markov-chain-meetup-dojo

README.md

• Source http://tinyurl.com/alice-clj

generated-sentences.txt

into it   llustration  you can t s the
 ho cares for   he fee as set forth
or the court   round     
t seem to feel which seemed   went on a
of her    lice felt quite   
 llustration    of white   iterary 
than a    ventured  perhaps not claim a
archbishop find another moment    hat size  round
  hite  he  said  ound  
times as well as she did the animals that lay on
them all the    begged the   ouse
of or immediate adoption of the     lice
by e ooks   by that they re a tidy
     ouse in any work  providing
  uchess     cross and one 
     quiver all the little quicker 
in a minute or really  what does it was labeled
of  hy   only wish to open any part
it quite a bit hurt the  in  rench 
as she heard this very good naturedly began running about for
that it  at her sister  roofreading   
          
  sentence set out of time    
    llustration  arroll    
over with his e ook of footsteps    
  and conquest   lice to   llustration
provisions   from the face and rapped loudly  
 dear   he game    said 
 lice     sir  roduced by the
 she ran out of   ouse

markov_chain.clj

(ns markov-dojo.core
  (:require [clojure.java.io :as io]
            [clojure.string :as s]))

(defn load-resource
  [res-name]
  (slurp (io/resource res-name)))

(def end-of-sentence #".*[.;:?!]$")

(defn split-by-word
  [txt]
  (s/split txt #"[ -_]"))

(defn foo
  "I don't do a whole lot."
  [x]
  (println x "Hello, World!"))

(defn pairs
  [s]
  (map vector s (drop 1 s)))

(defn comb
  [acc [a b]]
  (let [merge-fn (fn [one two] (apply conj one two))]
    (merge-with merge-fn acc {a [b]})))

(defn markov-chain
  [pairs]
  (reduce comb {} pairs))

(defn random-token
  [vals]
  (nth vals (rand-int (count vals))))

(defn end-of-sentence?
  [token]
  (re-matches end-of-sentence token))

(defn random-sentence
  [chain max-tokens]
  (let [tokens (keys chain)]
    (loop [sentence []
           token (random-token tokens)]
      (let [next-token (random-token (get chain token))
            next-sentence (conj sentence next-token)]
        (if (or (end-of-sentence? next-token)
                (<= max-tokens (count next-sentence)))
          (s/join " " sentence)
          (recur next-sentence next-token))))))