################################################################################
#
#	Bayesian Filter - 3.23.09 - Joe Mariglio <www.joemariglio.com>
#
#===============================================================================
#
#	classify text documents.  it's slow but it works.
#
#	supply as many docs as you want.
#
################################################################################

import sys, string, os.path, array, math
class Category:
	count = {}
	total = 0
	name = ""
	def train(self, word):
		if word in self.count.keys():
			self.count[word] +=1
		else:
			self.count[word] = 1
		self.total+=1
	def percentage(self, word):
		if word in self.count.keys():
			return self.count[word] / float(self.total)
		else:
			return 0.01 / total
	def score(self, src, totalAll):
		score_val = 0
		for word in src:
			percent = self.percentage(word)
			score_val += math.log(percent)
		score_val += math.log(self.total / float(totalAll))
		return score_val
	def relevance(self, word, categories):
		per_sum = 0
		for cat in categories:
			per_sum += cat.percentage(word)
		return (percentage(self, word) / per_sum)
class Classifier:
	cats = []
	uniq = []
	def addCat(self, fname):
		cat = Category()
		cat.name = fname
		infile = open(fname, "r")
		lines = infile.readlines()
		infile.close()
		for line in lines:
			words = line.split(" ")
			for word in words:
				cat.train(word)
#				if word not in self.uniq:
#					self.uniq.append(word)
		self.cats.append(cat)
def main(argv):
	c = Classifier()
	cat_w_t = 0
	for name in sys.argv[1:]:
		c.addCat(name)
	for cat in c.cats:
		cat_w_t += cat.total
	for cat in c.cats:
		score = cat.score(c.uniq, cat_w_t)
		print(cat.name + ": " + str(score))
if __name__ == '__main__':
	import sys, string, os.path, random, array
	main(sys.argv[1:])