def test_filtered_cmu_entries(self):
entries = entries_from_cmudict()
entries = filter_entries(entries, 'Austen')
self.assertEqual(
entries[:2],
[(u'a', [u'AH0']), (u'a', [u'EY1'])]
)
def _generate_words_letters(number_of_words=100,
order=1,
max_transitions=30,
words_cache=None):
if words_cache is None:
words_cache = WordsCache()
entries = entries_from_cmudict()
existing_words = [word for word, phones in entries]
entries = filter_entries(entries, 'Austen')
sequences = [[letter for letter in word] for word, phones in entries]
markov_tree = generate_markov_tree(sequences, order=order)
number_generated = 0
while number_generated < number_of_words:
cached_words = words_cache.get_words('letters')
new_sequence = generate_new_sequence(markov_tree, max_transitions)
new_word = ''.join(new_sequence)
# Reject words already in the corpus
if new_word in existing_words or new_word in cached_words:
continue
else:
words_cache.add_word('letters', new_word)
print "New word: %s" % new_word
number_generated += 1
return words_cache
def _generate_words_tuples(number_of_words=100,
tuple_length=2,
words_cache=None):
if words_cache is None:
words_cache = WordsCache()
entries = entries_from_cmudict()
existing_words = [word for word, phones in entries]
entries = filter_entries(entries, 'Austen')
entry_words = [entry[0] for entry in entries]
number_generated = 0
while number_generated < number_of_words:
cached_words = words_cache.get_words('tuples')
words_tuple = []
for i in range(tuple_length):
word = random.choice(entry_words)
words_tuple.append(word)
new_word = ''.join(words_tuple)
# Reject words in the corpus
if new_word in cached_words or new_word in existing_words:
continue
else:
words_cache.add_word('tuples', new_word)
print "New word: %s" % new_word
number_generated += 1
return words_cache
def _generate_words_letters(number_of_words=100,
order=1,
max_transitions=30,
words_cache=None):
if words_cache is None:
words_cache = WordsCache()
entries = entries_from_cmudict()
existing_words = [word for word, phones in entries]
entries = filter_entries(entries, 'Austen')
sequences = [[letter for letter in word] for word, phones in entries]
markov_tree = generate_markov_tree(sequences, order=order)
number_generated = 0
while number_generated < number_of_words:
cached_words = words_cache.get_words('letters')
new_sequence = generate_new_sequence(markov_tree, max_transitions)
new_word = ''.join(new_sequence)
# Reject words already in the corpus
if new_word in existing_words or new_word in cached_words:
continue
else:
words_cache.add_word('letters', new_word)
print "New word: %s" % new_word
number_generated += 1
return words_cache
def _generate_words_tuples(number_of_words=100,
tuple_length=2,
words_cache=None):
if words_cache is None:
words_cache = WordsCache()
entries = entries_from_cmudict()
existing_words = [word for word, phones in entries]
entries = filter_entries(entries, 'Austen')
entry_words = [entry[0] for entry in entries]
number_generated = 0
while number_generated < number_of_words:
cached_words = words_cache.get_words('tuples')
words_tuple = []
for i in range(tuple_length):
word = random.choice(entry_words)
words_tuple.append(word)
new_word = ''.join(words_tuple)
# Reject words in the corpus
if new_word in cached_words or new_word in existing_words:
continue
else:
words_cache.add_word('tuples', new_word)
print "New word: %s" % new_word
number_generated += 1
return words_cache
def test_words_correct(self):
"""Test the accuracy of the spelling against existing words."""
entries = entries_from_cmudict()
entries = filter_entries(entries, 'Austen')
number_correct = 0
for entry in entries:
word, phones = entry
# clean_phone modifies phones in-place, so
cleaned_phones = _clean_phones(phones)
calculated_word = phones_to_word(cleaned_phones)
if word == calculated_word:
number_correct += 1
else:
print "Incorrect spelling. Expected %s, got %s. Phones: %s" % (
word, calculated_word, phones
)
self.assertEqual(number_correct, 1001)
def test_entries_from_cmudict(self):
entries = entries_from_cmudict()
self.assertEqual(
entries[:2],
[(u'a', [u'AH0']), (u'a.', [u'EY1'])]
)
def test_entries_from_cmudict(self):
entries = entries_from_cmudict()
self.assertEqual(entries[:2], [(u'a', [u'AH0']), (u'a.', [u'EY1'])])
def test_filtered_cmu_entries(self):
entries = entries_from_cmudict()
entries = filter_entries(entries, 'Austen')
self.assertEqual(entries[:2], [(u'a', [u'AH0']), (u'a', [u'EY1'])])