary.grep(Regexp.new(“^#{@word}$”))
set_probability()
end

def set_probability
alphabet = (‘a’…‘z’).to_a
@probabilities = {}
alphabet.each { |l| @probabilities[l] = 0 }
@dictionary.each do |word|
word.each_byte do |letter|
#p letter
l = letter.chr
@probabilities[l] += 1 if alphabet.include?(l)
end
end

@probabilities = @probabilities.sort {|a,b| a[1]<=>b[1]}

end

def load_dictionary()
return YAML.load_file(DICTIONARYFILE)
end
end #of Player

def random_word
words = YAML.load_file(“dictionary.yaml”)
return words[rand(words.length)]
end

def check_for_letters(word, guess, masked_word)
if word.include?(guess) then
#puts “#{guess} is in #{word}”
word.length.times do |i|
if word[i].chr == guess then
masked_word[i] = guess
end
end
end

return masked_word
end

def play_game(word = “”, lifes = 6, give_output = false)
#user given word
word = random_word if word == “”
masked_word = word.gsub(/\w/, “.”)
guess = “”

player = Player.new(masked_word)

while(lifes > 0)
#AI guesses a letter or word
puts “AI is looking for >#{masked_word}<” if give_output
guess = player.guess()
new_word = “”
won = false
puts “AI guessed ‘#{guess}’” if give_output
if guess.length == 1 then
masked_word = check_for_letters(word, guess, masked_word)

else
  if guess.length > 1 then
    break if guess == word
    lifes -= 1
    next
  else
    #nil
  end
end

#wrong guess
if not masked_word.include?(guess) then
  lifes -= 1
  puts "AI lost a life. #{lifes} lifes left."
else
  #found word
  if masked_word == word then
    break
  else
    #found a letter
    player.word = masked_word
    next
  end
end

end

if lifes > 0 then
won = true
else
#give word to player to extend dictionary
player.word = word
won = false
end

return won, word, lifes
end #of play_game

won = false
word = “”
lifes = 6
if ARGV.length > 0
ARGV.each do |arg|
option = arg.split(“=”)
case option[0]
when “-w”
word = option[1]
when “-l”
lifes = option[1].to_i
end
end
end

won, word, lifes = play_game(word, lifes, true)

if won then
puts “AI won! It guessed "#{word}" with #{lifes} lifes left.”
else
puts “Awww! Lost! AI couldn’t guess "#{word}".”
end

On Jul 11, 2007, at 1:47 PM, James Edward G. II wrote:

A second version of my code.

One last tiny tweak.

WORDS = File.open( ARGV.find { |arg| arg =~ /\A[^-]/ } ||
“/usr/share/dict/words” ) do |dict|
dict.inject(Hash.new) do |all, word|
all.update(word.delete("^A-Za-z").downcase => true)
end.keys

Changing that line to:

 end.keys.sort_by { |w| [w.length, w] }

helps me see where I’m at in the tests.

end
File.open(WORDS_CASH_FILE, “w”) { |file| Marshal.dump(WORDS, file) }
end

END

James Edward G. II

Last tweak, I promise!

Here’s a new version of the guesser script that fairs a bit better.
I found the improvement by trying some changes and running my testing
script:

#!/usr/bin/env ruby -wKU

puts “One moment…”
puts
require “words”

def frequency(words)
freq = Hash.new(0)
words.each do |word|
word.split("").each { |letter| freq[letter] += word.count(letter) }
end
freq
end
FREQ = frequency(WORDS).sort_by { |_, count| -count }.map { |letter,
_| letter }

choices = WORDS
guesses = Array.new

loop do
puts guesses.empty? ?
“Please enter a word pattern (_ _ _ _ for example):” :
“Please update your pattern according to my guess (_ i _ _
for example):”
$stdout.flush
pattern = $stdin.gets.to_s.delete("^A-Za-z_")

if (guesses - pattern.delete("").split("")).size > 5 and
pattern.include? "
"
puts “I’m out of guesses. You win.”
elsif not pattern.include? “"
puts “I guessed your word. Pretty smart, huh?”
else
choices = choices.grep(/\A#{pattern.tr("
”, “.”)}\Z/i)
guess = frequency(choices).
reject { |letter, _| guesses.include? letter }.
sort_by { |letter, count| [-count, FREQ.index(letter)] }.
first.first rescue nil

 guesses << guess
 puts "I guess the letter '#{guess}'."
 puts
 next

end

puts
if ARGV.include? “–loop”
choices = WORDS
guesses = Array.new
else
break
end
end

END

James Edward G. II

On Jul 11, 2007, at 3:42 PM, James Edward G. II wrote:

Last tweak, I promise!

This time I really mean last. :wink:

I found another change that improves the algorithm:

#!/usr/bin/env ruby -wKU

puts “One moment…”
puts
require “words”

def frequency(words)
freq = Hash.new(0)
words.each do |word|
word.split("").each { |letter| freq[letter] += word.count(letter) }
end
freq
end
FREQ = frequency(WORDS).sort_by { |_, count| -count }.map { |letter,
_| letter }

choices = WORDS
guesses = Array.new

loop do
puts guesses.empty? ?
“Please enter a word pattern (_ _ _ _ for example):” :
“Please update your pattern according to my guess (_ i _ _
for example):”
$stdout.flush
pattern = $stdin.gets.to_s.delete("^A-Za-z_")

bad_guesses = guesses - pattern.delete("").split("")
if bad_guesses.size > 5 and pattern.include? "
"
puts “I’m out of guesses. You win.”
elsif not pattern.include? “"
puts “I guessed your word. Pretty smart, huh?”
else
choices = choices.grep(
bad_guesses.empty? ?
/\A#{pattern.tr("
”, “.”)}\Z/i :
/\A(?!.*[#{bad_guesses.join}])#{pattern.tr("_", “.”)}
\Z/i
)
guess = frequency(choices).
reject { |letter, _| guesses.include? letter }.
sort_by { |letter, count| [-count, FREQ.index(letter)] }.
first.first rescue nil

 guesses << guess
 puts "I guess the letter '#{guess}'."
 puts
 next

end

puts
if ARGV.include? “–loop”
choices = WORDS
guesses = Array.new
else
break
end
end

END

James Edward G. II

Sorry, it is late

Please replace

word = “Salawzander”
with
word = “”

Sorry it’s a little late; I’m on vacation and have spent limited time on
a computer with Ruby installed.

My solution reads a dictionary stored in “words.txt.” It uses this to
construct an array of arrays containing the possible words for each word
in the pattern, which is constantly updated according to feedback.

I thought about making something based on making sure each word has a
vowel and hardcoding in the general English language letter frequencies,
but I realized that that would be compeltely unnecessary, as I could
constantly calculate the true letter frequency for the possible letters.
This is highly effective and surprisingly simple, although it could
definitely benefit from some basic knowledge of English grammar and
parts of speech.

Unfortunately, this program is definitely not unbeatqable. Likewise,
with only 6 lives, probably none of the others are as well. As expected,
when facing the word “cwm,” the program guesses AEOIUY [“A” is more
frequent than “E” for three-letter words, apparently], and loses before
it has a chance to guess correctly.

$Words = (file=File.new(“words.txt”)).read.upcase.split(/\n/)
file.close
def hangman_start
puts “Please enter word pattern.”
word_pattern = gets.chomp
possible_words = []
word_pattern.split.length.times do |t|
possible_words << $Words.select{ |word|
word_pattern.split[t].length == word.length}
end

hangman_round word_pattern, possible_words
end
$avail_letters= (“A”…“Z”).to_a
def hangman_round(word_pattern, possible_words, lives=6)
guess(word_pattern, possible_words)
puts word_pattern
puts “Are there any #{$guess}s?\t\tComputer lives=#{lives}”
if gets.chomp==“y”
puts “Please indicate all positions with a #{$guess}”
puts “(0-indexed, comma-delimited)”
gets.chomp.split(/,/).each{|pstr| word_pattern[pstr.to_i] = $guess}
possible_words.each_index do |i|
possible_words[i] = possible_words[i].select{|word|
word.gsub(/[^#{$guess}]/, ‘’) ==
word_pattern.split[i].gsub(/[^#{$guess}]/, '
’)}
end
else
lives -= 1
possible_words.each {|words| words.reject! {|word| word.index $guess}}
end
if word_pattern !~ /_/
puts word_pattern
puts “I win”
elsif lives > 0
hangman_round(word_pattern, possible_words, lives)
else
puts “You win”
end
end
#Guesses by frequency analysis. If a letter appears in a possible
word, it’s a vote for
#that letter. If a word is possible more than once, that’s multiple
votes, but not
#if the letter appears multiple times in a possible word (it’s still one
possibility)
#It then removes that letter from $avail_letters and stores the guess
into $guess
#for convenience
def guess(word_pattern, possible_words)
all_words = possible_words.flatten
guess = $avail_letters.sort_by {|c|
all_words.select{|w|w.index c}.length}.last
$avail_letters -= [guess]
$guess = guess
end
hangman_start