def test_encode_without_permutation(self):
corpus = 'hi my name is pat what is your name'
d = DeLorean(corpus, bucket_size=2, n=1)
self.assertEqual(d.encode_without_permutation('00'), ['hi', 'name'])
self.assertEqual(d.encode_without_permutation('10'), ['my', 'name'])
self.assertEqual(d.encode_without_permutation('11'), ['my', 'is'])
def test_encode_without_permutation_padding(self):
corpus = 'hi my name is pat what is your name please tell me'
d = DeLorean(corpus, bucket_size=4, n=2)
self.assertEqual(d.encode_without_permutation('0'), ['hi'])
self.assertEqual(d.encode_without_permutation('1'), ['name'])
self.assertEqual(d.encode_without_permutation('000'), ['hi', 'pat'])
self.assertEqual(d.encode_without_permutation('001'), ['hi', 'your'])
def test_end_to_end(self):
with open('script.txt', 'r') as f:
corpus = f.read()
d = DeLorean(corpus)
for i in range(12):
inp = ''.join(random.choice(['0', '1']) for _ in range(d.max_bits))
inp = inp[:-i]
encoded = d.encode_without_permutation(inp)
random.shuffle(encoded)
decoded = d.decode(encoded)
while len(inp) % d.n != 0:
inp += '0'
self.assertEqual(decoded, inp)