def test_apply_axis(self, x):
""" Tests apply along axis """
x1 = ds.apply_along_axis(_sum_and_mult, 0, x)
self.assertTrue(x1.shape, (1, 3))
self.assertTrue(x1._reg_shape, (1, 2))
self.assertTrue(_equal_arrays(x1.collect(), np.array([12, 15, 18])))
self.assertTrue(_validate_array(x1))
x1 = ds.apply_along_axis(_sum_and_mult, 1, x)
self.assertTrue(x1.shape, (3, 1))
self.assertTrue(x1._reg_shape, (2, 1))
self.assertTrue(_equal_arrays(x1.collect(False),
np.array([[6], [15], [24]])))
self.assertTrue(_validate_array(x1))
x1 = ds.apply_along_axis(_sum_and_mult, 1, x, 2)
self.assertTrue(x1.shape, (3, 1))
self.assertTrue(x1._reg_shape, (2, 1))
self.assertTrue(_equal_arrays(x1.collect(False),
np.array([[8], [17], [26]])))
self.assertTrue(_validate_array(x1))
x1 = ds.apply_along_axis(_sum_and_mult, 1, x, b=2)
self.assertTrue(x1.shape, (3, 1))
self.assertTrue(x1._reg_shape, (2, 1))
self.assertTrue(_equal_arrays(x1.collect(False),
np.array([[12], [30], [48]])))
self.assertTrue(_validate_array(x1))
x1 = ds.apply_along_axis(_sum_and_mult, 1, x, 1, b=2)
self.assertTrue(x1.shape, (3, 1))
self.assertTrue(x1._reg_shape, (2, 1))
self.assertTrue(_equal_arrays(x1.collect(False),
np.array([[14], [32], [50]])))
self.assertTrue(_validate_array(x1))
def test_apply_sparse(self):
""" Tests apply with sparse data """
x_d = ds.array(np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]]),
block_size=(2, 2))
x_sp = ds.array(csr_matrix([[1, 2, 3], [4, 5, 6], [7, 8, 9]]),
block_size=(2, 2))
x1 = ds.apply_along_axis(_sum_and_mult, 0, x_d)
x2 = ds.apply_along_axis(_sum_and_mult, 0, x_sp)
self.assertTrue(np.array_equal(x1.collect(), x2.collect()))
def fit(self, x):
""" Compute the min and max values for later scaling.
Parameters
----------
x : ds-array, shape=(n_samples, n_features)
Returns
-------
self : MinMaxScaler
"""
self.data_min_ = ds.apply_along_axis(np.min, 0, x)
self.data_max_ = ds.apply_along_axis(np.max, 0, x)
return self
def test_apply_axis(self):
""" Tests apply along axis"""
x = ds.array(np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]]),
block_size=(2, 2))
x1 = ds.apply_along_axis(_sum_and_mult, 0, x)
self.assertTrue(x1.shape, (1, 3))
self.assertTrue(x1._reg_shape, (1, 2))
self.assertTrue(np.array_equal(x1.collect(), np.array([12, 15, 18])))
x1 = ds.apply_along_axis(_sum_and_mult, 1, x)
self.assertTrue(x1.shape, (3, 1))
self.assertTrue(x1._reg_shape, (2, 1))
self.assertTrue(np.array_equal(x1.collect(), np.array([6, 15, 24])))
x1 = ds.apply_along_axis(_sum_and_mult, 1, x, 2)
self.assertTrue(x1.shape, (3, 1))
self.assertTrue(x1._reg_shape, (2, 1))
self.assertTrue(np.array_equal(x1.collect(), np.array([8, 17, 26])))
x1 = ds.apply_along_axis(_sum_and_mult, 1, x, b=2)
self.assertTrue(x1.shape, (3, 1))
self.assertTrue(x1._reg_shape, (2, 1))
self.assertTrue(np.array_equal(x1.collect(), np.array([12, 30, 48])))
x1 = ds.apply_along_axis(_sum_and_mult, 1, x, 1, b=2)
self.assertTrue(x1.shape, (3, 1))
self.assertTrue(x1._reg_shape, (2, 1))
self.assertTrue(np.array_equal(x1.collect(), np.array([14, 32, 50])))
x = ds.array(sp.csr_matrix([[1, 0, -1], [0, 5, 0], [7, 8, 0]]),
block_size=(2, 2))
x1 = ds.apply_along_axis(_sum_and_mult, 0, x, 1, b=2)
self.assertTrue(x1.shape, (1, 3))
self.assertTrue(x1._reg_shape, (1, 2))
self.assertTrue(np.array_equal(x1.collect(), np.array([18, 28, 0])))
x = ds.array(sp.csr_matrix([[1, 0, -1], [0, 5, 0], [7, 8, 0]]),
block_size=(2, 2))
x1 = ds.apply_along_axis(_sum_and_mult, 0, x, 1, b=2)
self.assertTrue(x1.shape, (1, 3))
self.assertTrue(x1._reg_shape, (1, 2))
self.assertTrue((x1.collect() == np.array([18, 28, 0])).all())
def fit(self, x):
""" Compute the mean and std to be used for later scaling.
Parameters
----------
x : ds-array, shape=(n_samples, n_features)
Returns
-------
self : StandardScaler
"""
self.mean_ = ds.apply_along_axis(np.mean, 0, x)
var_blocks = [[]]
for row, m_row in zip(x._iterator(1), self.mean_._iterator(1)):
var_blocks[0].append(_compute_var(row._blocks, m_row._blocks))
self.var_ = Array(var_blocks,
top_left_shape=self.mean_._top_left_shape,
reg_shape=self.mean_._reg_shape,
shape=self.mean_.shape, sparse=x._sparse)
return self