def test_FBPD_Norm1_cvx(self):
if not cvx_not_installable:
opt = {'memopt': True}
# Problem data.
m = 30
n = 20
np.random.seed(1)
Amat = np.random.randn(m, n)
A = LinearOperatorMatrix(Amat)
bmat = np.random.randn(m)
bmat.shape = (bmat.shape[0], 1)
# A = Identity()
# Change n to equal to m.
b = DataContainer(bmat)
# Regularization parameter
lam = 10
opt = {'memopt': True}
# Create object instances with the test data A and b.
f = Norm2sq(A, b, c=0.5, memopt=True)
g0 = ZeroFun()
# Initial guess
x_init = DataContainer(np.zeros((n, 1)))
# Create 1-norm object instance
g1 = Norm1(lam)
# Compare to CVXPY
# Construct the problem.
x1 = Variable(n)
objective1 = Minimize(0.5 * sum_squares(Amat * x1 - bmat.T[0]) +
lam * norm(x1, 1))
prob1 = Problem(objective1)
# The optimal objective is returned by prob.solve().
result1 = prob1.solve(verbose=False, solver=SCS, eps=1e-9)
# The optimal solution for x is stored in x.value and optimal objective value
# is in result as well as in objective.value
print(
"CVXPY least squares plus 1-norm solution and objective value:"
)
print(x1.value)
print(objective1.value)
# Now try another algorithm FBPD for same problem:
x_fbpd1, itfbpd1, timingfbpd1, criterfbpd1 = FBPD(
x_init, Identity(), None, f, g1)
print(x_fbpd1)
print(criterfbpd1[-1])
self.assertNumpyArrayAlmostEqual(numpy.squeeze(x_fbpd1.array),
x1.value, 6)
else:
self.assertTrue(cvx_not_installable)
def test_FISTA_cvx(self):
if not cvx_not_installable:
# Problem data.
m = 30
n = 20
np.random.seed(1)
Amat = np.random.randn(m, n)
A = LinearOperatorMatrix(Amat)
bmat = np.random.randn(m)
bmat.shape = (bmat.shape[0], 1)
# A = Identity()
# Change n to equal to m.
b = DataContainer(bmat)
# Regularization parameter
lam = 10
opt = {'memopt': True}
# Create object instances with the test data A and b.
f = Norm2sq(A, b, c=0.5, memopt=True)
g0 = ZeroFun()
# Initial guess
x_init = DataContainer(np.zeros((n, 1)))
f.grad(x_init)
# Run FISTA for least squares plus zero function.
x_fista0, it0, timing0, criter0 = FISTA(x_init, f, g0, opt=opt)
# Print solution and final objective/criterion value for comparison
print(
"FISTA least squares plus zero function solution and objective value:"
)
print(x_fista0.array)
print(criter0[-1])
# Compare to CVXPY
# Construct the problem.
x0 = Variable(n)
objective0 = Minimize(0.5 * sum_squares(Amat * x0 - bmat.T[0]))
prob0 = Problem(objective0)
# The optimal objective is returned by prob.solve().
result0 = prob0.solve(verbose=False, solver=SCS, eps=1e-9)
# The optimal solution for x is stored in x.value and optimal objective value
# is in result as well as in objective.value
print(
"CVXPY least squares plus zero function solution and objective value:"
)
print(x0.value)
print(objective0.value)
self.assertNumpyArrayAlmostEqual(numpy.squeeze(x_fista0.array),
x0.value, 6)
else:
self.assertTrue(cvx_not_installable)
def create_DataContainer(self, X,Y,Z, value=1):
steps = [timer()]
a = value * numpy.ones((X, Y, Z), dtype='float32')
steps.append(timer())
t0 = dt(steps)
#print("a refcount " , sys.getrefcount(a))
ds = DataContainer(a, False, ['X', 'Y', 'Z'])
return ds
def prox(self,x,tau):
pars = {'algorithm' : ROF_TV, \
'input' : np.asarray(x.as_array(), dtype=np.float32),\
'regularization_parameter':self.lambdaReg*tau, \
'number_of_iterations' :self.iterationsTV ,\
'time_marching_parameter':self.time_marchstep}
out = regularisers.ROF_TV(pars['input'],
pars['regularization_parameter'],
pars['number_of_iterations'],
pars['time_marching_parameter'], self.device)
return DataContainer(out)
def binary_multiply(self):
print("Test binary multiply")
X, Y, Z = 1024, 512, 512
X, Y, Z = 256, 512, 512
steps = [timer()]
a = numpy.ones((X, Y, Z), dtype='float32')
steps.append(timer())
t0 = dt(steps)
#print("a refcount " , sys.getrefcount(a))
ds = DataContainer(a, False, ['X', 'Y', 'Z'])
ds1 = ds.copy()
steps.append(timer())
ds.multiply(ds1, out=ds)
steps.append(timer())
t1 = dt(steps)
print("ds.multiply(ds1, out=ds)", dt(steps))
steps.append(timer())
ds2 = ds.multiply(ds1)
steps.append(timer())
t2 = dt(steps)
print("ds2 = ds.multiply(ds1)", dt(steps))
#self.assertLess(t1, t2)
ds0 = ds
ds0.multiply(2, out=ds0)
steps.append(timer())
print("ds0.multiply(2,out=ds0)", dt(
steps), 2., ds0.as_array()[0][0][0])
self.assertEqual(2., ds0.as_array()[0][0][0])
dt1 = dt(steps)
ds3 = ds0.multiply(2)
steps.append(timer())
print("ds3 = ds0.multiply(2)", dt(steps), 4., ds3.as_array()[0][0][0])
dt2 = dt(steps)
#self.assertLess(dt1, dt2)
self.assertEqual(4., ds3.as_array()[0][0][0])
self.assertEqual(2., ds.as_array()[0][0][0])
ds.multiply(2.5, out=ds0)
self.assertEqual(2.5*2., ds0.as_array()[0][0][0])
def test_creation_copy(self):
shape = (2, 3, 4, 5)
size = shape[0]
for i in range(1, len(shape)):
size = size * shape[i]
#print("a refcount " , sys.getrefcount(a))
a = numpy.asarray([i for i in range(size)])
#print("a refcount " , sys.getrefcount(a))
a = numpy.reshape(a, shape)
#print("a refcount " , sys.getrefcount(a))
ds = DataContainer(a, True, ['X', 'Y', 'Z', 'W'])
#print("a refcount " , sys.getrefcount(a))
self.assertEqual(sys.getrefcount(a), 2)
def test_creation_nocopy(self):
shape = (2, 3, 4, 5)
size = shape[0]
for i in range(1, len(shape)):
size = size * shape[i]
#print("a refcount " , sys.getrefcount(a))
a = numpy.asarray([i for i in range(size)])
#print("a refcount " , sys.getrefcount(a))
a = numpy.reshape(a, shape)
#print("a refcount " , sys.getrefcount(a))
ds = DataContainer(a, False, ['X', 'Y', 'Z', 'W'])
#print("a refcount " , sys.getrefcount(a))
self.assertEqual(sys.getrefcount(a), 3)
self.assertEqual(ds.dimension_labels, {0: 'X', 1: 'Y', 2: 'Z', 3: 'W'})
def binary_subtract(self):
print("Test binary subtract")
X, Y, Z = 512, 512, 512
steps = [timer()]
a = numpy.ones((X, Y, Z), dtype='float32')
steps.append(timer())
t0 = dt(steps)
#print("a refcount " , sys.getrefcount(a))
ds = DataContainer(a, False, ['X', 'Y', 'Z'])
ds1 = ds.copy()
steps.append(timer())
ds.subtract(ds1, out=ds)
steps.append(timer())
t1 = dt(steps)
print("ds.subtract(ds1, out=ds)", dt(steps))
self.assertEqual(0., ds.as_array()[0][0][0])
steps.append(timer())
ds2 = ds.subtract(ds1)
self.assertEqual(-1., ds2.as_array()[0][0][0])
steps.append(timer())
t2 = dt(steps)
print("ds2 = ds.subtract(ds1)", dt(steps))
self.assertLess(t1, t2)
del ds1
ds0 = ds.copy()
steps.append(timer())
ds0.subtract(2, out=ds0)
#ds0.__isub__( 2 )
steps.append(timer())
print("ds0.subtract(2,out=ds0)", dt(
steps), -2., ds0.as_array()[0][0][0])
self.assertEqual(-2., ds0.as_array()[0][0][0])
dt1 = dt(steps)
ds3 = ds0.subtract(2)
steps.append(timer())
print("ds3 = ds0.subtract(2)", dt(steps), 0., ds3.as_array()[0][0][0])
dt2 = dt(steps)
self.assertLess(dt1, dt2)
self.assertEqual(-2., ds0.as_array()[0][0][0])
self.assertEqual(-4., ds3.as_array()[0][0][0])
def test_exp_log(self):
a0 = numpy.asarray([1 for i in range(2*3*4)])
ds0 = DataContainer(numpy.reshape(a0,(2,3,4)), suppress_warning=True)
# ds1 = DataContainer(numpy.reshape(a1,(2,3,4)), suppress_warning=True)
b = ds0.exp().log()
self.assertNumpyArrayEqual(ds0.as_array(), b.as_array())
self.assertEqual(ds0.exp().as_array()[0][0][0], numpy.exp(1))
self.assertEqual(ds0.log().as_array()[0][0][0], 0.)
def prox(self,x,tau):
pars = {'algorithm' : SB_TV, \
'input' : np.asarray(x.as_array(), dtype=np.float32),\
'regularization_parameter':self.lambdaReg*tau, \
'number_of_iterations' :self.iterationsTV ,\
'tolerance_constant':self.tolerance,\
'methodTV': self.methodTV ,\
'printingOut': self.printing}
out = regularisers.SB_TV(pars['input'],
pars['regularization_parameter'],
pars['number_of_iterations'],
pars['tolerance_constant'],
pars['methodTV'],
pars['printingOut'], self.device)
return DataContainer(out)
def binary_divide(self):
print("Test binary divide")
X, Y, Z = 1024, 512, 512
X, Y, Z = 256, 512, 512
steps = [timer()]
a = numpy.ones((X, Y, Z), dtype='float32')
steps.append(timer())
t0 = dt(steps)
#print("a refcount " , sys.getrefcount(a))
ds = DataContainer(a, False, ['X', 'Y', 'Z'])
ds1 = ds.copy()
t1 = 0
t2 = 0
for i in range(10):
steps.append(timer())
ds.divide(ds1, out=ds)
steps.append(timer())
t1 += dt(steps)/10.
print("ds.divide(ds1, out=ds)", dt(steps))
steps.append(timer())
ds2 = ds.divide(ds1)
steps.append(timer())
t2 += dt(steps)/10.
print("ds2 = ds.divide(ds1)", dt(steps))
#self.assertLess(t1, t2)
self.assertEqual(ds.as_array()[0][0][0], 1.)
ds0 = ds
ds0.divide(2, out=ds0)
steps.append(timer())
print("ds0.divide(2,out=ds0)", dt(steps), 0.5, ds0.as_array()[0][0][0])
self.assertEqual(0.5, ds0.as_array()[0][0][0])
dt1 = dt(steps)
ds3 = ds0.divide(2)
steps.append(timer())
print("ds3 = ds0.divide(2)", dt(steps), 0.25, ds3.as_array()[0][0][0])
dt2 = dt(steps)
#self.assertLess(dt1, dt2)
self.assertEqual(.25, ds3.as_array()[0][0][0])
self.assertEqual(.5, ds.as_array()[0][0][0])
def test_dot(self):
a0 = numpy.asarray([i for i in range(2*3*4)])
a1 = numpy.asarray([2*i for i in range(2*3*4)])
ds0 = DataContainer(numpy.reshape(a0,(2,3,4)))
ds1 = DataContainer(numpy.reshape(a1,(2,3,4)))
numpy.testing.assert_equal(ds0.dot(ds1), a0.dot(a1))
a2 = numpy.asarray([2*i for i in range(2*3*5)])
ds2 = DataContainer(numpy.reshape(a2,(2,3,5)))
# it should fail if the shape is wrong
try:
ds2.dot(ds0)
self.assertTrue(False)
except ValueError as ve:
self.assertTrue(True)
def process(self):
data = self.get_input()
#stats, i = add_sample(0)
N = data.get_dimension_size(self.axis)
ax = data.get_dimension_axis(self.axis)
stats = None
i = 0
while i < N:
stats , i = DataStatMoments.add_sample(data, i, self.axis, stats, offset=self.offset)
self.offset += N
labels = ['StatisticalMoments']
labels += [data.dimension_labels[i] \
for i in range(len(data.dimension_labels)) if i != ax]
y = DataContainer( stats[:4] , False,
dimension_labels=labels)
return y
def testGb_creation_nocopy(self):
X, Y, Z = 512, 512, 512
X, Y, Z = 256, 512, 512
steps = [timer()]
a = numpy.ones((X, Y, Z), dtype='float32')
steps.append(timer())
t0 = dt(steps)
print("test clone")
#print("a refcount " , sys.getrefcount(a))
ds = DataContainer(a, False, ['X', 'Y', 'Z'])
#print("a refcount " , sys.getrefcount(a))
self.assertEqual(sys.getrefcount(a), 3)
ds1 = ds.copy()
self.assertNotEqual(aid(ds.as_array()), aid(ds1.as_array()))
ds1 = ds.clone()
self.assertNotEqual(aid(ds.as_array()), aid(ds1.as_array()))
def binary_add(self):
print("Test binary add")
X, Y, Z = 512, 512, 512
#X, Y, Z = 1024, 512, 512
steps = [timer()]
a = numpy.ones((X, Y, Z), dtype='float32')
steps.append(timer())
t0 = dt(steps)
#print("a refcount " , sys.getrefcount(a))
ds = DataContainer(a, False, ['X', 'Y', 'Z'])
ds1 = ds.copy()
out = ds.copy()
steps.append(timer())
ds.add(ds1, out=out)
steps.append(timer())
t1 = dt(steps)
print("ds.add(ds1, out=ds)", dt(steps))
steps.append(timer())
ds2 = ds.add(ds1)
steps.append(timer())
t2 = dt(steps)
print("ds2 = ds.add(ds1)", dt(steps))
#self.assertLess(t1, t2)
self.assertEqual(out.as_array()[0][0][0], 2.)
self.assertNumpyArrayEqual(out.as_array(), ds2.as_array())
ds0 = ds
dt1 = 0
dt2 = 0
for i in range(10):
steps.append(timer())
ds0.add(2, out=out)
steps.append(timer())
print("ds0.add(2,out=out)", dt(steps), 3, ds0.as_array()[0][0][0])
self.assertEqual(3., out.as_array()[0][0][0])
dt1 += dt(steps)/10
steps.append(timer())
ds3 = ds0.add(2)
steps.append(timer())
print("ds3 = ds0.add(2)", dt(steps), 5, ds3.as_array()[0][0][0])
dt2 += dt(steps)/10
self.assertNumpyArrayEqual(out.as_array(), ds3.as_array())
def create_image_data(self):
inputsize = self.size()[1]
return DataContainer(numpy.random.randn(self.volume_geometry.channels,
inputsize[0],
inputsize[1]))
def test_subset(self):
shape = (4, 3, 2)
a = [i for i in range(2*3*4)]
a = numpy.asarray(a)
a = numpy.reshape(a, shape)
ds = DataContainer(a, True, ['X', 'Y', 'Z'])
sub = ds.subset(['X'])
res = True
try:
numpy.testing.assert_array_equal(sub.as_array(),
numpy.asarray([0, 6, 12, 18]))
except AssertionError as err:
res = False
print(err)
self.assertTrue(res)
sub = ds.subset(['X'], Y=2, Z=0)
res = True
try:
numpy.testing.assert_array_equal(sub.as_array(),
numpy.asarray([4, 10, 16, 22]))
except AssertionError as err:
res = False
print(err)
self.assertTrue(res)
sub = ds.subset(['Y'])
try:
numpy.testing.assert_array_equal(
sub.as_array(), numpy.asarray([0, 2, 4]))
res = True
except AssertionError as err:
res = False
print(err)
self.assertTrue(res)
sub = ds.subset(['Z'])
try:
numpy.testing.assert_array_equal(
sub.as_array(), numpy.asarray([0, 1]))
res = True
except AssertionError as err:
res = False
print(err)
self.assertTrue(res)
sub = ds.subset(['Z'], X=1, Y=2)
try:
numpy.testing.assert_array_equal(
sub.as_array(), numpy.asarray([10, 11]))
res = True
except AssertionError as err:
res = False
print(err)
self.assertTrue(res)
print(a)
sub = ds.subset(['X', 'Y'], Z=1)
res = True
try:
numpy.testing.assert_array_equal(sub.as_array(),
numpy.asarray([[1, 3, 5],
[7, 9, 11],
[13, 15, 17],
[19, 21, 23]]))
except AssertionError as err:
res = False
print(err)
self.assertTrue(res)
def test_dot(self):
a0 = numpy.asarray([i for i in range(2*3*4)])
a1 = numpy.asarray([2*i for i in range(2*3*4)])
ds0 = DataContainer(numpy.reshape(a0,(2,3,4)))
ds1 = DataContainer(numpy.reshape(a1,(2,3,4)))
numpy.testing.assert_equal(ds0.dot(ds1), a0.dot(a1))
a2 = numpy.asarray([2*i for i in range(2*3*5)])
ds2 = DataContainer(numpy.reshape(a2,(2,3,5)))
# it should fail if the shape is wrong
try:
ds2.dot(ds0)
self.assertTrue(False)
except ValueError as ve:
self.assertTrue(True)
print ("test dot numpy")
n0 = (ds0 * ds1).sum()
n1 = ds0.as_array().ravel().dot(ds1.as_array().ravel())
self.assertEqual(n0, n1)
def test_unary_operations(self):
print("Test unary operations")
X, Y, Z = 1024, 512, 512
X, Y, Z = 256, 512, 512
steps = [timer()]
a = -numpy.ones((X, Y, Z), dtype='float32')
steps.append(timer())
t0 = dt(steps)
print(t0)
#print("a refcount " , sys.getrefcount(a))
ds = DataContainer(a, False, ['X', 'Y', 'Z'])
ds.sign(out=ds)
steps.append(timer())
print(dt(steps))
self.assertEqual(ds.as_array()[0][0][0], -1.)
ds.abs(out=ds)
steps.append(timer())
print(dt(steps))
self.assertEqual(ds.as_array()[0][0][0], 1.)
ds.__imul__(2)
ds.sqrt(out=ds)
steps.append(timer())
print(dt(steps))
self.assertEqual(ds.as_array()[0][0][0],
numpy.sqrt(2., dtype='float32'))
import numpy as np
import matplotlib.pyplot as plt
# Problem data.
m = 30
n = 20
np.random.seed(1)
Amat = np.random.randn(m, n)
A = LinearOperatorMatrix(Amat)
bmat = np.random.randn(m)
bmat.shape = (bmat.shape[0], 1)
# A = Identity()
# Change n to equal to m.
b = DataContainer(bmat)
# Regularization parameter
lam = 10
# Create object instances with the test data A and b.
f = Norm2sq(A, b, c=0.5, memopt=True)
g0 = ZeroFun()
# Initial guess
x_init = DataContainer(np.zeros((n, 1)))
f.grad(x_init)
opt = {'memopt': True}
# Run FISTA for least squares plus zero function.
x_fista0, it0, timing0, criter0 = FISTA(x_init, f, g0)
ig = ImageGeometry(n, n)
ag = AcquisitionGeometry('parallel', '2D', angles, n, 1)
Aop = AstraProjectorSimple(ig, ag, 'gpu')
# loop to reconstruct energy channels
REC_chan = np.zeros((totChannels, n, n), 'float32')
for i in range(0, totChannels, 1):
sino_channel = sino_all_channels[:,
i, :] # extract a sinogram for i-th channel
print("Initial guess")
x_init = ImageData(geometry=ig)
# Create least squares object instance with projector and data.
print("Create least squares object instance with projector and data.")
f = Norm2sq(Aop, DataContainer(sino_channel), c=0.5)
print("Run FISTA-TV for least squares")
lamtv = 5
opt = {'tol': 1e-4, 'iter': 200}
g_fgp = FGP_TV(lambdaReg=lamtv,
iterationsTV=50,
tolerance=1e-6,
methodTV=0,
nonnegativity=0,
printing=0,
device='gpu')
x_fista_fgp, it1, timing1, criter_fgp = FISTA(x_init, f, g_fgp, opt)
REC_chan[i, :, :] = x_fista_fgp.array
"""
def test_DataProcessorChaining(self):
shape = (2, 3, 4, 5)
size = shape[0]
for i in range(1, len(shape)):
size = size * shape[i]
#print("a refcount " , sys.getrefcount(a))
a = numpy.asarray([i for i in range(size)])
a = numpy.reshape(a, shape)
ds = DataContainer(a, False, ['X', 'Y', 'Z', 'W'])
c = ds.subset(['Z', 'W', 'X'])
arr = c.as_array()
#[ 0 60 1 61 2 62 3 63 4 64 5 65 6 66 7 67 8 68 9 69 10 70 11 71
# 12 72 13 73 14 74 15 75 16 76 17 77 18 78 19 79]
ax = AX()
ax.scalar = 2
ax.set_input(c)
#ax.apply()
print("ax in {0} out {1}".format(
c.as_array().flatten(),
ax.get_output().as_array().flatten()))
numpy.testing.assert_array_equal(ax.get_output().as_array(), arr * 2)
cast = CastDataContainer(dtype=numpy.float32)
cast.set_input(c)
out = cast.get_output()
self.assertTrue(out.as_array().dtype == numpy.float32)
out *= 0
axm = AX()
axm.scalar = 0.5
axm.set_input(c)
axm.get_output(out)
numpy.testing.assert_array_equal(out.as_array(), arr * 0.5)
# check out in DataSetProcessor
#a = numpy.asarray([i for i in range( size )])
# create a PixelByPixelDataProcessor
#define a python function which will take only one input (the pixel value)
pyfunc = lambda x: -x if x > 20 else x
clip = PixelByPixelDataProcessor()
clip.pyfunc = pyfunc
clip.set_input(c)
#clip.apply()
v = clip.get_output().as_array()
self.assertTrue(v.max() == 19)
self.assertTrue(v.min() == -79)
print("clip in {0} out {1}".format(c.as_array(),
clip.get_output().as_array()))
#dsp = DataProcessor()
#dsp.set_input(ds)
#dsp.input = a
# pipeline
chain = AX()
chain.scalar = 0.5
chain.set_input_processor(ax)
print("chain in {0} out {1}".format(ax.get_output().as_array(),
chain.get_output().as_array()))
numpy.testing.assert_array_equal(chain.get_output().as_array(), arr)
def test_FISTA_Norm1_cvx(self):
if not cvx_not_installable:
try:
opt = {'memopt': True}
# Problem data.
m = 30
n = 20
np.random.seed(1)
Amat = np.random.randn(m, n)
A = LinearOperatorMatrix(Amat)
bmat = np.random.randn(m)
bmat.shape = (bmat.shape[0], 1)
# A = Identity()
# Change n to equal to m.
b = DataContainer(bmat)
# Regularization parameter
lam = 10
opt = {'memopt': True}
# Create object instances with the test data A and b.
f = Norm2sq(A, b, c=0.5, memopt=True)
g0 = ZeroFun()
# Initial guess
x_init = DataContainer(np.zeros((n, 1)))
# Create 1-norm object instance
g1 = Norm1(lam)
g1(x_init)
g1.prox(x_init, 0.02)
# Combine with least squares and solve using generic FISTA implementation
x_fista1, it1, timing1, criter1 = FISTA(x_init, f, g1, opt=opt)
# Print for comparison
print(
"FISTA least squares plus 1-norm solution and objective value:"
)
print(x_fista1.as_array().squeeze())
print(criter1[-1])
# Compare to CVXPY
# Construct the problem.
x1 = Variable(n)
objective1 = Minimize(0.5 *
sum_squares(Amat * x1 - bmat.T[0]) +
lam * norm(x1, 1))
prob1 = Problem(objective1)
# The optimal objective is returned by prob.solve().
result1 = prob1.solve(verbose=False, solver=SCS, eps=1e-9)
# The optimal solution for x is stored in x.value and optimal objective value
# is in result as well as in objective.value
print(
"CVXPY least squares plus 1-norm solution and objective value:"
)
print(x1.value)
print(objective1.value)
self.assertNumpyArrayAlmostEqual(numpy.squeeze(x_fista1.array),
x1.value, 6)
except SolverError as se:
print(str(se))
self.assertTrue(True)
else:
self.assertTrue(cvx_not_installable)
def testInlineAlgebra(self):
print("Test Inline Algebra")
X, Y, Z = 1024, 512, 512
X, Y, Z = 256, 512, 512
steps = [timer()]
a = numpy.ones((X, Y, Z), dtype='float32')
steps.append(timer())
t0 = dt(steps)
print(t0)
#print("a refcount " , sys.getrefcount(a))
ds = DataContainer(a, False, ['X', 'Y', 'Z'])
#ds.__iadd__( 2 )
ds += 2
steps.append(timer())
print(dt(steps))
self.assertEqual(ds.as_array()[0][0][0], 3.)
#ds.__isub__( 2 )
ds -= 2
steps.append(timer())
print(dt(steps))
self.assertEqual(ds.as_array()[0][0][0], 1.)
#ds.__imul__( 2 )
ds *= 2
steps.append(timer())
print(dt(steps))
self.assertEqual(ds.as_array()[0][0][0], 2.)
#ds.__idiv__( 2 )
ds /= 2
steps.append(timer())
print(dt(steps))
self.assertEqual(ds.as_array()[0][0][0], 1.)
ds1 = ds.copy()
#ds1.__iadd__( 1 )
ds1 += 1
#ds.__iadd__( ds1 )
ds += ds1
steps.append(timer())
print(dt(steps))
self.assertEqual(ds.as_array()[0][0][0], 3.)
#ds.__isub__( ds1 )
ds -= ds1
steps.append(timer())
print(dt(steps))
self.assertEqual(ds.as_array()[0][0][0], 1.)
#ds.__imul__( ds1 )
ds *= ds1
steps.append(timer())
print(dt(steps))
self.assertEqual(ds.as_array()[0][0][0], 2.)
#ds.__idiv__( ds1 )
ds /= ds1
steps.append(timer())
print(dt(steps))
self.assertEqual(ds.as_array()[0][0][0], 1.)
def create_image_data(self):
inputsize = self.size()[1]
return DataContainer(numpy.random.randn(inputsize[0],
inputsize[1]))
import numpy as np
import matplotlib.pyplot as plt
# Problem data.
m = 30
n = 20
np.random.seed(1)
Amat = np.random.randn(m, n)
A = LinearOperatorMatrix(Amat)
bmat = np.random.randn(m)
bmat.shape = (bmat.shape[0],1)
# A = Identity()
# Change n to equal to m.
b = DataContainer(bmat)
# Regularization parameter
lam = 10
opt = {'memopt':True}
# Create object instances with the test data A and b.
f = Norm2sq(A,b,c=0.5, memopt=True)
g0 = ZeroFun()
# Initial guess
x_init = DataContainer(np.zeros((n,1)))
f.grad(x_init)
# Run FISTA for least squares plus zero function.
x_fista0, it0, timing0, criter0 = FISTA(x_init, f, g0 , opt=opt)
def allocate_adjoint(self):
'''allocate the memory to hold the result of direct'''
#numpy.dot(self.A.transpose(),x.as_array())
M_A, N_A = self.A.shape
out = numpy.zeros((M_A, 1))
return DataContainer(out)
