Simple Markov Chain programme and helper utilities

markov.py

#!/usr/bin/env python

import sys
import random

class Markov:
	def __init__(self):
		self.h = {}

	def add( self, x, y ):
		if x not in self.h:
			self.h[x] = {}

		if y not in self.h[x]:
			self.h[x][y] = 0

		self.h[x][y] = self.h[x][y] + 1

	def next( self, x ):
		# Calculate which is the next word
		total = 0
		for k in self.h[x]:
			total = total + self.h[x][k]

		r = random.randint( 0, total -1 )

		i = 0
		for k in self.h[x]:
			i = i + self.h[x][k]
			if r < i:
				return k


if __name__ == '__main__':
	markov = Markov()

	for line in sys.stdin.readlines():
		line = line.rstrip("\n")

		( x, y ) = line.split()

		markov.add( x, y )

	#print markov.h

	i = 0
	n = "the"

	if len(sys.argv) > 1:
		n = sys.argv[1]

	while i < 100:
		n = markov.next( n )
		print n,
		i = i + 1

markov.sh

#!/bin/bash

SOURCE=$1

if [ -z $SOURCE ]; then
	# use the king james bible
	SOURCE="http://www.gutenberg.org/ebooks/10.txt.utf-8"
fi

FN=$(echo $SOURCE | awk -F\/ '{print $NF}')
if [ ! -f $FN ]; then
	echo $SOURCE | egrep -q -e '^http://' && wget $SOURCE
fi

if [ ! -f /tmp/markov.list ]; then
	./split.pl $FN > /tmp/markov.list
fi

./markov.py < /tmp/markov.list

split.pl

#!/usr/bin/perl -s
use strict;

local $/;
my $text = <>;
chomp $text;
my @a = split /\s+/ => $text;

while( scalar @a > 1 ){
	my $x = shift @a;
	print "$x $a[0]\n";
}