def test_levenshtein(self):
self.assertEqual(0, Distance.levenshtein('', ''))
self.assertEqual(3, Distance.levenshtein('', 'abc'))
self.assertEqual(2, Distance.levenshtein('abcde', 'acd'))
self.assertEqual(4, Distance.levenshtein('abcabcabc', 'bcabc'))
self.assertEqual(5, Distance.levenshtein('abcabcabc', 'bcac'))
self.assertEqual(5, Distance.levenshtein('bcac', 'abcabcabc'))
for _ in range(100):
s = Sequence.randomDnaKmer(100)
tmp = list(s)
edited = []
edits = 0
i = 0
while i < len(tmp):
r = random()
if r < 10 / 100:
# deletion
edits += 1
elif r < 20 / 100:
# insertion
edited.append('-')
i -= 1 # do not move index
edits += 1
elif r < 30 / 100:
# substitution
edited.append('*')
edits += 1
else:
edited.append(tmp[i])
i += 1
edited = ''.join(edited)
self.assertLessEqual(Distance.levenshtein(edited, s), edits)
def test_hash_certainty(self):
for M in (8, 16, 32, 64):
for K in range(8, 257, 8):
dkh = DnaKmerHasher(K, M)
dna = Sequence.randomDnaKmer(K)
h = dkh.hash(dna)
for _ in range(100):
self.assertEqual(h, dkh.hash(dna))
def test_randomDnaKmer(self):
c = Counter()
for k in range(0, 101):
seq = Sequence.randomDnaKmer(k)
self.assertEqual(k, len(seq))
c += Counter(seq)
self.assertEqual(tuple(sorted(c.keys())), tuple('ACGT'))
self.assertGreaterEqual(chisquare(list(c.values()))[1], 0.05)
def test_to_string(self):
for M in (1, 2, 4, 8, 11, 16, 17, 32, 64):
for K in range(4, 257, 4):
for _ in range(100):
dkh = DnaKmerHasher(K, M)
dna = Sequence.randomDnaKmer(K)
hash = dkh.hash(dna)
dkh = DnaKmerHasher(dkh.str())
self.assertEqual(hash, dkh.hash(dna))
def test_hash_uniform1(self):
for M in (8, 16, 32, 64):
upper_bound = 2**M - 1
for K in range(8, 257, 8):
dna = Sequence.randomDnaKmer(K)
hashes = []
for _ in range(10000):
dkh = DnaKmerHasher(K, M)
hashes.append(dkh.hash(dna) / upper_bound)
self.assertAlmostEqual(min(hashes), 0, 2)
self.assertAlmostEqual(max(hashes), 1, 2)
hist, edges = np.histogram(hashes, bins=50, density=True)
self.assertGreater(chisquare(hist)[1], 0.05)
def test_hamming(self):
self.assertEqual(Distance.hamming('', ''), 0)
self.assertEqual(Distance.hamming("123456789", "12345678*"), 1)
with self.assertRaises(RuntimeError):
Distance.hamming('12', '123')
for i in range(100):
s = Sequence.randomDnaKmer(100)
tmp = list(s)
for j in range(i):
tmp[j] = '*'
t = ''.join(tmp)
self.assertEqual(Distance.hamming(s, t), i)
def test_hash_uniform0(self):
for M in (8, 16, 32, 64):
upper_bound = 2**M - 1
for K in range(8, 257, 8):
hashes = []
dkh = DnaKmerHasher(K, M)
for _ in range(10000):
dna = Sequence.randomDnaKmer(K)
hashes.append(dkh.hash(dna) / upper_bound)
# for the random hash values to reach the boundary (0 and 1), k should be bigger than M/2
self.assertAlmostEqual(0, min(hashes), 2)
self.assertAlmostEqual(1, max(hashes), 2)
hist, edges = np.histogram(hashes, bins=50, density=True)
self.assertGreater(chisquare(hist)[1], 0.05)
def test_reverseComplementRna(self):
self.assertEqual('auauauaaaucuuguuuCUAGa', Sequence.reverseComplementRna('uCUAGaaacaagauuuauauau'))
for k in range(0, 101):
rna = Sequence.randomRnaKmer(k)
self.assertEqual(Sequence.complementRna(Sequence.reverse(rna)), Sequence.reverseComplementRna(rna))
self.assertEqual(Sequence.reverse(Sequence.complementRna(rna)), Sequence.reverseComplementRna(rna))
def test_reverseComplementDna(self):
self.assertEqual('atatataaatcttgtttCTAGa', Sequence.reverseComplementDna('tCTAGaaacaagatttatatat'))
for k in range(0, 101):
dna = Sequence.randomDnaKmer(k)
self.assertEqual(Sequence.complementDna(Sequence.reverse(dna)), Sequence.reverseComplementDna(dna))
self.assertEqual(Sequence.reverse(Sequence.complementDna(dna)), Sequence.reverseComplementDna(dna))
def test_reverse(self):
for k in range(0, 101):
dna = Sequence.randomDnaKmer(k)
self.assertEqual(Sequence.reverse(dna), dna[::-1])