def print_sentence(sentence):
single_sentence = ""
for (word, tag) in sentence:
if not (penn_punct.__contains__(tag) or brown_punct.__contains__(tag)):
single_sentence += str(word)+" "
else:
single_sentence = single_sentence[:len(single_sentence)-1]+str(word)+" "
print single_sentence
def single_sent_word_length(sentence):
total_num_of_letters = 0
word_count = 0
for (word, tag) in sentence:
if not (penn_punct.__contains__(tag) or brown_punct.__contains__(tag)):
word_count += 1.0
for letter in word:
total_num_of_letters+=1.0
average_word_length = total_num_of_letters/word_count
return average_word_length
def word_length_average(tagged_corpus):
total_num_of_letters = 0
word_count = 0
for sent in tagged_corpus:
for (word, tag) in sent:
if not (penn_punct.__contains__(tag) or brown_punct.__contains__(tag)):
word_count += 1.0
for letter in word:
total_num_of_letters+=1.0
average_word_length = total_num_of_letters/word_count
return average_word_length
def count_pronouns_per_words(tagged_corpus):
num_of_pronouns = 0
num_of_words = 0
for sent in tagged_corpus:
for (word, tag) in sent:
if not (penn_punct.__contains__(tag) or brown_punct.__contains__(tag)):
num_of_words += 1.0
if (brown_pron_tags.__contains__(tag)) or (penn_pron_tags.__contains__(tag)):
num_of_pronouns += 1.0
if not num_of_pronouns == 0:
pron_per_word_total = num_of_words/num_of_pronouns
else:
pron_per_word_total = 0.0
return pron_per_word_total
