A python script to generate a dictionary to be used with NaiveBayes

Dictionary.py

#This script reads all files in all directories of the folder taken from the openClassroom site and generates the dictionary, which we can then store in a file 
folders = ["spam-train","spam-test","nonspam-train","nonspam-test"]
import os,sys
#We need a dictionary to store word occurences. What we can do is create a default dict and then update the frequencies. Write it all into a file all at once. 
from collections import * 
dictionary = defaultdict(int)
fdict = open(sys.argv[2],'w') #File to write all the entries in the dictionary 
for root,dirnames,filenames in os.walk(sys.argv[1]):
	for d in dirnames: #For each directory 
		for f in os.listdir(d):
			data = open ( os.path.join(sys.argv[1],d,f),'r')
			for line in data:
				words = line.split(" ")#Split words on space 
				for w in words:
					dictionary[w] += 1
for k,v in dictionary.iteritems():
	fdict.write(k +" "+str(dictionary[k])+"\n")