1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
|
#!/usr/bin/env python
from __future__ import division, print_function, absolute_import
from itertools import product
from functools import reduce
import operator
import numpy as np
from numpy.testing import (assert_allclose, assert_, assert_raises,
assert_equal)
import pywt
def test_traversing_tree_nd():
x = np.array([[1, 2, 3, 4, 5, 6, 7, 8]] * 8, dtype=np.float64)
wp = pywt.WaveletPacketND(data=x, wavelet='db1', mode='symmetric')
assert_(np.all(wp.data == x))
assert_(wp.path == '')
assert_(wp.level == 0)
assert_(wp.maxlevel == 3)
assert_allclose(wp['aa'].data, np.array([[3., 7., 11., 15.]] * 4),
rtol=1e-12)
assert_allclose(wp['da'].data, np.zeros((4, 4)), rtol=1e-12, atol=1e-14)
assert_allclose(wp['ad'].data, -np.ones((4, 4)), rtol=1e-12, atol=1e-14)
assert_allclose(wp['dd'].data, np.zeros((4, 4)), rtol=1e-12, atol=1e-14)
assert_allclose(wp['aa'*2].data, np.array([[10., 26.]] * 2), rtol=1e-12)
# __getitem__ using a tuple access instead
assert_allclose(wp[('aa', 'aa')].data, np.array([[10., 26.]] * 2),
rtol=1e-12)
assert_(wp['aa']['aa'].data is wp['aa'*2].data)
assert_allclose(wp['aa'*3].data, np.array([[36.]]), rtol=1e-12)
assert_raises(IndexError, lambda: wp['aa'*(wp.maxlevel+1)])
assert_raises(ValueError, lambda: wp['f'])
# getitem input must be a string or tuple of strings
assert_raises(TypeError, wp.__getitem__, (5, 3))
assert_raises(TypeError, wp.__getitem__, 5)
def test_accessing_node_attributes_nd():
x = np.array([[1, 2, 3, 4, 5, 6, 7, 8]] * 8, dtype=np.float64)
wp = pywt.WaveletPacketND(data=x, wavelet='db1', mode='symmetric')
assert_allclose(wp['aa'+'ad'].data, np.zeros((2, 2)) - 4, rtol=1e-12)
assert_(wp['aa'+'ad'].path == 'aa'+'ad')
assert_(wp['aa'+'ad'].node_name == 'ad')
assert_(wp['aa'+'ad'].parent.path == 'aa')
assert_allclose(wp['aa'+'ad'].parent.data,
np.array([[3., 7., 11., 15.]] * 4), rtol=1e-12)
# can also index via a tuple instead of concatenated strings
assert_(wp[('aa', 'ad')].level == 2)
assert_(wp[('aa', 'ad')].maxlevel == 3)
assert_(wp[('aa', 'ad')].mode == 'symmetric')
# can access a node's path as either a single string or in tuple form
node = wp[('ad', 'dd')]
assert_(node.path == 'addd')
assert_(node.path_tuple == ('ad', 'dd'))
def test_collecting_nodes_nd():
x = np.array([[1, 2, 3, 4, 5, 6, 7, 8]] * 8, dtype=np.float64)
wp = pywt.WaveletPacketND(data=x, wavelet='db1', mode='symmetric')
assert_(len(wp.get_level(0)) == 1)
assert_(wp.get_level(0)[0].path == '')
# First level
assert_(len(wp.get_level(1)) == 4)
assert_(
[node.path for node in wp.get_level(1)] == ['aa', 'ad', 'da', 'dd'])
# Second and third levels
for lev in [2, 3]:
assert_(len(wp.get_level(lev)) == (2**x.ndim)**lev)
paths = [node.path for node in wp.get_level(lev)]
expected_paths = [
reduce(operator.add, p) for
p in sorted(product(['aa', 'ad', 'da', 'dd'], repeat=lev))]
assert_(paths == expected_paths)
def test_data_reconstruction_delete_nodes_nd():
x = np.array([[1, 2, 3, 4, 5, 6, 7, 8]] * 8, dtype=np.float64)
wp = pywt.WaveletPacketND(data=x, wavelet='db1', mode='symmetric')
# The user must supply either data or axes
assert_raises(ValueError, pywt.WaveletPacketND, data=None, wavelet='db1',
axes=None)
new_wp = pywt.WaveletPacketND(data=None, wavelet='db1', mode='symmetric',
axes=range(x.ndim))
new_wp['ad'+'da'] = wp['ad'+'da'].data
new_wp['ad'*2] = wp['ad'+'da'].data
new_wp['ad'+'dd'] = np.zeros((2, 2), dtype=np.float64)
new_wp['aa'] = [[3.0, 7.0, 11.0, 15.0]] * 4
new_wp['dd'] = np.zeros((4, 4), dtype=np.float64)
new_wp['da'] = wp['da'] # all zeros
assert_allclose(new_wp.reconstruct(update=False),
np.array([[1.5, 1.5, 3.5, 3.5, 5.5, 5.5, 7.5, 7.5]] * 8),
rtol=1e-12)
new_wp['ad'+'aa'] = wp['ad'+'aa'].data
assert_allclose(new_wp.reconstruct(update=False), x, rtol=1e-12)
del(new_wp['ad'+'aa'])
# TypeError on accessing deleted node
assert_raises(TypeError, lambda: new_wp['ad'+'aa'])
new_wp['ad'+'aa'] = wp['ad'+'aa'].data
assert_(new_wp.data is None)
assert_allclose(new_wp.reconstruct(update=True), x, rtol=1e-12)
assert_allclose(new_wp.data, x, rtol=1e-12)
# TODO: decompose=True
def test_wavelet_packet_dtypes():
shape = (16, 8, 8)
for dtype in [np.float32, np.float64, np.complex64, np.complex128]:
x = np.random.randn(*shape).astype(dtype)
if np.iscomplexobj(x):
x = x + 1j*np.random.randn(*shape).astype(x.real.dtype)
wp = pywt.WaveletPacketND(data=x, wavelet='db1', mode='symmetric')
# no unnecessary copy made
assert_(wp.data is x)
# full decomposition
wp.get_level(wp.maxlevel)
# reconstruction from coefficients should preserve dtype
r = wp.reconstruct(False)
assert_equal(r.dtype, x.dtype)
assert_allclose(r, x, atol=1e-6, rtol=1e-6)
def test_wavelet_packet_axes():
rstate = np.random.RandomState(0)
shape = (32, 16, 8)
x = rstate.standard_normal(shape)
for axes in [(0, 1), 1, (-3, -2, -1), (0, 2), (1, )]:
wp = pywt.WaveletPacketND(data=x, wavelet='db1', mode='symmetric',
axes=axes)
# partial decomposition
nodes = wp.get_level(1)
# size along the transformed axes has changed
for ax2 in range(x.ndim):
if ax2 in tuple(np.atleast_1d(axes) % x.ndim):
nodes[0].data.shape[ax2] < x.shape[ax2]
else:
nodes[0].data.shape[ax2] == x.shape[ax2]
# recontsruction from coefficients should preserve dtype
r = wp.reconstruct(False)
assert_equal(r.dtype, x.dtype)
assert_allclose(r, x, atol=1e-12, rtol=1e-12)
# must have non-duplicate axes
assert_raises(ValueError, pywt.WaveletPacketND, data=x, wavelet='db1',
axes=(0, 0))
|
