def test_exception_on_incorrect_average(self):
'''
Test exception on incorrect average.
'''
matrix = cm()
matrix['p']['p'] += 1
matrix['p']['f'] += 1
self.assertRaises(
SegmentationMetricError, fmeasure, matrix, version='incorrect')
def test_setitem(self):
'''
Ensure that __setitem__ raises an AttributeError
'''
exception = False
matrix = cm()
try:
matrix['a'] = 0
except AttributeError:
exception = True
self.assertTrue(exception, 'AttributeError not raised')
def test_matrix_set_add(self):
'''
Test matrix.
'''
matrix = cm()
matrix['p']['p'] += 2
matrix['p']['n'] = 3
self.assertEqual(matrix['p']['p'], 2)
self.assertEqual(matrix['p']['n'], 3)
self.assertEqual(matrix['p']['f'], 0)
self.assertEqual(matrix['a']['b'], 0)
def test_setitem(self):
'''
Ensure that __setitem__ raises an AttributeError
'''
exception = False
matrix = cm()
try:
matrix['a'] = 0
except AttributeError:
exception = True
self.assertTrue(exception, 'AttributeError not raised')
def test_exception_on_incorrect_average(self):
'''
Test exception on incorrect average.
'''
matrix = cm()
matrix['p']['p'] += 1
matrix['p']['f'] += 1
self.assertRaises(SegmentationMetricError,
fmeasure,
matrix,
version='incorrect')
def test_matrix_set_add(self):
'''
Test matrix.
'''
matrix = cm()
matrix['p']['p'] += 2
matrix['p']['n'] = 3
self.assertEqual(matrix['p']['p'], 2)
self.assertEqual(matrix['p']['n'], 3)
self.assertEqual(matrix['p']['f'], 0)
self.assertEqual(matrix['a']['b'], 0)
def test_matrix_classes(self):
'''
Test matrix.
'''
matrix = cm()
matrix['p']['p'] += 2
matrix['p']['n'] = 3
self.assertEqual(matrix['p']['p'], 2)
self.assertEqual(matrix['p']['n'], 3)
self.assertEqual(matrix['p']['f'], 0)
self.assertEqual(matrix['a']['b'], 0)
self.assertEqual(matrix.classes(), set(['p', 'n', 'a', 'b', 'f']))
def test_matrix_classes(self):
'''
Test matrix.
'''
matrix = cm()
matrix['p']['p'] += 2
matrix['p']['n'] = 3
self.assertEqual(matrix['p']['p'], 2)
self.assertEqual(matrix['p']['n'], 3)
self.assertEqual(matrix['p']['f'], 0)
self.assertEqual(matrix['a']['b'], 0)
self.assertEqual(matrix.classes(), set(['p', 'n', 'a', 'b', 'f']))
def test_fmeasure(self):
'''
Test FMeasure.
'''
matrix = cm()
matrix['p']['p'] += 1
matrix['p']['f'] += 1
self.assertAlmostEqual(
__fmeasure__(matrix, 'p'), Decimal('0.66666'), 4)
self.assertEqual(__fmeasure__(matrix, 'f'), Decimal('0'))
self.assertAlmostEqual(fmeasure(matrix, version=Average.micro),
Decimal('0.66666'), 4)
self.assertAlmostEqual(fmeasure(matrix, version=Average.macro),
Decimal('0.33333'), 4)
self.assertAlmostEqual(fmeasure(matrix, classification='p'),
Decimal('0.66666'), 4)
matrix = cm()
matrix['p']['p'] += 1
matrix['p']['f'] += 3
matrix['f']['p'] += 1
self.assertAlmostEqual(fmeasure(matrix, version=Average.micro),
Decimal('0.33333'), 4)
self.assertAlmostEqual(fmeasure(matrix, version=Average.macro),
Decimal('0.16666'), 4)
self.assertAlmostEqual(__fmeasure__(matrix, 'p'),
Decimal('0.33333'), 4)
self.assertAlmostEqual(fmeasure(matrix, classification='p'),
Decimal('0.33333'), 4)
self.assertEqual(__fmeasure__(matrix, 'f'), Decimal('0'))
matrix = cm()
matrix['p']['p'] += 5
matrix['p']['f'] += 2
matrix['f']['p'] += 1
matrix['f']['f'] += 2
self.assertAlmostEqual(fmeasure(matrix, version=Average.micro),
Decimal('0.68421'), 4)
self.assertAlmostEqual(fmeasure(matrix, version=Average.macro),
Decimal('0.67032'), 4)
self.assertAlmostEqual(__fmeasure__(matrix, 'p'),
Decimal('0.76923'), 4)
self.assertAlmostEqual(__fmeasure__(matrix, 'f'),
Decimal('0.57142'), 4)
matrix = cm()
matrix['p']['f'] += 2
self.assertEqual(fmeasure(matrix), Decimal('0'))
self.assertEqual(__fmeasure__(matrix, 'p'), Decimal('0'))
self.assertEqual(__fmeasure__(matrix, 'f'), Decimal('0'))
matrix = cm()
matrix['p']['p'] += 2
self.assertEqual(fmeasure(matrix), Decimal('1'))
self.assertEqual(__fmeasure__(matrix, 'p'), Decimal('1'))
self.assertEqual(__fmeasure__(matrix, 'f'), Decimal('0'))
matrix = cm()
self.assertEqual(fmeasure(matrix), Decimal('0'))
self.assertEqual(__fmeasure__(matrix, 'p'), Decimal('0'))
self.assertEqual(__fmeasure__(matrix, 'f'), Decimal('0'))
def test_fmeasure(self):
'''
Test FMeasure.
'''
matrix = cm()
matrix['p']['p'] += 1
matrix['p']['f'] += 1
self.assertAlmostEqual(__fmeasure__(matrix, 'p'), Decimal('0.66666'),
4)
self.assertEqual(__fmeasure__(matrix, 'f'), Decimal('0'))
self.assertAlmostEqual(fmeasure(matrix, version=Average.micro),
Decimal('0.66666'), 4)
self.assertAlmostEqual(fmeasure(matrix, version=Average.macro),
Decimal('0.33333'), 4)
self.assertAlmostEqual(fmeasure(matrix, classification='p'),
Decimal('0.66666'), 4)
matrix = cm()
matrix['p']['p'] += 1
matrix['p']['f'] += 3
matrix['f']['p'] += 1
self.assertAlmostEqual(fmeasure(matrix, version=Average.micro),
Decimal('0.33333'), 4)
self.assertAlmostEqual(fmeasure(matrix, version=Average.macro),
Decimal('0.16666'), 4)
self.assertAlmostEqual(__fmeasure__(matrix, 'p'), Decimal('0.33333'),
4)
self.assertAlmostEqual(fmeasure(matrix, classification='p'),
Decimal('0.33333'), 4)
self.assertEqual(__fmeasure__(matrix, 'f'), Decimal('0'))
matrix = cm()
matrix['p']['p'] += 5
matrix['p']['f'] += 2
matrix['f']['p'] += 1
matrix['f']['f'] += 2
self.assertAlmostEqual(fmeasure(matrix, version=Average.micro),
Decimal('0.68421'), 4)
self.assertAlmostEqual(fmeasure(matrix, version=Average.macro),
Decimal('0.67032'), 4)
self.assertAlmostEqual(__fmeasure__(matrix, 'p'), Decimal('0.76923'),
4)
self.assertAlmostEqual(__fmeasure__(matrix, 'f'), Decimal('0.57142'),
4)
matrix = cm()
matrix['p']['f'] += 2
self.assertEqual(fmeasure(matrix), Decimal('0'))
self.assertEqual(__fmeasure__(matrix, 'p'), Decimal('0'))
self.assertEqual(__fmeasure__(matrix, 'f'), Decimal('0'))
matrix = cm()
matrix['p']['p'] += 2
self.assertEqual(fmeasure(matrix), Decimal('1'))
self.assertEqual(__fmeasure__(matrix, 'p'), Decimal('1'))
self.assertEqual(__fmeasure__(matrix, 'f'), Decimal('0'))
matrix = cm()
self.assertEqual(fmeasure(matrix), Decimal('0'))
self.assertEqual(__fmeasure__(matrix, 'p'), Decimal('0'))
self.assertEqual(__fmeasure__(matrix, 'f'), Decimal('0'))
def test_precision(self):
'''
Test precision.
'''
matrix = cm()
matrix['p']['p'] += 1
matrix['p']['f'] += 1
self.assertEqual(__precision__(matrix, 'p'), Decimal('0.5'))
self.assertEqual(__precision__(matrix, 'f'), Decimal('0'))
self.assertEqual(
precision(matrix, version=Average.micro), Decimal('0.5'))
self.assertEqual(
precision(matrix, version=Average.macro), Decimal('0.25'))
matrix = cm()
matrix['p']['p'] += 1
matrix['p']['f'] += 3
matrix['f']['p'] += 1
self.assertEqual(
precision(matrix, version=Average.micro), Decimal('0.2'))
self.assertEqual(
precision(matrix, version=Average.macro), Decimal('0.125'))
self.assertEqual(__precision__(matrix, 'p'), Decimal('0.25'))
self.assertEqual(__precision__(matrix, 'f'), Decimal('0'))
matrix = cm()
matrix['p']['p'] += 5
matrix['p']['f'] += 2
matrix['f']['p'] += 1
matrix['f']['f'] += 2
self.assertEqual(
precision(matrix, version=Average.micro), Decimal('0.7'))
self.assertAlmostEqual(precision(matrix, version=Average.macro),
Decimal('0.69047'), 4)
self.assertAlmostEqual(__precision__(matrix, 'p'),
Decimal('0.71428'), 4)
self.assertAlmostEqual(__precision__(matrix, 'f'),
Decimal('0.66666'), 4)
matrix = cm()
matrix['p']['f'] += 2
self.assertEqual(precision(matrix), Decimal('0'))
self.assertEqual(__precision__(matrix, 'p'), Decimal('0'))
self.assertEqual(__precision__(matrix, 'f'), Decimal('0'))
matrix = cm()
matrix['p']['p'] += 2
self.assertEqual(precision(matrix), Decimal('1'))
self.assertEqual(__precision__(matrix, 'p'), Decimal('1'))
self.assertEqual(__precision__(matrix, 'f'), Decimal('0'))
matrix = cm()
self.assertEqual(precision(matrix), Decimal('0'))
self.assertEqual(__precision__(matrix, 'p'), Decimal('0'))
self.assertEqual(__precision__(matrix, 'f'), Decimal('0'))
def test_precision(self):
'''
Test precision.
'''
matrix = cm()
matrix['p']['p'] += 1
matrix['p']['f'] += 1
self.assertEqual(__precision__(matrix, 'p'), Decimal('0.5'))
self.assertEqual(__precision__(matrix, 'f'), Decimal('0'))
self.assertEqual(precision(matrix, version=Average.micro),
Decimal('0.5'))
self.assertEqual(precision(matrix, version=Average.macro),
Decimal('0.25'))
matrix = cm()
matrix['p']['p'] += 1
matrix['p']['f'] += 3
matrix['f']['p'] += 1
self.assertEqual(precision(matrix, version=Average.micro),
Decimal('0.2'))
self.assertEqual(precision(matrix, version=Average.macro),
Decimal('0.125'))
self.assertEqual(__precision__(matrix, 'p'), Decimal('0.25'))
self.assertEqual(__precision__(matrix, 'f'), Decimal('0'))
matrix = cm()
matrix['p']['p'] += 5
matrix['p']['f'] += 2
matrix['f']['p'] += 1
matrix['f']['f'] += 2
self.assertEqual(precision(matrix, version=Average.micro),
Decimal('0.7'))
self.assertAlmostEqual(precision(matrix, version=Average.macro),
Decimal('0.69047'), 4)
self.assertAlmostEqual(__precision__(matrix, 'p'), Decimal('0.71428'),
4)
self.assertAlmostEqual(__precision__(matrix, 'f'), Decimal('0.66666'),
4)
matrix = cm()
matrix['p']['f'] += 2
self.assertEqual(precision(matrix), Decimal('0'))
self.assertEqual(__precision__(matrix, 'p'), Decimal('0'))
self.assertEqual(__precision__(matrix, 'f'), Decimal('0'))
matrix = cm()
matrix['p']['p'] += 2
self.assertEqual(precision(matrix), Decimal('1'))
self.assertEqual(__precision__(matrix, 'p'), Decimal('1'))
self.assertEqual(__precision__(matrix, 'f'), Decimal('0'))
matrix = cm()
self.assertEqual(precision(matrix), Decimal('0'))
self.assertEqual(__precision__(matrix, 'p'), Decimal('0'))
self.assertEqual(__precision__(matrix, 'f'), Decimal('0'))
def __boundary_confusion_matrix__(*args, **kwargs):
'''
Create a confusion matrix using boundary edit distance.
'''
# Trim kwargs
metric_kwargs = dict(kwargs)
del metric_kwargs['return_parts']
del metric_kwargs['one_minus']
# Obtain statistics
statistics = __boundary_statistics__(*args, **metric_kwargs)
# Get parameters
n_t = kwargs['n_t']
weight = kwargs['weight']
# Initialize
matrix = cm()
fnc_weight_t = weight[2]
# Add matches
for match in statistics['matches']:
matrix[match][match] += 1
# Add weighted near misses
for transposition in statistics['transpositions']:
match = transposition[2]
matrix[match][match] += fnc_weight_t([transposition], n_t)
# Add confusion errors
for substitution in statistics['substitutions']:
hyp, ref = substitution
matrix[hyp][ref] += 1
# Add full misses
for addition in statistics['additions']:
hyp, ref = None, None
boundary_type, side = addition
if side == 'a':
hyp = None
ref = boundary_type
else: # side == 'b'
hyp = boundary_type
ref = None
assert side == 'a' or side == 'b'
matrix[hyp][ref] += 1
return matrix
def __boundary_confusion_matrix__(*args, **kwargs):
'''
Create a confusion matrix using boundary edit distance.
'''
# Trim kwargs
metric_kwargs = dict(kwargs)
del metric_kwargs['return_parts']
del metric_kwargs['one_minus']
# Obtain statistics
statistics = __boundary_statistics__(*args, **metric_kwargs)
# Get parameters
n_t = kwargs['n_t']
weight = kwargs['weight']
# Initialize
matrix = cm()
fnc_weight_t = weight[2]
# Add matches
for match in statistics['matches']:
matrix[match][match] += 1
# Add weighted near misses
for transposition in statistics['transpositions']:
match = transposition[2]
matrix[match][match] += fnc_weight_t([transposition], n_t)
# Add confusion errors
for substitution in statistics['substitutions']:
hyp, ref = substitution
matrix[hyp][ref] += 1
# Add full misses
for addition in statistics['additions']:
hyp, ref = None, None
boundary_type, side = addition
if side == 'a':
hyp = None
ref = boundary_type
else: # side == 'b'
hyp = boundary_type
ref = None
assert side == 'a' or side == 'b'
matrix[hyp][ref] += 1
return matrix
def test_recall(self):
'''
Test recall.
'''
matrix = cm()
matrix['p']['p'] += 1
matrix['p']['f'] += 1
self.assertEqual(__recall__(matrix, 'p'), Decimal('1.0'))
self.assertEqual(__recall__(matrix, 'f'), Decimal('0'))
self.assertEqual(recall(matrix, version=Average.micro), Decimal('0.5'))
self.assertEqual(recall(matrix, version=Average.macro), Decimal('0.5'))
matrix = cm()
matrix['p']['p'] += 1
matrix['p']['f'] += 3
matrix['f']['p'] += 1
self.assertEqual(recall(matrix, version=Average.micro), Decimal('0.2'))
self.assertEqual(
recall(matrix, version=Average.macro), Decimal('0.25'))
self.assertEqual(__recall__(matrix, 'p'), Decimal('0.5'))
self.assertEqual(__recall__(matrix, 'f'), Decimal('0'))
matrix = cm()
matrix['p']['p'] += 5
matrix['p']['f'] += 2
matrix['f']['p'] += 1
matrix['f']['f'] += 2
self.assertEqual(recall(matrix, version=Average.micro), Decimal('0.7'))
self.assertAlmostEqual(recall(matrix, version=Average.macro),
Decimal('0.66666'), 4)
self.assertAlmostEqual(__recall__(matrix, 'p'),
Decimal('0.83333'), 4)
self.assertAlmostEqual(__recall__(matrix, 'f'),
Decimal('0.5'), 4)
matrix = cm()
matrix['p']['f'] += 2
self.assertEqual(recall(matrix), Decimal('0'))
self.assertEqual(__recall__(matrix, 'p'), Decimal('0'))
self.assertEqual(__recall__(matrix, 'f'), Decimal('0'))
matrix = cm()
matrix['p']['p'] += 2
self.assertEqual(recall(matrix), Decimal('1'))
self.assertEqual(__recall__(matrix, 'p'), Decimal('1'))
self.assertEqual(__recall__(matrix, 'f'), Decimal('0'))
matrix = cm()
self.assertEqual(recall(matrix), Decimal('0'))
self.assertEqual(__recall__(matrix, 'p'), Decimal('0'))
self.assertEqual(__recall__(matrix, 'f'), Decimal('0'))
def test_recall(self):
'''
Test recall.
'''
matrix = cm()
matrix['p']['p'] += 1
matrix['p']['f'] += 1
self.assertEqual(__recall__(matrix, 'p'), Decimal('1.0'))
self.assertEqual(__recall__(matrix, 'f'), Decimal('0'))
self.assertEqual(recall(matrix, version=Average.micro), Decimal('0.5'))
self.assertEqual(recall(matrix, version=Average.macro), Decimal('0.5'))
matrix = cm()
matrix['p']['p'] += 1
matrix['p']['f'] += 3
matrix['f']['p'] += 1
self.assertEqual(recall(matrix, version=Average.micro), Decimal('0.2'))
self.assertEqual(recall(matrix, version=Average.macro),
Decimal('0.25'))
self.assertEqual(__recall__(matrix, 'p'), Decimal('0.5'))
self.assertEqual(__recall__(matrix, 'f'), Decimal('0'))
matrix = cm()
matrix['p']['p'] += 5
matrix['p']['f'] += 2
matrix['f']['p'] += 1
matrix['f']['f'] += 2
self.assertEqual(recall(matrix, version=Average.micro), Decimal('0.7'))
self.assertAlmostEqual(recall(matrix, version=Average.macro),
Decimal('0.66666'), 4)
self.assertAlmostEqual(__recall__(matrix, 'p'), Decimal('0.83333'), 4)
self.assertAlmostEqual(__recall__(matrix, 'f'), Decimal('0.5'), 4)
matrix = cm()
matrix['p']['f'] += 2
self.assertEqual(recall(matrix), Decimal('0'))
self.assertEqual(__recall__(matrix, 'p'), Decimal('0'))
self.assertEqual(__recall__(matrix, 'f'), Decimal('0'))
matrix = cm()
matrix['p']['p'] += 2
self.assertEqual(recall(matrix), Decimal('1'))
self.assertEqual(__recall__(matrix, 'p'), Decimal('1'))
self.assertEqual(__recall__(matrix, 'f'), Decimal('0'))
matrix = cm()
self.assertEqual(recall(matrix), Decimal('0'))
self.assertEqual(__recall__(matrix, 'p'), Decimal('0'))
self.assertEqual(__recall__(matrix, 'f'), Decimal('0'))