def test_streaming_process_diagonal(self):
mmp = MultidimensionalMatrixProfileLR()
mmp.initialise(3, 5, 7)
available_data = np.full((3, 5, 7), np.inf)
a = self.dist_matrix[:, 0:5, 0:7]
available_data[diag3(available_data, 0)] = a[diag3(a, 0)]
available_data[diag3(available_data, 4)] = a[diag3(a, 4)]
available_data[diag3(available_data, -3)] = a[diag3(a, -3)]
mmp.process_diagonal(0, a[diag3(a, 0)])
mmp.process_diagonal(4, a[diag3(a, 4)])
mmp.process_diagonal(-3, a[diag3(a, -3)])
correct_mp, correct_index, correct_dims = self.bruteforce_mmp(
available_data)
npt.assert_allclose(mmp.md_matrix_profile(), correct_mp)
npt.assert_equal(mmp.md_profile_index(), correct_index)
for i, dims in enumerate(correct_dims):
sorted_mmp_dims = np.sort(mmp.md_profile_dimensions()[i], axis=0)
npt.assert_equal(sorted_mmp_dims, dims)
mmp.shift_series(3)
mmp.shift_query(2)
available_data = np.full((3, 7, 10), np.inf)
a_win = available_data[:, 0:5, 0:7]
a = self.dist_matrix[:, 0:5, 0:7]
a_win[diag3(a_win, 0)] = a[diag3(a, 0)]
a_win[diag3(a_win, 4)] = a[diag3(a, 4)]
a_win[diag3(a_win, -3)] = a[diag3(a, -3)]
correct_mp, correct_index, correct_dims = self.bruteforce_mmp(
available_data)
npt.assert_allclose(mmp.md_matrix_profile(), correct_mp[:, 3:])
npt.assert_equal(mmp.md_profile_index(), correct_index[:, 3:])
for i, dims in enumerate(correct_dims):
sorted_mmp_dims = np.sort(mmp.md_profile_dimensions()[i], axis=0)
npt.assert_equal(sorted_mmp_dims, dims[:, 3:])
a_win = available_data[:, 2:7, 3:10]
a = self.dist_matrix[:, 2:7, 3:10]
a_win[diag3(a_win, -1)] = a[diag3(a, -1)] # Old diag 0
a_win[diag3(a_win, 3)] = a[diag3(a, 3)] # Old diag 4
a_win[diag3(a_win, -4)] = a[diag3(a, -4)] # Old diag -3
a_win[diag3(a_win, 2)] = a[diag3(a, 2)]
mmp.process_diagonal(-1, a[diag3(a, -1)])
mmp.process_diagonal(3, a[diag3(a, 3)])
mmp.process_diagonal(-4, a[diag3(a, -4)])
mmp.process_diagonal(2, a[diag3(a, 2)])
correct_mp, correct_index, correct_dims = self.bruteforce_mmp(
available_data)
npt.assert_allclose(mmp.md_matrix_profile(), correct_mp[:, 3:])
npt.assert_equal(mmp.md_profile_index(), correct_index[:, 3:])
for i, dims in enumerate(correct_dims):
sorted_mmp_dims = np.sort(mmp.md_profile_dimensions()[i], axis=0)
npt.assert_equal(sorted_mmp_dims, dims[:, 3:])
def test_streaming_process_column(self):
mmp = MultidimensionalMatrixProfileLR()
mmp.initialise(3, 5, 7)
available_data = np.full((3, 5, 7), np.inf)
available_data[:, 0:5, 0] = self.dist_matrix[:, 0:5, 0]
available_data[:, 0:5, 4] = self.dist_matrix[:, 0:5, 4]
available_data[:, 0:5, 5] = self.dist_matrix[:, 0:5, 5]
mmp.process_column(0, self.dist_matrix[:, 0:5, 0])
mmp.process_column(4, self.dist_matrix[:, 0:5, 4])
mmp.process_column(5, self.dist_matrix[:, 0:5, 5])
correct_mp, correct_index, correct_dims = self.bruteforce_mmp(
available_data)
npt.assert_allclose(mmp.md_matrix_profile(), correct_mp)
npt.assert_equal(mmp.md_profile_index(), correct_index)
for i, dims in enumerate(correct_dims):
sorted_mmp_dims = np.sort(mmp.md_profile_dimensions()[i], axis=0)
npt.assert_equal(sorted_mmp_dims, dims)
mmp.shift_series(2)
mmp.shift_query(3)
available_data = np.full((3, 8, 9), np.inf)
available_data[:, 0:5, 4] = self.dist_matrix[:, 0:5, 4]
available_data[:, 0:5, 5] = self.dist_matrix[:, 0:5, 5]
correct_mp, correct_index, correct_dims = self.bruteforce_mmp(
available_data)
npt.assert_allclose(mmp.md_matrix_profile(), correct_mp[:, 2:])
npt.assert_equal(mmp.md_profile_index(), correct_index[:, 2:])
for i, dims in enumerate(correct_dims):
sorted_mmp_dims = np.sort(mmp.md_profile_dimensions()[i], axis=0)
npt.assert_equal(sorted_mmp_dims, dims[:, 2:])
mmp.process_column(0, self.dist_matrix[:, 3:8, 2])
mmp.process_column(2, self.dist_matrix[:, 3:8, 4])
mmp.process_column(3, self.dist_matrix[:, 3:8, 5])
mmp.process_column(5, self.dist_matrix[:, 3:8, 7])
available_data[:, 3:8, 2] = self.dist_matrix[:, 3:8, 2]
available_data[:, 3:8, 4] = self.dist_matrix[:, 3:8, 4]
available_data[:, 3:8, 5] = self.dist_matrix[:, 3:8, 5]
available_data[:, 3:8, 7] = self.dist_matrix[:, 3:8, 7]
correct_mp, correct_index, correct_dims = self.bruteforce_mmp(
available_data)
npt.assert_allclose(mmp.md_matrix_profile(), correct_mp[:, 2:])
npt.assert_equal(mmp.md_profile_index(), correct_index[:, 2:])
for i, dims in enumerate(correct_dims):
sorted_mmp_dims = np.sort(mmp.md_profile_dimensions()[i], axis=0)
npt.assert_equal(sorted_mmp_dims, dims[:, 2:])