def test_fit(self):
a = EncodedBond()
a.fit(ALL_DATA)
self.assertEqual(
a._element_pairs,
set([('H', 'O'), ('O', 'O'), ('N', 'O'), ('C', 'O'), ('C', 'H'),
('H', 'N'), ('H', 'H'), ('C', 'C'), ('C', 'N'), ('N', 'N')]))
def test_get_labels(self):
a = EncodedBond(segments=2, start=0., end=1.)
X = a.fit_transform([METHANE])
labels = a.get_labels()
self.assertEqual(X.shape[1], len(labels))
expected = (
'C-H_0.0', 'C-H_1.0',
'H-H_0.0', 'H-H_1.0',
)
self.assertEqual(labels, expected)
def test_max_depth_neg(self):
a = EncodedBond(max_depth=-1)
# This is a cheap test to prevent needing all the values here
expected_results = numpy.array([
0.503237244954, # mean
0.857850829564, # std
0., # min
7.15861023, # max
])
try:
m = a.fit_transform([BIG])
assert_close_statistics(m, expected_results)
except AssertionError as e:
self.fail(e)
def test_fit_transform(self):
a = EncodedBond()
# This is a cheap test to prevent needing all the values here
expected_results = numpy.array([
0.042672, # mean
0.246663, # std
0., # min
2.392207, # max
])
try:
m = a.fit_transform([METHANE])
assert_close_statistics(m, expected_results)
except AssertionError as e:
self.fail(e)
def test_spacing_log(self):
a = EncodedBond(spacing="log")
# This is a cheap test to prevent needing all the values here
expected_results = numpy.array([
0.072768, # mean
0.318508, # std
0., # min
2.339376, # max
])
try:
m = a.fit_transform([METHANE])
assert_close_statistics(m, expected_results)
except AssertionError as e:
self.fail(e)
def test_spacing_inverse(self):
a = EncodedBond(spacing="inverse")
# This is a cheap test to prevent needing all the values here
expected_results = numpy.array([
0.051207, # mean
0.269248, # std
0., # min
2.387995, # max
])
try:
m = a.fit_transform([METHANE])
assert_close_statistics(m, expected_results)
except AssertionError as e:
self.fail(e)
def test_max_depth_1(self):
a = EncodedBond(max_depth=1)
# This is a cheap test to prevent needing all the values here
expected_results = numpy.array([
0.0443793, # mean
0.33766942, # std
0., # min
5.76559336, # max
])
try:
m = a.fit_transform([BIG])
assert_close_statistics(m, expected_results)
except AssertionError as e:
self.fail(e)
def test_smoothing_function(self):
a = EncodedBond(smoothing="norm_cdf")
# This is a cheap test to prevent needing all the values here
expected_results = numpy.array([
3.859534e+000, # mean
2.182923e+000, # std
0., # min
6.000000e+000, # max
])
try:
m = a.fit_transform([METHANE])
assert_close_statistics(m, expected_results)
except AssertionError as e:
self.fail(e)
def test_small_to_large_transform(self):
a = EncodedBond()
a.fit([METHANE])
# This is a cheap test to prevent needing all the values here
expected_results = numpy.array([
9.207308e-001, # mean
1.062388e+000, # std
0., # min
5.023670e+000, # max
])
try:
m = a.transform([BIG])
assert_close_statistics(m, expected_results)
except AssertionError as e:
self.fail(e)
def test_large_to_small_transform(self):
a = EncodedBond()
a.fit([MID])
# This is a cheap test to prevent needing all the values here
expected_results = numpy.array([
0.014224, # mean
0.143824, # std
0., # min
2.392207, # max
])
try:
m = a.transform([METHANE])
assert_close_statistics(m, expected_results)
except AssertionError as e:
self.fail(e)
def test_max_depth_3(self):
a = EncodedBond(max_depth=3)
# This is a cheap test to prevent needing all the values here
expected_results = numpy.array([
0.18434482, # mean
0.62589799, # std
0., # min
7.15861023, # max
])
try:
m = a.fit_transform([BIG])
assert_close_statistics(m, expected_results)
except AssertionError as e:
self.fail(e)
def test_form_0(self):
a = EncodedBond(form=0)
# This is a cheap test to prevent needing all the values here
expected_results = numpy.array([
0.085345, # mean
0.343574, # std
0., # min
2.392207, # max
])
try:
m = a.fit_transform([METHANE])
self.assertEqual(m.shape, (1, 100))
assert_close_statistics(m, expected_results)
except AssertionError as e:
self.fail(e)
def test_add_unknown(self):
a = EncodedBond(add_unknown=True)
a.fit([METHANE])
# This is a cheap test to prevent needing all the values here
expected_results = numpy.array([
0.09105, # mean
0.231761, # std
0., # min
1.869012, # max
])
try:
m = a.transform([MID])
self.assertEqual(m.shape, (1, 300))
assert_close_statistics(m, expected_results)
except AssertionError as e:
self.fail(e)
def test_spacing_invalid(self):
a = EncodedBond(spacing="not valid")
with self.assertRaises(KeyError):
a.fit_transform([METHANE])
def test_smoothing_function_error(self):
a = EncodedBond(smoothing="not valid")
with self.assertRaises(KeyError):
a.fit_transform([METHANE])
def test_transform_before_fit(self):
a = EncodedBond()
with self.assertRaises(ValueError):
a.transform(ALL_DATA)