def test_type_mismatch(self):
seq1 = Sequence('ABC')
seq2 = DNA('ACG')
with self.assertRaisesRegex(TypeError,
r'Sequence.*does not match.*DNA'):
hamming(seq1, seq2)
def test_type_mismatch(self):
seq1 = Sequence('ABC')
seq2 = DNA('ACG')
with self.assertRaisesRegex(TypeError,
'Sequence.*does not match.*DNA'):
hamming(seq1, seq2)
def test_non_sequence(self):
seq1 = Sequence('abc')
seq2 = 'abc'
with self.assertRaisesRegex(TypeError, r'seq1.*seq2.*Sequence.*str'):
hamming(seq1, seq2)
with self.assertRaisesRegex(TypeError, r'seq1.*seq2.*Sequence.*str'):
hamming(seq2, seq1)
def test_non_sequence(self):
seq1 = Sequence('abc')
seq2 = 'abc'
with self.assertRaisesRegex(TypeError, 'seq1.*seq2.*Sequence.*str'):
hamming(seq1, seq2)
with self.assertRaisesRegex(TypeError, 'seq1.*seq2.*Sequence.*str'):
hamming(seq2, seq1)
def test_maximum_distance(self):
seq1 = Sequence('ABC')
seq2 = Sequence('CAB')
distance = hamming(seq1, seq2)
self.assertEqual(distance, 1.0)
def test_sequence_subclass(self):
seq1 = DNA('ACG-T')
seq2 = DNA('ACCTT')
distance = hamming(seq1, seq2)
self.assertEqual(distance, 2.0/5.0)
def test_empty_sequences(self):
seq1 = Sequence('')
seq2 = Sequence('')
distance = hamming(seq1, seq2)
npt.assert_equal(distance, np.nan)
def test_empty_sequences(self):
seq1 = Sequence('')
seq2 = Sequence('')
distance = hamming(seq1, seq2)
npt.assert_equal(distance, np.nan)
def test_maximum_distance(self):
seq1 = Sequence('ABC')
seq2 = Sequence('CAB')
distance = hamming(seq1, seq2)
self.assertEqual(distance, 1.0)
def test_mid_range_distance(self):
seq1 = Sequence("abcdefgh")
seq2 = Sequence("1b23ef45")
distance = hamming(seq1, seq2)
self.assertEqual(distance, 5.0/8.0)
def test_sequence_subclass(self):
seq1 = DNA('ACG-T')
seq2 = DNA('ACCTT')
distance = hamming(seq1, seq2)
self.assertEqual(distance, 2.0 / 5.0)
def test_mid_range_distance(self):
seq1 = Sequence("abcdefgh")
seq2 = Sequence("1b23ef45")
distance = hamming(seq1, seq2)
self.assertEqual(distance, 5.0 / 8.0)
def p_distance(seq1, seq2):
'''Modified hamming distance to include only non-gap sites'''
from skbio.sequence import Sequence
from numpy import isnan
myseq1 = str(seq1)
myseq2 = str(seq2)
degapped1 = []
degapped2 = []
for i in range(len(myseq1)):
if myseq1[i] != "-":
if myseq2[i] != "-":
degapped1.append(myseq1[i])
degapped2.append(myseq2[i])
degapped1 = "".join(degapped1)
degapped2 = "".join(degapped2)
#sys.stderr.write(degapped1)
#sys.stderr.write(degapped2)
hamming_dist = hamming(Sequence(degapped1), Sequence(degapped2))
#sys.stderr.write(hamming_dist)
if isnan(hamming_dist):
#sys.stderr.write(seq1.metadata["id"], seq2.metadata["id"])
return 0.0
else:
return hamming_dist
def test_return_type(self):
seq1 = Sequence('ABC')
seq2 = Sequence('ABC')
distance = hamming(seq1, seq2)
self.assertIsInstance(distance, float)
self.assertEqual(distance, 0.0)
def test_return_type(self):
seq1 = Sequence('ABC')
seq2 = Sequence('ABC')
distance = hamming(seq1, seq2)
self.assertIsInstance(distance, float)
self.assertEqual(distance, 0.0)
def test_sequences_with_metadata(self):
# test for #1254
seqs1 = [
Sequence("ACGT"),
Sequence("ACGT", metadata={'id': 'abc'}),
Sequence("ACGT", positional_metadata={'qual': range(4)})
]
seqs2 = [
Sequence("AAAA"),
Sequence("AAAA", metadata={'id': 'def'}),
Sequence("AAAA", positional_metadata={'qual': range(4, 8)})
]
for seqs in seqs1, seqs2:
for seq1, seq2 in itertools.product(seqs, repeat=2):
distance = hamming(seq1, seq2)
self.assertEqual(distance, 0.0)
for seq1, seq2 in itertools.product(seqs1, seqs2):
distance = hamming(seq1, seq2)
self.assertEqual(distance, 0.75)
def test_sequences_with_metadata(self):
# test for #1254
seqs1 = [
Sequence("ACGT"),
Sequence("ACGT", metadata={'id': 'abc'}),
Sequence("ACGT", positional_metadata={'qual': range(4)})
]
seqs2 = [
Sequence("AAAA"),
Sequence("AAAA", metadata={'id': 'def'}),
Sequence("AAAA", positional_metadata={'qual': range(4, 8)})
]
for seqs in seqs1, seqs2:
for seq1, seq2 in itertools.product(seqs, repeat=2):
distance = hamming(seq1, seq2)
self.assertEqual(distance, 0.0)
for seq1, seq2 in itertools.product(seqs1, seqs2):
distance = hamming(seq1, seq2)
self.assertEqual(distance, 0.75)
def calculateHammingDistance(seq1, seq2):
"""Returns hamming distance between two equal length sequences"""
seq1 = Sequence(seq1)
seq2 = Sequence(seq2)
result = hamming(seq1, seq2)
return result
def test_length_mismatch(self):
seq1 = Sequence('ABC')
seq2 = Sequence('ABCD')
with self.assertRaisesRegex(ValueError, 'equal length.*3 != 4'):
hamming(seq1, seq2)
def test_single_character_sequences(self):
seq1 = Sequence('a')
seq2 = Sequence('b')
self.assertEqual(hamming(seq1, seq1), 0.0)
self.assertEqual(hamming(seq1, seq2), 1.0)
def test_single_character_sequences(self):
seq1 = Sequence('a')
seq2 = Sequence('b')
self.assertEqual(hamming(seq1, seq1), 0.0)
self.assertEqual(hamming(seq1, seq2), 1.0)
def test_length_mismatch(self):
seq1 = Sequence('ABC')
seq2 = Sequence('ABCD')
with self.assertRaisesRegex(ValueError, r'equal length.*3 != 4'):
hamming(seq1, seq2)
