def test_get_coulomb_matrix(self):
res = get_coulomb_matrix([1, 1], [[0.0, 0.0, 0.0], [0.0, 0.0, 1.0]])
expected_results = numpy.array([[0.5, 1.0], [1.0, 0.5]])
try:
numpy.testing.assert_array_almost_equal(res, expected_results)
except AssertionError as e:
self.fail(e)
def test_get_coulomb_matrix_alpha(self):
nums = [1, 1]
coords = [[0.0, 0.0, 0.0], [0.0, 0.0, .5]]
res = get_coulomb_matrix(nums, coords, alpha=2)
expected_results = numpy.array([[0.5, 4.], [4., 0.5]])
try:
numpy.testing.assert_array_almost_equal(res, expected_results)
except AssertionError as e:
self.fail(e)
def test_get_coulomb_matrix_use_decay(self):
nums = [1, 1, 1]
coords = [[0.0, 0.0, 0.0], [0.0, 0.0, .5], [0.0, 0.5, 0.0]]
res = get_coulomb_matrix(nums, coords, use_decay=True)
expected_results = numpy.array([[0.5, 1., 1.], [1., 0.5, 0.585786],
[1., 0.585786, 0.5]])
try:
numpy.testing.assert_array_almost_equal(res, expected_results)
except AssertionError as e:
self.fail(e)
def test_get_coulomb_matrix(self):
res = get_coulomb_matrix([1, 1], [[0.0, 0.0, 0.0], [0.0, 0.0, 1.0]])
expected_results = numpy.array([
[0.5, 1.0],
[1.0, 0.5]])
try:
numpy.testing.assert_array_almost_equal(
res,
expected_results)
except AssertionError as e:
self.fail(e)
def test_get_coulomb_matrix_alpha(self):
nums = [1, 1]
coords = [[0.0, 0.0, 0.0], [0.0, 0.0, .5]]
res = get_coulomb_matrix(nums, coords, alpha=2)
expected_results = numpy.array([
[0.5, 4.],
[4., 0.5]])
try:
numpy.testing.assert_array_almost_equal(
res,
expected_results)
except AssertionError as e:
self.fail(e)
def test_get_coulomb_matrix_use_decay(self):
nums = [1, 1, 1]
coords = [[0.0, 0.0, 0.0], [0.0, 0.0, .5], [0.0, 0.5, 0.0]]
res = get_coulomb_matrix(nums, coords, use_decay=True)
expected_results = numpy.array([
[0.5, 1., 1.],
[1., 0.5, 0.585786],
[1., 0.585786, 0.5]])
try:
numpy.testing.assert_array_almost_equal(
res,
expected_results)
except AssertionError as e:
self.fail(e)
def _para_transform(self, X):
"""
A single instance of the transform procedure.
This is formulated in a way that the transformations can be done
completely parallel with map.
Parameters
----------
X : object
An object to use for the transform
Returns
-------
value : array
The features extracted from the molecule
Raises
------
ValueError
If the transformer has not been fit.
ValueError
If the size of the transforming molecules are larger than the fit.
"""
if self._max_size is None:
msg = "This %s instance is not fitted yet. Call 'fit' first."
raise ValueError(msg % type(self).__name__)
data = self.convert_input(X)
if len(data.numbers) > self._max_size:
msg = "The fit molecules (%d) were not as large as the ones that"
msg += " are being transformed (%d)."
raise ValueError(msg % (self._max_size, len(data.numbers)))
padding_difference = self._max_size - len(data.numbers)
values = get_coulomb_matrix(data.numbers, data.coords)
if self.sort:
order = numpy.argsort(values.sum(0))[::-1]
values = values[order, :][:, order]
if self.eigen:
values = numpy.linalg.eig(values)[0]
values = numpy.pad(values,
(0, padding_difference),
mode="constant")
return values