madlib_generatorP3.py

# code developed by Jackie Cohen; revised by Paul Resnick
# further revised by Colleen van Lent for Python3
import nltk # requires some downloading/installing dependencies to use all its features; numpy is especially tricky to install
import random

# import nltk
nltk.download('punkt')

from nltk import word_tokenize,sent_tokenize

debug = False #True

# get file from user to make mad lib out of
if debug:
	print ("Getting information from file madlib_test.txt...\n")
fname = "madlibtest2.txt" # need a file with this name in directory

f = open(fname, 'r')
para = f.read()
tokens = nltk.word_tokenize(para)
print("TOKENS")
print(tokens)
tagged_tokens = nltk.pos_tag(tokens) # gives us a tagged list of tuples
print("TAGGED TOKENS")
print(tagged_tokens)
if debug:
	print ("First few tagged tokens are:")
	for tup in tagged_tokens[:5]:
		print (tup)

tagmap = {"NN":"a noun","NNS":"a plural noun","VB":"a verb","JJ":"an adjective"}
substitution_probabilities = {"NN":.1,"NNS":.2,"VB":.25,"JJ":.25}

def spaced(word):
	if word in [",", ".", "?", "!", ":"]:
		return word
	else:
		return " " + word

final_words = []


for (word, tag) in tagged_tokens:
	if tag not in substitution_probabilities or random.random() > substitution_probabilities[tag]:
		final_words.append(spaced(word))
	else:
		new_word = input("Please enter %s:\n" % (tagmap[tag]))
		final_words.append(spaced(new_word))

print ("".join(final_words))