[SOLUTION] Hangman (#130) -- Guesser Class


My “Guesser” class will guess the letter and the position based on the
words that are “x” characters in length. It will suggest a particular
letter in a particular position that has the highest frequency. It will
filter out words based on characters known in a particular position and
characters known not to be in a particular position.

It can find almost all words within 20 guesses and many is less than 10;
I’ve found that words of 6 characters in length are typically the
toughest to discover with the position and character guessing method.
Once it filters down to one word it will return the word in a string.
If it doesn’t have a word that matches, it will begin guessing based on
random character guesses from the first unknown position.

Note: I did use Unix command “egrep” to quickly parse the word database.
You can fix that in the “load_database” method.

Example Usage based on the term “writhe”

@known = string of characters; “-” is the wildcard

@unknown = array of array of letters that letters

that have been ruled out

@file = Dictionary File - one word per line

require ‘guesser’

guess = Guesser.new(:known => “------”,
:unknown => [[], [], [], [], [], []]
:file => “words.txt”)

guess.list.length # Total Words search from with six characters

guess.guess # Returns a guess of [letter, position]
[:s, 5] # position is based on an index of 0
# I’ve found the best guess to be “s”
# for the last character because it rules out
# all plurals

guess.list.length # The method “guess” created a list of words from the

11153 # database.
# All the words are 6 characters in length.
# Not necessary, just showing you how it filters down
# the words

guess.create_regex # Returns the Regexp used to find the word
/^.{5}$/ # Not necessary, just for example purposes

guess.unknown = [[], [], [], [], [], [“s”]] # Insert an “s” to signify
# that it is not
# in that position of the
# word.

guess.guess # Filters the words and guesses again
[:e, 4] #

guess.list.length # Filtered by the method “guess” to have fewer words.
7045 # Notice 4108 words did not have the last letter “s”

guess.create_regex # Returns the Regexp used to find the list
/^.{4}[^s]$/ #

guess.unknown = [[], [], [], [], [“e”], [“s”]]

[:e, 5]

guess.list.length # Guess filters out 2400 more words based on the
# Regexp
4568 # Note that the “list” is a filtering list and cannot
# be recreated in full
# once it has been filtered. That was a conscious
# decision to preserve resources.

guess.create_regex # Not necessary, just shows for example purposes.

guess.known = “-----e”.split(//) # Must be submitted as an array.
# Converts the string “-----e” into an

                              # array
                              # Probably need to fix this in the
                              # class

[:l, 4]

Eventually (16 guesses down the road), once the word is filtered down it
would look something list this:

guess.unknown = [[“s”, “t”], [“l”, “o”], [], [“i”, “a”, “u”, “o”, “e”,
“n”, “r”], [“e”, “l”, “i”, “u”], [“s”]]

guess.known = “—the”.split(//)