def test_rectangular(self):
lons = numpy.array(range(100)).reshape((10, 10))
lats = numpy.negative(lons)
mesh = RectangularMesh(lons, lats, depths=None)
bounding_mesh = mesh._get_bounding_mesh()
expected_lons = numpy.array([
0, 1, 2, 3, 4, 5, 6, 7, 8, 9,
19, 29, 39, 49, 59, 69, 79, 89,
99, 98, 97, 96, 95, 94, 93, 92, 91,
90, 80, 70, 60, 50, 40, 30, 20, 10
])
expected_lats = numpy.negative(expected_lons)
self.assertTrue((bounding_mesh.lons == expected_lons).all())
self.assertTrue((bounding_mesh.lats == expected_lats).all())
self.assertIsNone(bounding_mesh.depths)
depths = lons + 10
mesh = RectangularMesh(lons, lats, depths)
expected_depths = expected_lons + 10
bounding_mesh = mesh._get_bounding_mesh()
self.assertIsNotNone(bounding_mesh.depths)
self.assertTrue((bounding_mesh.depths
== expected_depths.flatten()).all())
bounding_mesh = mesh._get_bounding_mesh(with_depths=False)
self.assertIsNone(bounding_mesh.depths)
def unmask_temperature(self, signal, order='nested', seed=None):
"""
Given the harmonic sphere map ``signal`` as the underlying signal,
provide a map where the mask has been removed and replaced with the
contents of signal. Noise consistent with the noise properties
of the observation (without the mask) will be added.
"""
Nside, lmin, lmax = self.Nside, signal.lmin, signal.lmax
random_state = as_random_state(seed)
temperature = self.load_temperature_mutable(order)
inverse_mask = (self.properties.load_mask_mutable(order) == 1).view(np.ndarray)
np.negative(inverse_mask, inverse_mask) # invert the mask in-place
# First, smooth the signal with the beam and pixel window
smoothed_signal = self.properties.load_beam_transfer_matrix(lmin, lmax) * signal
pixwin = load_temperature_pixel_window_matrix(Nside, lmin, lmax)
smoothed_signal = pixwin * smoothed_signal
# Create map from signal, and replace unmasked values in temperature map
signal_map = smoothed_signal.to_pixel(self.Nside)
signal_map.change_order_inplace(order)
temperature[inverse_mask] = signal_map[inverse_mask]
# Finally, add RMS to unmasked area
rms_in_mask = self.properties.load_rms(order)[inverse_mask]
temperature[inverse_mask] += random_state.normal(scale=rms_in_mask)
return temperature
def test_ufunc_out(self):
from numpy import array, negative, zeros, sin
from math import sin as msin
a = array([[1, 2], [3, 4]])
c = zeros((2,2,2))
b = negative(a + a, out=c[1])
#test for view, and also test that forcing out also forces b
assert (c[:, :, 1] == [[0, 0], [-4, -8]]).all()
assert (b == [[-2, -4], [-6, -8]]).all()
#Test broadcast, type promotion
b = negative(3, out=a)
assert (a == -3).all()
c = zeros((2, 2), dtype=float)
b = negative(3, out=c)
assert b.dtype.kind == c.dtype.kind
assert b.shape == c.shape
a = array([1, 2])
b = sin(a, out=c)
assert(c == [[msin(1), msin(2)]] * 2).all()
b = sin(a, out=c+c)
assert (c == b).all()
#Test shape agreement
a = zeros((3,4))
b = zeros((3,5))
raises(ValueError, 'negative(a, out=b)')
b = zeros((1,4))
raises(ValueError, 'negative(a, out=b)')
def logpdf(x, nu, s2=1):
"""Log of the scaled inverse chi-squared probability density function.
Parameters
----------
x : array_like
quantiles
nu : array_like
degrees of freedom
s2 : array_like, optional
scale (default 1)
Returns
-------
logpdf : ndarray
Log of the probability density function evaluated at `x`.
"""
x = np.asarray(x)
nu = np.asarray(nu)
s2 = np.asarray(s2)
nu_2 = nu/2
y = np.log(x)
y *= (nu_2 +1)
np.negative(y, out=y)
y -= (nu_2*s2)/x
y += np.log(s2)*nu_2
y -= gammaln(nu_2)
y += np.log(nu_2)*nu_2
return y
def forward_cpu(self, inputs):
self.retain_inputs((0, 1))
x, gy = inputs
gx = utils.force_array(numpy.sin(x))
numpy.negative(gx, out=gx)
gx *= gy
return gx,
def testInitializerFunction(self):
value = [[-42], [133.7]]
shape = [2, 1]
with self.test_session():
initializer = lambda: tf.constant(value)
with self.assertRaises(ValueError):
# Checks that dtype must be specified.
tf.Variable(initializer)
v1 = tf.Variable(initializer, dtype=tf.float32)
self.assertEqual(shape, v1.get_shape())
self.assertAllClose(value, v1.initial_value.eval())
with self.assertRaises(tf.errors.FailedPreconditionError):
v1.eval()
v2 = tf.Variable(tf.neg(v1.initialized_value()), dtype=tf.float32)
self.assertEqual(v1.get_shape(), v2.get_shape())
self.assertAllClose(np.negative(value), v2.initial_value.eval())
# Once v2.initial_value.eval() has been called, v1 has effectively been
# initialized.
self.assertAllClose(value, v1.eval())
with self.assertRaises(tf.errors.FailedPreconditionError):
v2.eval()
tf.initialize_all_variables().run()
self.assertAllClose(np.negative(value), v2.eval())
def testInitializerFunction(self):
value = [[-42], [133.7]]
shape = [2, 1]
with self.test_session():
initializer = lambda: constant_op.constant(value)
v1 = variables.Variable(initializer, dtype=dtypes.float32)
self.assertEqual(shape, v1.get_shape())
self.assertEqual(shape, v1.shape)
self.assertAllClose(value, v1.initial_value.eval())
with self.assertRaises(errors_impl.FailedPreconditionError):
v1.eval()
v2 = variables.Variable(
math_ops.negative(v1.initialized_value()), dtype=dtypes.float32)
self.assertEqual(v1.get_shape(), v2.get_shape())
self.assertEqual(v1.shape, v2.shape)
self.assertAllClose(np.negative(value), v2.initial_value.eval())
# Once v2.initial_value.eval() has been called, v1 has effectively been
# initialized.
self.assertAllClose(value, v1.eval())
with self.assertRaises(errors_impl.FailedPreconditionError):
v2.eval()
variables.global_variables_initializer().run()
self.assertAllClose(np.negative(value), v2.eval())
def test_negative(self):
from numpy import array, negative
a = array([-5.0, 0.0, 1.0])
b = negative(a)
for i in range(3):
assert b[i] == -a[i]
a = array([-5.0, 1.0])
b = negative(a)
a[0] = 5.0
assert b[0] == 5.0
a = array(range(30))
assert negative(a + a)[3] == -6
a = array([[1, 2], [3, 4]])
b = negative(a + a)
assert (b == [[-2, -4], [-6, -8]]).all()
class Obj(object):
def __neg__(self):
return "neg"
x = Obj()
assert type(negative(x)) is str
def neg(target):
a = pyext.Buffer(target)
# in place transformation (see Python array ufuncs)
N.negative(a[:],a[:])
# must mark buffer content as dirty to update graph
# (no explicit assignment occurred)
a.dirty()
def construct_uvn_frame(n, u, b=None, flip_to_match_image=True):
""" Returns an orthonormal 3x3 frame from a normal and one in-plane vector """
n = normalized(n)
u = normalized(np.array(u) - np.dot(n, u) * n)
v = normalized_cross(n, u)
# flip to match image orientation
if flip_to_match_image:
if abs(u[1]) > abs(v[1]):
u, v = v, u
if u[0] < 0:
u = np.negative(u)
if v[1] < 0:
v = np.negative(v)
if b is None:
if n[2] < 0:
n = np.negative(n)
else:
if np.dot(n, b) > 0:
n = np.negative(n)
# return uvn matrix, column major
return np.matrix([
[u[0], v[0], n[0]],
[u[1], v[1], n[1]],
[u[2], v[2], n[2]],
])
def negateVal():
"""negate a boolean, change the sign of a float inplace"""
Z=np.random.randint(0,2,100)
np.logical_not(Z,out=Z)
print Z
W=np.random.uniform(-1.0,1.0,100)
np.negative(Z,out=Z)
print Z
def backward_cpu(self, x, gy):
gx = utils.force_array(numpy.square(x[0]))
numpy.negative(gx, out=gx)
gx += 1
numpy.sqrt(gx, out=gx)
numpy.reciprocal(gx, out=gx)
gx *= gy[0]
return gx,
def _quaternion_from_matrix(matrix, isprecise=False):
"""Summary
Args:
matrix (TYPE): Description
isprecise (bool, optional): Description
Returns:
TYPE: Description
"""
M = np.array(matrix, dtype=np.float64, copy=False)[:4, :4]
if isprecise:
q = np.empty((4, ))
t = np.trace(M)
if t > M[3, 3]:
q[0] = t
q[3] = M[1, 0] - M[0, 1]
q[2] = M[0, 2] - M[2, 0]
q[1] = M[2, 1] - M[1, 2]
else:
i, j, k = 1, 2, 3
if M[1, 1] > M[0, 0]:
i, j, k = 2, 3, 1
if M[2, 2] > M[i, i]:
i, j, k = 3, 1, 2
t = M[i, i] - (M[j, j] + M[k, k]) + M[3, 3]
q[i] = t
q[j] = M[i, j] + M[j, i]
q[k] = M[k, i] + M[i, k]
q[3] = M[k, j] - M[j, k]
q *= 0.5 / math.sqrt(t * M[3, 3])
# NEED MORMALIZE
else:
m00 = M[0, 0]
m01 = M[0, 1]
m02 = M[0, 2]
m10 = M[1, 0]
m11 = M[1, 1]
m12 = M[1, 2]
m20 = M[2, 0]
m21 = M[2, 1]
m22 = M[2, 2]
# symmetric matrix K
K = np.array([[m00-m11-m22, 0.0, 0.0, 0.0],
[m01+m10, m11-m00-m22, 0.0, 0.0],
[m02+m20, m12+m21, m22-m00-m11, 0.0],
[m21-m12, m02-m20, m10-m01, m00+m11+m22]])
K /= 3.0
# quaternion is eigenvector of K that corresponds to largest eigenvalue
w, V = np.linalg.eigh(K)
q = V[[3, 0, 1, 2], np.argmax(w)]
if q[0] < 0.0:
np.negative(q, q)
n = np.linalg.norm(q)
if n > 1.0:
q = q/n
# taken from ransformations.py so w first
return (q[0], q[1:])
def backProp_epoch(I,T,W_IH,W_HO,A_O,
DeltaW_IH,DeltaW_HO,
net_H=None,net_O=None,A_H=None,
Delta_H=None,Delta_O=None,
sigma_H=(afs.sigmoid,afs.sigmoid_prime),
sigma_O=(afs.sigmoid,afs.sigmoid_prime),
errorF=(efs.sumSquaredError,efs.sumSquaredError_prime)):
"""
net: node input function result
A: node activation
sigma: activation function
errorF: ( error function(target,output),error function derivative(target,output) )
"""
(M_H,M_I) = W_IH.shape; M_I-=1;
(M_O,blah) = W_HO.shape;
(M,N) = I.shape
if net_H == None:
net_H = np.empty((M_H,1))
if net_O == None:
net_O = np.empty((M_O,1))
if A_H == None:
A_H = np.empty_like(net_H)
if Delta_H == None:
Delta_H = np.empty_like(net_H)
if Delta_O == None:
Delta_O = np.empty_like(net_O)
# compute hidden layer inputs
np.dot(W_IH[:,:-1],I,net_H) # net_H is M_H x N
np.add(net_H,np.dot(W_IH[:,-1:],np.ones((1,N))),net_H) # bias
# compute hidden layer activations
sigma_H[0](net_H,A_H) # A_H is M_H x N
# compute output layer inputs
np.dot(W_HO[:,:-1],A_H,net_O)
np.add(net_O,np.dot(W_HO[:,-1:],np.ones((1,N))),net_O) # bias
# compute output layer activations
sigma_O[0](net_O,A_O)
# compute output error
errorVal = errorF[0](A_O,T)
# compute output error gradient
errorF[1](T,A_O,Delta_O) # Delta_O holds tmp value
np.negative(Delta_O,Delta_O)
sigma_O[1](A_O,net_O) # reusing net_O matrix as tmp storage
np.multiply(Delta_O,net_O,Delta_O)
# compute output weight update
tmpA_H = np.append(A_H,np.ones((1,N)),axis=0) # add bias inputs
np.dot(Delta_O,tmpA_H.T,DeltaW_HO) # TODO: compute using tanspose for speed-up?
# compute hidden error gradient
sigma_H[1](A_H,Delta_H) # Delta_H holds tmp value
np.multiply(Delta_H,
np.dot(W_HO[:,:-1].T,Delta_O), # TODO: W^T*Delta_O too wasteful
Delta_H)
# compute hidden weight update
tmpI = np.append(I,np.ones((1,N)),axis=0) # add bias inputs
np.dot(Delta_H,tmpI.T,DeltaW_IH) # TODO: compute using transpose for speed-up?
# np.multiply(DeltaW_IH,alpha,DeltaW_IH) # apply learning rate
# TODO: force garbage collection at key areas where temporaries are created
return errorVal
def __neg__(self):
"""
return negated
"""
self.A = numpy.negative(self.A)
self.bX = numpy.negative(self.bX)
self.bY = numpy.negative(self.bY)
self.bZ = numpy.negative(self.bZ)
return self
def generate_unit_phase_shifts(shape, float_type=float):
"""
Computes the complex phase shift's angle due to a unit spatial shift.
This is meant to be a helper function for ``register_mean_offsets``. It
does this by computing a table of the angle of the phase of a unit
shift in each dimension (with a factor of :math:`2\pi`).
This allows arbitrary phase shifts to be made in each dimensions by
multiplying these angles by the size of the shift and added to the
existing angle to induce the proper phase shift in fourier space, which
is equivalent to the spatial translation.
Args:
shape(tuple of ints): shape of the data to be shifted.
float_type(real type): phase type (default numpy.float64)
Returns:
(numpy.ndarray): an array containing the angle of the
complex phase shift to use for each
dimension.
Examples:
>>> generate_unit_phase_shifts((2,4))
array([[[-0. , -0. , -0. , -0. ],
[-3.14159265, -3.14159265, -3.14159265, -3.14159265]],
<BLANKLINE>
[[-0. , -1.57079633, -3.14159265, -4.71238898],
[-0. , -1.57079633, -3.14159265, -4.71238898]]])
"""
# Convert to `numpy`-based type if not done already.
float_type = numpy.dtype(float_type).type
# Must be of type float.
assert issubclass(float_type, numpy.floating)
assert numpy.dtype(float_type).itemsize >= 4
# Get the negative wave vector
negative_wave_vector = numpy.asarray(shape, dtype=float_type)
numpy.reciprocal(negative_wave_vector, out=negative_wave_vector)
negative_wave_vector *= 2*numpy.pi
numpy.negative(negative_wave_vector, out=negative_wave_vector)
# Get the indices for each point in the selected space.
indices = xnumpy.cartesian_product([numpy.arange(_) for _ in shape])
# Determine the phase offset for each point in space.
complex_angle_unit_shift = indices * negative_wave_vector
complex_angle_unit_shift = complex_angle_unit_shift.T.copy()
complex_angle_unit_shift = complex_angle_unit_shift.reshape(
(len(shape),) + shape
)
return(complex_angle_unit_shift)
def getMountainWeights(bgDiff, mountainCenter, halfLife = 0.1):
mountainWeights = bgDiff - mountainCenter
k = np.log(2) / halfLife
mountainWeights *= k
np.abs(mountainWeights, out = mountainWeights)
np.negative(mountainWeights, out = mountainWeights)
np.exp(mountainWeights, out = mountainWeights)
return mountainWeights
def forward_cpu(self, inputs):
self.retain_inputs((0, 1))
x, gy = inputs
gx = utils.force_array(numpy.square(x))
numpy.negative(gx, out=gx)
gx += 1
numpy.sqrt(gx, out=gx)
numpy.reciprocal(gx, out=gx)
gx *= gy
return gx,
def __neg__(self):
"""
return negated
"""
if __sparse__:
self.A = -self.A
else:
self.A = numpy.negative(self.A)
self.b = numpy.negative(self.b)
return self
def test_lower_align(self):
# check data that is not aligned to element size
# i.e doubles are aligned to 4 bytes on i386
d = np.zeros(23 * 8, dtype=np.int8)[4:-4].view(np.float64)
assert_equal(np.abs(d), d)
assert_equal(np.negative(d), -d)
np.negative(d, out=d)
np.negative(np.ones_like(d), out=d)
np.abs(d, out=d)
np.abs(np.ones_like(d), out=d)
def decompose(self):
M = numpy.array(self._data, dtype = numpy.float64, copy = True).T
if abs(M[3, 3]) < self._EPS:
raise ValueError("M[3, 3] is zero")
M /= M[3, 3]
P = M.copy()
P[:, 3] = 0.0, 0.0, 0.0, 1.0
if not numpy.linalg.det(P):
raise ValueError("matrix is singular")
scale = numpy.zeros((3, ))
shear = [0.0, 0.0, 0.0]
angles = [0.0, 0.0, 0.0]
mirror = [1, 1, 1]
translate = M[3, :3].copy()
M[3, :3] = 0.0
row = M[:3, :3].copy()
scale[0] = math.sqrt(numpy.dot(row[0], row[0]))
row[0] /= scale[0]
shear[0] = numpy.dot(row[0], row[1])
row[1] -= row[0] * shear[0]
scale[1] = math.sqrt(numpy.dot(row[1], row[1]))
row[1] /= scale[1]
shear[0] /= scale[1]
shear[1] = numpy.dot(row[0], row[2])
row[2] -= row[0] * shear[1]
shear[2] = numpy.dot(row[1], row[2])
row[2] -= row[1] * shear[2]
scale[2] = math.sqrt(numpy.dot(row[2], row[2]))
row[2] /= scale[2]
shear[1:] /= scale[2]
if numpy.dot(row[0], numpy.cross(row[1], row[2])) < 0:
numpy.negative(scale, scale)
numpy.negative(row, row)
# If the scale was negative, we give back a seperate mirror vector to indicate this.
if M[0, 0] < 0:
mirror[0] = -1
if M[1, 1] < 0:
mirror[1] = -1
if M[2, 2] < 0:
mirror[2] = -1
angles[1] = math.asin(-row[0, 2])
if math.cos(angles[1]):
angles[0] = math.atan2(row[1, 2], row[2, 2])
angles[2] = math.atan2(row[0, 1], row[0, 0])
else:
angles[0] = math.atan2(-row[2, 1], row[1, 1])
angles[2] = 0.0
return Vector(data = scale), Vector(data = shear), Vector(data = angles), Vector(data = translate), Vector(data = mirror)
def computeDeltaW(tmpDeltaW,A_O,y,DeltaW,errorF):
# init values and allocate memory
M_O = A_O.shape[0] # num vector elements
errors = np.zeros_like(y)
# compute errors
errorF[1](y,A_O,errors)
np.negative(errors,errors)
# compute Delta_W
for i in range(M_O):
tmpDeltaW[i,:,:] *= errors[i]
np.add(DeltaW,tmpDeltaW[i,:,:],DeltaW)
def __neg__(self):
"""Operator for inverting the phase of the spectrogram with ``-spectrogram``.
:returns: a :class:`sumpf.Spectrogram` instance
"""
channels = sumpf_internal.allocate_array(shape=self.shape(), dtype=numpy.complex128)
numpy.negative(self._channels, out=channels)
return Spectrogram(channels=channels,
resolution=self.__resolution,
sampling_rate=self.__sampling_rate,
offset=self.__offset,
labels=self._labels)
def quaternion_conjugate(quaternion):
"""Return conjugate of quaternion.
>>> q0 = random_quaternion()
>>> q1 = quaternion_conjugate(q0)
>>> q1[0] == q0[0] and all(q1[1:] == -q0[1:])
True
"""
q = numpy.array(quaternion, dtype=numpy.float64, copy=True)
numpy.negative(q[1:], q[1:])
return q
def quaternion_inverse(quaternion):
"""Return inverse of quaternion.
>>> q0 = random_quaternion()
>>> q1 = quaternion_inverse(q0)
>>> numpy.allclose(quaternion_multiply(q0, q1), [1, 0, 0, 0])
True
"""
q = numpy.array(quaternion, dtype=numpy.float64, copy=True)
numpy.negative(q[1:], q[1:])
return q / numpy.dot(q, q)
def test_negative(self):
from numpy import array, negative
a = array([-5.0, 0.0, 1.0])
b = negative(a)
for i in range(3):
assert b[i] == -a[i]
a = array([-5.0, 1.0])
b = negative(a)
a[0] = 5.0
assert b[0] == 5.0
def plot_hists(nus=[143,353],
map1_name=None,
map2_name=None,
maskname='wmap_temperature_kq85_analysis_mask_r10_9yr_v5.fits',
nside=2048,
fwhm=0.0,
bins=100,normed=True,
atol=1e-6, ymin=0.01, ymax=None,
xmin=-0.001, xmax=0.005):
if map1_name is None:
map1_name = 'HFI_SkyMap_{}_2048_R2.02_full.fits'.format(nus[0])
label1 = '{} GHz'.format(nus[0])
if map2_name is None:
map2_name = 'HFI_SkyMap_{}_2048_R2.02_full.fits'.format(nus[1])
label2 = '{} GHz'.format(nus[1])
map1 = prepare_map( map1_name, field=0,
maskname=maskname,
nside_out=nside, fwhm=fwhm )
map2 = prepare_map( map2_name, field=0,
maskname=maskname,
nside_out=nside, fwhm=fwhm )
y1,x1 = pl.histogram(map1[np.where(np.negative(np.isclose(map1,0.,atol=atol)))],
bins=bins,normed=normed)
bin1 = (x1[:-1] + x1[1:]) / 2.
y2,x2 = pl.histogram(map2[np.where(np.negative(np.isclose(map2,0.,atol=atol)))],
bins=bins,normed=normed)
bin2 = (x2[:-1] + x2[1:]) / 2.
#return bin1,y1,bin2,y2
fig = plt.figure()
ax = plt.gca()
ax.semilogy(bin1, y1, lw=3, label=label1,color='red')
ax.semilogy(bin2, y2, lw=3, label=label2,color='gray')
ax.set_xlim(xmin=xmin,xmax=xmax)
ax.set_ylim(ymin=ymin, ymax=ymax)
#ax.set_yscale('log')
ax.set_xlabel('$\mu K$', fontsize=20)
ax.set_yticks([])
plt.draw()
plt.legend(frameon=False, fontsize=20)
plt.savefig('pdfs_{}GHz_{}GHz_fwhm{:.3}rad.pdf'.format(nus[0],nus[1],fwhm))
def from_matrix(matrix, isprecise=False):
"""Return quaternion from rotation matrix.
If isprecise is True, the input matrix is assumed to be a precise rotation
matrix and a faster algorithm is used.
"""
M = numpy.array(matrix, dtype=numpy.float64, copy=False)[:4, :4]
if isprecise:
q = numpy.empty((4, ))
t = numpy.trace(M)
if t > M[3, 3]:
q[0] = t
q[3] = M[1, 0] - M[0, 1]
q[2] = M[0, 2] - M[2, 0]
q[1] = M[2, 1] - M[1, 2]
else:
i, j, k = 1, 2, 3
if M[1, 1] > M[0, 0]:
i, j, k = 2, 3, 1
if M[2, 2] > M[i, i]:
i, j, k = 3, 1, 2
t = M[i, i] - (M[j, j] + M[k, k]) + M[3, 3]
q[i] = t
q[j] = M[i, j] + M[j, i]
q[k] = M[k, i] + M[i, k]
q[3] = M[k, j] - M[j, k]
q *= 0.5 / math.sqrt(t * M[3, 3])
else:
m00 = M[0, 0]
m01 = M[0, 1]
m02 = M[0, 2]
m10 = M[1, 0]
m11 = M[1, 1]
m12 = M[1, 2]
m20 = M[2, 0]
m21 = M[2, 1]
m22 = M[2, 2]
# symmetric matrix K
K = numpy.array([[m00-m11-m22, 0.0, 0.0, 0.0],
[m01+m10, m11-m00-m22, 0.0, 0.0],
[m02+m20, m12+m21, m22-m00-m11, 0.0],
[m21-m12, m02-m20, m10-m01, m00+m11+m22]])
K /= 3.0
# quaternion is eigenvector of K that corresponds to largest eigenvalue
w, V = numpy.linalg.eigh(K)
q = V[[3, 0, 1, 2], numpy.argmax(w)]
if q[0] < 0.0:
numpy.negative(q, q)
#return q.tolist()
return [ q[1], q[2], q[3], q[0] ] # sofa order
def track_unavoided_crossings(overlaps: Tensor3D, nHOMO: int) -> Tuple:
"""
Track the index of the states if there is a crossing using the
algorithm described at:
J. Chem. Phys. 137, 014512 (2012); doi: 10.1063/1.4732536.
"""
# 3D array containing the costs
# Notice that the cost is compute on half of the overlap matrices
# correspoding to Sji_t, the other half corresponds to Sij_t
nOverlaps, nOrbitals, _ = overlaps.shape
# Indexes taking into account the crossing
# There are 2 Overlap matrices at each time t
indexes = np.empty((nOverlaps + 1, nOrbitals), dtype=np.int)
indexes[0] = np.arange(nOrbitals, dtype=np.int)
# Track the crossing using the overlap matrices
for k in range(nOverlaps):
# Cost matrix to track the corssings
logger.info("Tracking crossings at time: {}".format(k))
cost_mtx_homos = np.negative(overlaps[k, :nHOMO, :nHOMO] ** 2)
cost_mtx_lumos = np.negative(overlaps[k, nHOMO:, nHOMO:] ** 2)
# Compute the swap at time t + dt using two set of Orbitals:
# HOMOs and LUMOS
swaps_homos = linear_sum_assignment(cost_mtx_homos)[1]
swaps_lumos = linear_sum_assignment(cost_mtx_lumos)[1]
total_swaps = np.concatenate((swaps_homos, swaps_lumos + nHOMO))
indexes[k + 1] = total_swaps
# update the overlaps at times > t with the previous swaps
if k != (nOverlaps - 1): # last element
k1 = k + 1
# Update the matrix Sji at time t
overlaps[k] = swap_columns(overlaps[k], total_swaps)
# Update all the matrices Sji at time > t
overlaps[k1:] = swap_forward(overlaps[k1:], total_swaps)
# Accumulate the swaps
acc = indexes[0]
arr = np.empty(indexes.shape, dtype=np.int)
arr[0] = acc
# Fold accumulating the crossings
for i in range(nOverlaps):
acc = acc[indexes[i + 1]]
arr[i + 1] = acc
return overlaps, arr
def exceedance_graph(self, ax):
ax.set_title('Wave Height Distribution (m)')
lims = [0,1]
for x in range(0,self.n):
yh = [self.y[x], self.y[x]]
ax.plot(self.xh,yh,color='grey', alpha=0.5)
ax.text(self.x[0]-2*self.dx,self.y[x],self.poe[x])
lims[1] = self.y[x]
for x in range(0,self.m):
xv = [self.x[x],self.x[x]]
ax.plot(xv,self.yv,color='grey', alpha=0.5)
ax.text(self.x[x] - self.dx/4,self.y[0]-self.dy,self.x[x])
lims[0] = self.x[x]
zi = np.sqrt(np.negative(np.log(self.yi)))
ax.plot(self.xi,zi,'o',label='Observed Waves')
coef = np.polyfit(self.xi,zi,1)
xa = list([0.5 * np.min(self.xi)])
xb = list([1.5 * np.max(self.xi)])
xx = np.concatenate((np.array(xa),self.xi,np.array(xb)))
yy = np.polyval(coef, xx)
ax.plot(xx,yy, label='Rayleigh Fit')
H1 = self.get_bg_value(self.Htr/self.Hrms, 1)
H2 = self.get_bg_value(self.Htr/self.Hrms, 2)
xxx = self.array_utility(self.x[0],self.x[self.m - 1])
lxxx = len(xxx)
yyy = list()
for x in range(0,lxxx):
if xxx[x] < self.Htr:
yyy.append(1 - np.exp(np.negative(np.power(xxx[x]/(H1*self.Hrms),2.0))))
else:
yyy.append(1 - np.exp(np.negative(np.power(xxx[x]/(H2*self.Hrms),3.6))))
zzz = np.sqrt(np.negative(np.log(np.subtract(1,yyy))))
ax.plot(xxx,zzz,'r',label='BG Fit')
ax.legend()
plt.setp(ax.get_yticklabels(), visible=False)
plt.setp(ax.get_yticklines(),visible=False)
plt.setp(ax.get_xticklabels(), visible=False)
plt.setp(ax.get_xticklines(),visible=False)
plt.ylim(0,lims[1])
plt.xlim(0,lims[0])
def negIP(self):
"""Take the negation of F: F.negIP() => F(x) <- (-F(x)) (in-place)"""
np.negative(self.t, out=self.t)
return self
def strassen_matrix_multiplication(A, B):
n = len(A)
if n == 1:
return [[A[0][0] * B[0][0]]]
l = math.floor(n / 2)
Ak = array(A)
Bk = array(B)
#for i in range(l):
A11 = Ak[0:l, 0:l]
A12 = Ak[0:l, l:n]
A21 = Ak[l:n, 0:l]
A22 = Ak[l:n, l:n]
B11 = Bk[0:l, 0:l]
B12 = Bk[0:l, l:n]
B21 = Bk[l:n, 0:l]
B22 = Bk[l:n, l:n]
#print()
#print("A11",A11)
#print()
#print("A12",A12.tolist())
#print()
#print("A21",A21.tolist())
#print()
#print("A22",A22.tolist())
#print()
K1 = []
K2 = []
K3 = []
K4 = []
K5 = []
K6 = []
K7 = []
K8 = []
K9 = []
K10 = []
#M1= (A11+A22) (B11+B22)
#M2= (A21+A22)B11
#M3=A11(B12−B22)
#M4=A22(B21−B11)
#M5= (A11+A12)B22
#M6= (A21−A11)(B11+B12)
#M7= (A12−A22)(B21+B22)
K1 = numpy.zeros(shape=(l, l), dtype=int).tolist()
#print("K1",K1)
K1 = A11 + A22
K2 = numpy.zeros(shape=(l, l), dtype=int).tolist()
K2 = B11 + B22
K3 = numpy.zeros(shape=(l, l), dtype=int).tolist()
K3 = A21 + A22
K4 = numpy.zeros(shape=(l, l), dtype=int).tolist()
K4 = B12 - B22
K5 = numpy.zeros(shape=(l, l), dtype=int).tolist()
K5 = B21 - B11
K6 = numpy.zeros(shape=(l, l), dtype=int).tolist()
K6 = A11 + A12
K7 = numpy.zeros(shape=(l, l), dtype=int).tolist()
K7 = A21 - A11
K8 = numpy.zeros(shape=(l, l), dtype=int).tolist()
K8 = B11 + B12
K9 = numpy.zeros(shape=(l, l), dtype=int).tolist()
K9 = A12 - A22
K10 = numpy.zeros(shape=(l, l), dtype=int).tolist()
K10 = B21 + B22
P1 = numpy.zeros(shape=(l, l), dtype=int)
P1 = numpy.array(strassen_matrix_multiplication(K1, K2))
P2 = numpy.zeros(shape=(l, l), dtype=int)
P2 = numpy.array(strassen_matrix_multiplication(K3, B11))
P3 = numpy.zeros(shape=(l, l), dtype=int)
P3 = numpy.array(strassen_matrix_multiplication(A11, K4))
P4 = numpy.zeros(shape=(l, l), dtype=int)
P4 = numpy.array(strassen_matrix_multiplication(A22, K5))
P5 = numpy.zeros(shape=(l, l), dtype=int)
P5 = numpy.array(strassen_matrix_multiplication(K6, B22))
P6 = numpy.zeros(shape=(l, l), dtype=int)
P6 = numpy.array(strassen_matrix_multiplication(K7, K8))
P7 = numpy.zeros(shape=(l, l), dtype=int)
P7 = numpy.array(strassen_matrix_multiplication(K9, K10))
#C11=M1+M4−M5+M7
#C12=M3+M5
#C21=M2+M4
#C22=M1−M2+M3 +M6
L1 = numpy.zeros(shape=(l, l), dtype=int)
L1 = P1 + P4
L2 = numpy.zeros(shape=(l, l), dtype=int)
L2 = P7 + numpy.negative(P5)
C11 = numpy.zeros(shape=(l, l), dtype=int)
C11 = L1 + L2
C12 = numpy.zeros(shape=(l, l), dtype=int)
C12 = P3 + P5
C21 = numpy.zeros(shape=(l, l), dtype=int)
C21 = P2 + P4
L3 = numpy.zeros(shape=(l, l), dtype=int)
L3 = P1 + numpy.negative(P2)
L4 = numpy.zeros(shape=(l, l), dtype=int)
L4 = P3 + P6
C22 = numpy.zeros(shape=(l, l), dtype=int)
C22 = L3 + L4
print()
print()
print()
#print("C11",C11)
#print()
#print("C12",C12)
#print()
#print("C21",C21)
#print()
#print("C22",C22)
C11 = C11.tolist()
C12 = C12.tolist()
C21 = C21.tolist()
C22 = C22.tolist()
#print(type(C12))
C1 = numpy.zeros(shape=(l, l), dtype=int).tolist()
C2 = numpy.zeros(shape=(l, l), dtype=int).tolist()
C = []
for i in range(l):
C1[i] = C11[i][:] + C12[i][:]
C2[i] = C21[i][:] + C22[i][:]
for i in range(l):
C.append([0] * l)
#for i in range(l):
C = C1 + C2
#print("C1",C1)
#print("C2",C2)
#print()
return C
def np_sigmoid(x):
return np.divide(1, np.add(1, np.exp(np.negative(x))))
def test_half_ufuncs(self):
"""Test the various ufuncs"""
a = np.array([0, 1, 2, 4, 2], dtype=float16)
b = np.array([-2, 5, 1, 4, 3], dtype=float16)
c = np.array([0, -1, -np.inf, np.nan, 6], dtype=float16)
assert_equal(np.add(a, b), [-2, 6, 3, 8, 5])
assert_equal(np.subtract(a, b), [2, -4, 1, 0, -1])
assert_equal(np.multiply(a, b), [0, 5, 2, 16, 6])
assert_equal(np.divide(a, b), [0, 0.199951171875, 2, 1, 0.66650390625])
assert_equal(np.equal(a, b), [False, False, False, True, False])
assert_equal(np.not_equal(a, b), [True, True, True, False, True])
assert_equal(np.less(a, b), [False, True, False, False, True])
assert_equal(np.less_equal(a, b), [False, True, False, True, True])
assert_equal(np.greater(a, b), [True, False, True, False, False])
assert_equal(np.greater_equal(a, b), [True, False, True, True, False])
assert_equal(np.logical_and(a, b), [False, True, True, True, True])
assert_equal(np.logical_or(a, b), [True, True, True, True, True])
assert_equal(np.logical_xor(a, b), [True, False, False, False, False])
assert_equal(np.logical_not(a), [True, False, False, False, False])
assert_equal(np.isnan(c), [False, False, False, True, False])
assert_equal(np.isinf(c), [False, False, True, False, False])
assert_equal(np.isfinite(c), [True, True, False, False, True])
assert_equal(np.signbit(b), [True, False, False, False, False])
assert_equal(np.copysign(b, a), [2, 5, 1, 4, 3])
assert_equal(np.maximum(a, b), [0, 5, 2, 4, 3])
with suppress_warnings() as sup:
sup.record(RuntimeWarning)
x = np.maximum(b, c)
assert_(np.isnan(x[3]))
assert_equal(len(sup.log), 1)
x[3] = 0
assert_equal(x, [0, 5, 1, 0, 6])
assert_equal(np.minimum(a, b), [-2, 1, 1, 4, 2])
with suppress_warnings() as sup:
sup.record(RuntimeWarning)
x = np.minimum(b, c)
assert_(np.isnan(x[3]))
assert_equal(len(sup.log), 1)
x[3] = 0
assert_equal(x, [-2, -1, -np.inf, 0, 3])
assert_equal(np.fmax(a, b), [0, 5, 2, 4, 3])
assert_equal(np.fmax(b, c), [0, 5, 1, 4, 6])
assert_equal(np.fmin(a, b), [-2, 1, 1, 4, 2])
assert_equal(np.fmin(b, c), [-2, -1, -np.inf, 4, 3])
assert_equal(np.floor_divide(a, b), [0, 0, 2, 1, 0])
assert_equal(np.remainder(a, b), [0, 1, 0, 0, 2])
assert_equal(np.divmod(a, b), ([0, 0, 2, 1, 0], [0, 1, 0, 0, 2]))
assert_equal(np.square(b), [4, 25, 1, 16, 9])
assert_equal(np.reciprocal(b), [-0.5, 0.199951171875, 1, 0.25, 0.333251953125])
assert_equal(np.ones_like(b), [1, 1, 1, 1, 1])
assert_equal(np.conjugate(b), b)
assert_equal(np.absolute(b), [2, 5, 1, 4, 3])
assert_equal(np.negative(b), [2, -5, -1, -4, -3])
assert_equal(np.positive(b), b)
assert_equal(np.sign(b), [-1, 1, 1, 1, 1])
assert_equal(np.modf(b), ([0, 0, 0, 0, 0], b))
assert_equal(np.frexp(b), ([-0.5, 0.625, 0.5, 0.5, 0.75], [2, 3, 1, 3, 2]))
assert_equal(np.ldexp(b, [0, 1, 2, 4, 2]), [-2, 10, 4, 64, 12])
def create_psych_curve_plot(sessions):
data_mean = []
data_errorbar = []
data_fit = []
for session in sessions.fetch('KEY'):
contrasts, prob_right, prob_left, \
threshold, bias, lapse_low, lapse_high, \
n_trials, n_trials_right = \
(sessions & session).fetch1(
'signed_contrasts', 'prob_choose_right', 'prob_left',
'threshold', 'bias', 'lapse_low', 'lapse_high',
'n_trials_stim', 'n_trials_stim_right')
pars = [bias, threshold, lapse_low, lapse_high]
contrasts = contrasts * 100
contrasts_fit = np.arange(-100, 100)
prob_right_fit = psy.erf_psycho_2gammas(pars, contrasts_fit)
ci = smp.proportion_confint(
n_trials_right, n_trials, alpha=0.032,
method='normal') - prob_right
curve_color, error_color = get_color(prob_left, 0.3)
behavior_data = go.Scatter(
x=contrasts.tolist(),
y=prob_right.tolist(),
marker=dict(size=6,
color=curve_color,
line=dict(color='white', width=1)),
mode='markers',
name=f'p_left = {prob_left}, data with 68% CI')
behavior_errorbar = go.Scatter(x=contrasts.tolist(),
y=prob_right.tolist(),
error_y=dict(type='data',
array=ci[0].tolist(),
arrayminus=np.negative(
ci[1]).tolist(),
visible=True,
color=error_color),
marker=dict(size=6, ),
mode='none',
showlegend=False)
behavior_fit = go.Scatter(x=contrasts_fit.tolist(),
y=prob_right_fit.tolist(),
name=f'p_left = {prob_left} model fits',
marker=dict(color=curve_color))
data_mean.append(behavior_data)
data_errorbar.append(behavior_errorbar)
data_fit.append(behavior_fit)
layout = go.Layout(width=630,
height=350,
title=dict(text='Psychometric Curve', x=0.25, y=0.85),
xaxis=dict(title='Contrast (%)'),
yaxis=dict(title='Probability choosing right',
range=[-0.05, 1.05]),
template=dict(layout=dict(plot_bgcolor="white")))
data = data_errorbar
for element in data_fit:
data.append(element)
for element in data_mean:
data.append(element)
return go.Figure(data=data, layout=layout)
def neg(x):
return np.negative(x)
def __neg__(self):
neg_n_body_tensors = dict()
for key in self.n_body_tensors:
neg_n_body_tensors[key] = numpy.negative(self.n_body_tensors[key])
return PolynomialTensor(neg_n_body_tensors)
if r != 0: # A: singular
return False
for i in range(k):
if out[i] <= 0:
return False
val = out[-1]
if k == m:
return True
own_supp_flags = np.zeros(m, np.bool_)
own_supp_flags[own_supp] = True
for i in range(m):
if not own_supp_flags[i]:
payoff = 0
for j in range(k):
payoff += payoff_matrix[i, opp_supp[j]] * out[j]
if payoff > val:
return False
return True
playerA = np.array([[0, -1, 1], [1, 0, -1], [-1, 1, 0]])
playerB = np.negative(playerA)
rps = NormalFormGame(playerA)
#([[playerA, playerB]])
#print(playerA)
#print(playerB)
print(rps)
support_enumeration(rps)
def plot_summary(all_true_states, all_mean_belief, all_variance_belief, sample_period, all_kt=None):
time_steps = list(range(len(all_true_states)))
time_steps_in_seconds = [t*sample_period for t in time_steps]
all_true_states = np.array(all_true_states)
all_mean_belief = np.array(all_mean_belief)
all_variance_belief = np.array(all_variance_belief)
true_x = all_true_states[:, 0, 0]
true_y = all_true_states[:, 1, 0]
true_theta = all_true_states[:, 2, 0]
mean_beliefs_about_x = all_mean_belief[:, 0, 0]
mean_beliefs_about_y = all_mean_belief[:, 1, 0]
mean_beliefs_about_theta = all_mean_belief[:, 2, 0]
var_beliefs_about_x = all_variance_belief[:, 0, 0]
var_beliefs_about_y = all_variance_belief[:, 1, 0]
var_beliefs_about_theta = all_variance_belief[:, 2, 0]
# Add static plots
_, axes = plt.subplots(3, 2, figsize=(15, 15))
ax1 = axes[0, 0]
ax2 = axes[1, 0]
ax3 = axes[1, 1]
ax4 = axes[0, 1]
ax5 = axes[2, 0]
ax6 = axes[2, 1]
ax1.plot(time_steps_in_seconds, true_x)
ax1.plot(time_steps_in_seconds, mean_beliefs_about_x, '--')
ax1.plot(time_steps_in_seconds, true_y)
ax1.plot(time_steps_in_seconds, mean_beliefs_about_y, '--')
ax1.plot(time_steps_in_seconds, true_theta)
ax1.plot(time_steps_in_seconds, mean_beliefs_about_theta, '--')
ax1.set_title("State vs Mean belief about State")
ax1.set_xlabel("Time (s)")
ax1.legend(["Actual X", "Mean X Belief",
"Actual Y", "Mean Y Belief",
"Actual Theta", "Mean Theta Belief"])
x_error = [
(xt-mean_beliefs_about_x[i]) for i, xt in enumerate(true_x)]
ax2.plot(time_steps_in_seconds, x_error)
ax2.plot(time_steps_in_seconds, np.sqrt(var_beliefs_about_x)*2, 'b--')
ax2.plot(time_steps_in_seconds, np.negative(
np.sqrt(np.abs(var_beliefs_about_x))*2), 'b--')
ax2.legend(["X Error", "X Variance"])
ax2.set_title("Error from X and mean belief")
ax2.set_xlabel("Time (s)")
ax2.set_ylabel("X (m)")
ax2.set_ylim(-0.5, 0.5)
y_error = [
(vt-mean_beliefs_about_y[i])for i, vt in enumerate(true_y)]
ax3.plot(time_steps_in_seconds, y_error)
ax3.plot(time_steps_in_seconds,
np.sqrt(np.abs(var_beliefs_about_y))*2, 'y--')
ax3.plot(time_steps_in_seconds,
np.negative(np.sqrt(np.abs(var_beliefs_about_y))*2), 'y--')
ax3.legend(["Y Error", "Y Variance"])
ax3.set_title("Error from Y and mean belief")
ax3.set_xlabel("Time (s)")
ax3.set_ylabel("Y (m)")
ax3.set_ylim(-0.5, 0.5)
theta_error = [
(vt-mean_beliefs_about_theta[i])for i, vt in enumerate(true_theta)]
ax4.plot(time_steps_in_seconds, theta_error)
ax4.plot(time_steps_in_seconds,
np.sqrt(np.abs(var_beliefs_about_theta))*2, 'y--')
ax4.plot(time_steps_in_seconds,
np.negative(np.sqrt(np.abs(var_beliefs_about_theta))*2), 'y--')
ax4.legend(["Theta Error", "Theta Variance"])
ax4.set_title("Error from theta and mean belief")
ax4.set_xlabel("Time (s)")
ax4.set_ylabel("Theta (radians)")
ax4.set_ylim(-0.174, 0.174)
if all_kt is not None:
ax5.plot(time_steps_in_seconds, np.array(all_kt)[:, 0, 0])
ax5.plot(time_steps_in_seconds, np.array(all_kt)[:, 1, 0])
ax5.plot(time_steps_in_seconds, np.array(all_kt)[:, 2, 0])
ax5.plot(time_steps_in_seconds, np.array(all_kt)[:, 0, 1])
ax5.plot(time_steps_in_seconds, np.array(all_kt)[:, 1, 1])
ax5.plot(time_steps_in_seconds, np.array(all_kt)[:, 2, 1])
ax5.set_title("Kalman filter gain for position")
ax5.legend(["X kalman gain range", "Y kalman gain range", "Theta Kalman Gain range",
"X kalman gain bearing", "Y kalman gain bearing", "Theta Kalman Gain bearing"])
sc = plt.imshow(all_variance_belief[-1], cmap='Blues', interpolation='nearest', origin='lower')
plt.colorbar(sc)
plt.show()
plt.pause(200)
def gpdfitnew(x, sort=True, sort_in_place=False, return_quadrature=False):
"""Estimate the paramaters for the Generalized Pareto Distribution (GPD)
Returns empirical Bayes estimate for the parameters of the two-parameter
generalized Parato distribution given the data.
Parameters
----------
x : ndarray
One dimensional data array
sort : bool or ndarray, optional
If known in advance, one can provide an array of indices that would
sort the input array `x`. If the input array is already sorted, provide
False. If True (default behaviour), the array is sorted internally.
sort_in_place : bool, optional
If `sort` is True and `sort_in_place` is True, the array is sorted
in-place (False by default).
return_quadrature : bool, optional
If True, quadrature points and weight `ks` and `w` of the marginal posterior distribution of k are also calculated and returned. False by
default.
Returns
-------
k, sigma : float
estimated parameter values
ks, w : ndarray
Quadrature points and weights of the marginal posterior distribution
of `k`. Returned only if `return_quadrature` is True.
Notes
-----
This function returns a negative of Zhang and Stephens's k, because it is
more common parameterisation.
"""
if x.ndim != 1 or len(x) <= 1:
raise ValueError("Invalid input array.")
# check if x should be sorted
if sort is True:
if sort_in_place:
x.sort()
xsorted = True
else:
sort = np.argsort(x)
xsorted = False
elif sort is False:
xsorted = True
else:
xsorted = False
n = len(x)
PRIOR = 3
m = 30 + int(np.sqrt(n))
bs = np.arange(1, m + 1, dtype=float)
bs -= 0.5
np.divide(m, bs, out=bs)
np.sqrt(bs, out=bs)
np.subtract(1, bs, out=bs)
if xsorted:
bs /= PRIOR * x[int(n / 4 + 0.5) - 1]
bs += 1 / x[-1]
else:
bs /= PRIOR * x[sort[int(n / 4 + 0.5) - 1]]
bs += 1 / x[sort[-1]]
ks = np.negative(bs)
temp = ks[:, None] * x
np.log1p(temp, out=temp)
np.mean(temp, axis=1, out=ks)
L = bs / ks
np.negative(L, out=L)
np.log(L, out=L)
L -= ks
L -= 1
L *= n
temp = L - L[:, None]
np.exp(temp, out=temp)
w = np.sum(temp, axis=1)
np.divide(1, w, out=w)
# remove negligible weights
dii = w >= 10 * np.finfo(float).eps
if not np.all(dii):
w = w[dii]
bs = bs[dii]
# normalise w
w /= w.sum()
# posterior mean for b
b = np.sum(bs * w)
# Estimate for k, note that we return a negative of Zhang and
# Stephens's k, because it is more common parameterisation.
temp = (-b) * x
np.log1p(temp, out=temp)
k = np.mean(temp)
if return_quadrature:
np.negative(x, out=temp)
temp = bs[:, None] * temp
np.log1p(temp, out=temp)
ks = np.mean(temp, axis=1)
# estimate for sigma
sigma = -k / b * n / (n - 0)
# weakly informative prior for k
a = 10
k = k * n / (n + a) + a * 0.5 / (n + a)
if return_quadrature:
ks *= n / (n + a)
ks += a * 0.5 / (n + a)
if return_quadrature:
return k, sigma, ks, w
else:
return k, sigma
def iterateMatrix(matrix, goTerms, goEnrichment, background):
""" return a dictionary
Keyword arguments:
matrix -- numerical matrix of semantic similarities
goTerms -- list of goTerms
goEnrichment -- GO enrichment result
background -- flattened background: lists of genes and GO Terms
iterates through semantic similarity matrix in decreasing ss order
"""
numGenes = len(list(dict.fromkeys(background[0])))
max = goEnrichment.max().max()
min = goEnrichment.min().min()
# p-values are considered similar if they have a maximum difference of 5%
maxDiff = (max - min) * 0.05
# frequencies of GO temrs
frequencies = dict(Counter(background[1]))
frequencies = {k: v / numGenes for k, v in frequencies.items()}
avgs = dict()
# calculate averages for each term for uniqueness value
for index, term in enumerate(goTerms):
col = matrix[:, index]
row = matrix[index, :]
col = col[col != -1]
row = row[row != -1]
avgs[term] = (col.sum() + row.sum()) / (len(goTerms) - 1)
# stores tree structure
tree = dict()
# stores additional data for each GO term
goList = dict()
while len(goTerms) > 0:
maxValue = np.amax(np.ravel(matrix))
# get most similar pair of GO terms
indices = np.where(matrix == maxValue)
indices = list(zip(indices[0], indices[1]))[0]
termA = goTerms[indices[0]]
termB = goTerms[indices[1]]
# calculate which GO term is rejected
delete = testGoTerms(termA, termB, goEnrichment, background, frequencies, maxDiff)
toDelete = delete["term"]
if toDelete == termA:
deleteIndex = indices[0]
toKeep = termB
else:
deleteIndex = indices[1]
toKeep = termA
# add GO terms to current tree dict
if toKeep != toDelete:
if toDelete in tree:
if toKeep in tree:
# if both terms are in the tree dict, it means that they form non-connected subtrees.
# The tree dict of the rejected term is placed as the child of the kept term
tree[toKeep][toDelete] = tree[toDelete]
else:
# if the kept term is not in the tree dict but the rejected is,
# the kept term will be placed as the parent of the rejected term
tree[toKeep] = {}
tree[toKeep][toDelete] = tree[toDelete]
else:
if toKeep in tree:
# if the kept term is in the tree dict, but the rejected is not,
# the rejected term is added as a child of the kept term
tree[toKeep][toDelete] = toDelete
else:
# if none of the terms are in the tree dict, a new entry is created
tree[toKeep] = {toDelete: toDelete}
# not connected trees that have the rejected term as a parent are deleted
# as the rejected term is now incorporated in the final tree dict structure
tree.pop(toDelete, None)
else:
maxValue = 0
goList[toDelete] = {
"termID": toDelete,
"description": godag[toDelete].name,
"frequency": calculateFrequency(toDelete, frequencies),
"rejection": delete["rejection"],
"uniqueness": 1 - avgs[toDelete],
"dispensability": maxValue,
"pvalues": np.negative(2 * np.log(goEnrichment.loc[toDelete, :].values)).tolist()
}
# delete rejected term from list of GO terms and from ss matrix
goTerms = np.delete(goTerms, deleteIndex)
matrix = np.delete(matrix, deleteIndex, 0)
matrix = np.delete(matrix, deleteIndex, 1)
return {"tree": tree, "data": goList}
def logistic_func(X, bayta1, bayta2, bayta3, bayta4):
logisticPart = 1 + np.exp(
np.negative(np.divide(X - bayta3, np.abs(bayta4))))
yhat = bayta2 + np.divide(bayta1 - bayta2, logisticPart)
return yhat
rightXYZ = [0, 0, 0]
print("leftId : ", leftId)
print("rightId : ", rightId)
print("noseId : ", noseId)
pld.GetPoints().GetPoint(noseId, noseXYZ)
pld.GetPoints().GetPoint(leftId, leftXYZ)
pld.GetPoints().GetPoint(rightId, rightXYZ)
print("leftCoord : ", leftXYZ)
print("rightCoord : ", rightXYZ)
print("noseCoord : ", noseXYZ)
center = np.add(rightXYZ, leftXYZ) / 2
ex = np.add(rightXYZ, np.negative(leftXYZ))
ex = ex / np.linalg.norm(ex)
ey = np.add(center, np.negative(noseXYZ))
ey = ey / np.linalg.norm(ey)
# cross product to calculate a normal vector to plane of ex and ey
ez = np.cross(ex, ey)
ez = ez / np.linalg.norm(ez)
rotM = vtk.vtkMatrix4x4()
rotM.Identity()
for i in range(0, 3):
rotM.SetElement(0, i, ex[i])
rotM.SetElement(1, i, ey[i])
rotM.SetElement(2, i, ez[i])
def called_member(self, a):
return np.negative(a)
def quaternion_from_transformation_matrix(matrix, isprecise=False):
"""Return quaternion from rotation matrix.
If isprecise is True, the input matrix is assumed to be a precise rotation
matrix and a faster algorithm is used.
>>> q = quaternion_from_matrix(numpy.identity(4), True)
>>> numpy.allclose(q, [1, 0, 0, 0])
True
>>> q = quaternion_from_matrix(numpy.diag([1, -1, -1, 1]))
>>> numpy.allclose(q, [0, 1, 0, 0]) or numpy.allclose(q, [0, -1, 0, 0])
True
>>> R = rotation_matrix(0.123, (1, 2, 3))
>>> q = quaternion_from_matrix(R, True)
>>> numpy.allclose(q, [0.9981095, 0.0164262, 0.0328524, 0.0492786])
True
>>> R = [[-0.545, 0.797, 0.260, 0], [0.733, 0.603, -0.313, 0],
... [-0.407, 0.021, -0.913, 0], [0, 0, 0, 1]]
>>> q = quaternion_from_matrix(R)
>>> numpy.allclose(q, [0.19069, 0.43736, 0.87485, -0.083611])
True
>>> R = [[0.395, 0.362, 0.843, 0], [-0.626, 0.796, -0.056, 0],
... [-0.677, -0.498, 0.529, 0], [0, 0, 0, 1]]
>>> q = quaternion_from_matrix(R)
>>> numpy.allclose(q, [0.82336615, -0.13610694, 0.46344705, -0.29792603])
True
>>> R = random_rotation_matrix()
>>> q = quaternion_from_matrix(R)
>>> is_same_transform(R, quaternion_matrix(q))
True
>>> is_same_quaternion(quaternion_from_matrix(R, isprecise=False),
... quaternion_from_matrix(R, isprecise=True))
True
>>> R = euler_matrix(0.0, 0.0, numpy.pi/2.0)
>>> is_same_quaternion(quaternion_from_matrix(R, isprecise=False),
... quaternion_from_matrix(R, isprecise=True))
True
"""
M = np.array(matrix, dtype=np.float64, copy=False)[:4, :4]
if isprecise:
q = np.empty((4, ))
t = np.trace(M)
if t > M[3, 3]:
q[0] = t
q[3] = M[1, 0] - M[0, 1]
q[2] = M[0, 2] - M[2, 0]
q[1] = M[2, 1] - M[1, 2]
else:
i, j, k = 0, 1, 2
if M[1, 1] > M[0, 0]:
i, j, k = 1, 2, 0
if M[2, 2] > M[i, i]:
i, j, k = 2, 0, 1
t = M[i, i] - (M[j, j] + M[k, k]) + M[3, 3]
q[i] = t
q[j] = M[i, j] + M[j, i]
q[k] = M[k, i] + M[i, k]
q[3] = M[k, j] - M[j, k]
q = q[[3, 0, 1, 2]]
q *= 0.5 / math.sqrt(t * M[3, 3])
else:
m00 = M[0, 0]
m01 = M[0, 1]
m02 = M[0, 2]
m10 = M[1, 0]
m11 = M[1, 1]
m12 = M[1, 2]
m20 = M[2, 0]
m21 = M[2, 1]
m22 = M[2, 2]
# symmetric matrix K
K = np.array([[m00 - m11 - m22, 0.0, 0.0, 0.0],
[m01 + m10, m11 - m00 - m22, 0.0, 0.0],
[m02 + m20, m12 + m21, m22 - m00 - m11, 0.0],
[m21 - m12, m02 - m20, m10 - m01, m00 + m11 + m22]])
K /= 3.0
# quaternion is eigenvector of K that corresponds to largest eigenvalue
w, V = np.linalg.eigh(K)
q = V[[3, 0, 1, 2], np.argmax(w)]
if q[0] < 0.0:
np.negative(q, q)
return q
def __neg__(self):
"""Return the negative of F: G = -F => G(x) = -F(x) for all x"""
return Factor().__build(self.v.copy(), np.negative(self.t))
arr_i = buildArray('i.quadratic')
arr_u = buildArray('u.quadratic')
allMovieMetadata = getMetadata()
dotProds = []
watchedMovies = []
for i in range(0, 10000):
x = Weight((arr_u[i].id, arr_i[i].id),
np.dot(arr_u[i].weights, arr_i[i].weights))
dotProds.append(x)
watchedMovies.append((arr_u[i].id, arr_i[i].id))
# build -1/1 matrix (extremely sparse)
seenMovies = np.ones((totalUsers, totalMovies))
seenMovies = np.negative(seenMovies)
for watch in watchedMovies:
# print(str(watch[0]) + "\t" + str(watch[1]))
seenMovies[watch[0] - 1][watch[1] - 1] = 1
print(len(allMovieMetadata))
# assign metadata for logistic regression, this model is 30 GB!!
dfs = []
with open('data.vw', 'w') as f:
for i in range(totalUsers):
print("processing user: "******" ...")
for j in range(totalMovies):
# print(i , j)
# print("J", j)
m = allMovieMetadata[j - 1].metadata
def _multiply_no_nan(x, y, name=None): # pylint: disable=unused-argument
dtype = np.result_type(x, y)
# TODO(b/146385087): The gradient should be
# `lambda dz: [multiply_no_nan(dz, y), multiply_no_nan(x, dz)]`.
return np.where(np.equal(y, 0.), np.zeros((), dtype=dtype),
np.multiply(x, y))
multiply_no_nan = utils.copy_docstring('tf.math.multiply_no_nan',
_multiply_no_nan)
ndtri = utils.copy_docstring('tf.math.ndtri',
lambda x, name=None: scipy_special.ndtri(x))
negative = utils.copy_docstring('tf.math.negative',
lambda x, name=None: np.negative(x))
nextafter = utils.copy_docstring(
'tf.math.nextafter', lambda x1, x2, name=None: np.nextafter(x1, x2))
not_equal = utils.copy_docstring('tf.math.not_equal',
lambda x, y, name=None: np.not_equal(x, y))
polygamma = utils.copy_docstring(
'tf.math.polygamma',
lambda a, x, name=None: scipy_special.polygamma(np.int32(a), x).astype( # pylint: disable=unused-argument,g-long-lambda
utils.common_dtype([a, x], dtype_hint=np.float32)))
polyval = utils.copy_docstring(
'tf.math.polyval', lambda coeffs, x, name=None: np.polyval(coeffs, x))
def generate(self,
serial_number_list,
original_img_list,
generated_img_list,
defect_category_list,
bbox_list,
dataset_name,
is_difference_img_preprocessed,
is_removed=True,
properties=None):
ok_data_save_dir_path, ng_data_save_dir_path = self._create_save_directory(
directory_name=dataset_name, is_removed=is_removed)
serial_number_list = list(
map(lambda serial_number: serial_number.split("_")[0],
serial_number_list))
original_img_path_list = list()
for i, (serial_number, original_img, generated_img, defect_category,
bbox) in enumerate(
zip(serial_number_list, original_img_list,
generated_img_list, defect_category_list, bbox_list)):
difference_img = self.get_difference_img(
original_img=original_img,
generated_img=generated_img,
is_preprocessed=is_difference_img_preprocessed,
is_boundary_mask=False)
save_img = None
# Generate label.csv
# Controller with save category
if dataset_name == "difference_img":
pos_difference_img = np.where(difference_img >= 0,
difference_img, 0)
neg_difference_img = np.negative(
np.where(difference_img <= 0, difference_img, 0))
abs_difference_img = np.abs(difference_img)
# pos_difference_img = pos_difference_img.astype(np.uint8)
# neg_difference_img = neg_difference_img.astype(np.uint8)
# Save
if defect_category[0] == 1:
# Original image
original_save_path = os.path.join(
ok_data_save_dir_path,
f"{serial_number}_{i}_{0}_o.npy")
np.save(original_save_path, original_img)
# Generated image
generated_save_path = os.path.join(
ok_data_save_dir_path,
f"{serial_number}_{i}_{0}_g.npy")
np.save(generated_save_path, generated_img)
# Positive image
pos_save_path = os.path.join(
ok_data_save_dir_path,
f"{serial_number}_{i}_{0}_p.npy")
np.save(pos_save_path, pos_difference_img)
# cv2.imwrite(pos_save_path, pos_difference_img)
# Negative image
neg_save_path = os.path.join(
ok_data_save_dir_path,
f"{serial_number}_{i}_{0}_n.npy")
np.save(neg_save_path, neg_difference_img)
# cv2.imwrite(neg_save_path, neg_difference_img)
# Absolute image
abs_save_path = os.path.join(
ok_data_save_dir_path,
f"{serial_number}_{i}_{0}_a.npy")
np.save(abs_save_path, abs_difference_img)
# cv2.imwrite(abs_save_path, abs_difference_img)
else:
for category_idx, defect in enumerate(defect_category):
if defect == 1:
# Original image
original_save_path = os.path.join(
ng_data_save_dir_path,
f"{serial_number}_{i}_{category_idx}_o.npy")
np.save(original_save_path, original_img)
# Generated image
generated_save_path = os.path.join(
ng_data_save_dir_path,
f"{serial_number}_{i}_{category_idx}_g.npy")
np.save(generated_save_path, generated_img)
# Positive image
pos_save_path = os.path.join(
ng_data_save_dir_path,
f"{serial_number}_{i}_{category_idx}_p.npy")
np.save(pos_save_path, pos_difference_img)
# cv2.imwrite(pos_save_path, pos_difference_img)
# Negative image
neg_save_path = os.path.join(
ng_data_save_dir_path,
f"{serial_number}_{i}_{category_idx}_n.npy")
np.save(neg_save_path, neg_difference_img)
# cv2.imwrite(neg_save_path, neg_difference_img)
# Absolute image
abs_save_path = os.path.join(
ng_data_save_dir_path,
f"{serial_number}_{i}_{category_idx}_a.npy")
np.save(abs_save_path, abs_difference_img)
# cv2.imwrite(abs_save_path, abs_difference_img)
else:
continue
else:
if dataset_name == "defect_part":
"""
* properties
cut_size=32
"""
save_img = self.get_defect_img(
difference_img=difference_img, **properties)
save_img = save_img.astype(np.uint8)
elif dataset_name == "important_pixel":
"""
* properties
n=100
img_size=256
"""
save_img = self.get_important_pixel_img(
difference_img=difference_img, **properties)
save_img = save_img.astype(np.uint8)
# Save
if defect_category[0] == 1:
save_path = os.path.join(ok_data_save_dir_path,
f"{serial_number}_{i}_{0}.png")
cv2.imwrite(save_path, save_img)
else:
for category_idx, defect in enumerate(defect_category):
if defect == 1:
save_path = os.path.join(
ng_data_save_dir_path,
f"{serial_number}_{i}_{category_idx}.png")
cv2.imwrite(save_path, save_img)
else:
continue
class negative(UnaryElemwise):
_r = np.negative
_d = lambda self, x: np.negative(np.ones_like(x))
def gaussian(self, x):
sq = np.square(x)
neg = np.negative(sq)
return np.exp(neg)
def gpinv(p, k, sigma):
"""Inverse Generalised Pareto distribution function."""
x = np.empty(p.shape)
x.fill(np.nan)
if sigma <= 0:
return x
ok = (p > 0) & (p < 1)
if np.all(ok):
if np.abs(k) < np.finfo(float).eps:
np.negative(p, out=x)
np.log1p(x, out=x)
np.negative(x, out=x)
else:
np.negative(p, out=x)
np.log1p(x, out=x)
x *= -k
np.expm1(x, out=x)
x /= k
x *= sigma
else:
if np.abs(k) < np.finfo(float).eps:
# x[ok] = - np.log1p(-p[ok])
temp = p[ok]
np.negative(temp, out=temp)
np.log1p(temp, out=temp)
np.negative(temp, out=temp)
x[ok] = temp
else:
# x[ok] = np.expm1(-k * np.log1p(-p[ok])) / k
temp = p[ok]
np.negative(temp, out=temp)
np.log1p(temp, out=temp)
temp *= -k
np.expm1(temp, out=temp)
temp /= k
x[ok] = temp
x *= sigma
x[p == 0] = 0
if k >= 0:
x[p == 1] = np.inf
else:
x[p == 1] = -sigma / k
return x
np.insert(deltaThetaS, 0, np.amax(deltaThetaS)),
(len(deltaThetaS) + 1, 1))
zipSortPlusS = np.append(zipSortS, deltaThetaS, axis=1)
lineOutS = np.argmin(zipSortPlusS, axis=0)[-1]
zipSortMinusS = np.delete(zipSortPlusS, lineOutS, axis=0)
#Integrating the fluxes towards the cylinder
UrLocalS = zipSortMinusS[:,
-3] #Third column from end to beginning is Ur
rpLocalS = zipSortMinusS[:,
1] #Second column (starting from 0 index) is rpLocal
thetaLocalS = zipSortMinusS[:,
0] #First column (starting from 0 index) is thetaLocal
print("The size of thetaLocalS is: ", thetaLocalS.shape)
NTheta = np.size(thetaLocalS)
UrMClippedS = np.negative(np.clip(UrLocalS, -np.finfo('d').max, 0))
##kAreaS=deltaThetaS[0]*(rpLocalS[0]+1)*rCollector*deltaL[2]
##radialVolumeAccS=UrMClippedS*kAreaS*deltaT
#Writing the angles headers if it is the first line to write
if (existingLastEnd == -1000):
outputLine = np.reshape(np.insert(thetaLocalS, 0, [-1000, -1000]),
(1, NTheta + 2))
print("The shape of outputLine is: ", outputLine.shape)
fOut = open(outputName, mode='ab')
np.savetxt(fOut, outputLine)
fOut.close()
#End if existingLastEnd==-1000 (For printing the header
#Finding the beginning and end of accumulation for the first time
indexIni = startIndexTimes
def func(arr):
return np.negative(arr)
grid = np.arange(16).reshape((4, 4)) upper, lower = np.vsplit(grid, [2]) left, right = np.hsplit(grid, [2]) print(upper) print(lower) print(left) print(right) ########################################### # 通用函数 ########################################### np.arange(5) / np.arange(1, 6) x = np.arange(9).reshape((3, 3)) 2**x np.add(1, 2) np.subtract(1, 2) np.negative(2) np.multiply(2, 3) np.divide(6, 2) np.floor_divide(3, 2) np.power(2, 3) np.mod(9, 4) np.abs(-8) np.sin(np.pi) np.arctan(1) np.exp(4) np.log(10) np.log2(2) x = np.arange(5) y = np.empty(5) np.multiply(x, 10, out=y)
def __init__(self, T, K, sigma=0.5, _lambda=0.5):
self.SPEED_TO_ERPM_OFFSET = float(
rospy.get_param("/vesc/speed_to_erpm_offset", 0.0))
self.SPEED_TO_ERPM_GAIN = float(
rospy.get_param("/vesc/speed_to_erpm_gain", 4614.0))
self.STEERING_TO_SERVO_OFFSET = float(
rospy.get_param("/vesc/steering_angle_to_servo_offset", 0.5304))
self.STEERING_TO_SERVO_GAIN = float(
rospy.get_param("/vesc/steering_angle_to_servo_gain", -1.2135))
self.CAR_LENGTH = 0.33
self.last_pose = None
# MPPI params
self.T = T # Length of rollout horizon
self.K = K # Number of sample rollouts
self.sigma = sigma
self._lambda = _lambda
self.goal = None # Lets keep track of the goal pose (world frame) over time
self.lasttime = None
# PyTorch / GPU data configuration
# TODO
# you should pre-allocate GPU memory when you can, and re-use it when
# possible for arrays storing your controls or calculated MPPI costs, etc
model_name = rospy.get_param("~nn_model",
"myneuralnetisbestneuralnet.pt")
self.model = torch.load(model_name)
self.model.cuda() # tell torch to run the network on the GPU
self.dtype = torch.cuda.FloatTensor
print("Loading:", model_name)
print("Model:\n", self.model)
print("Torch Datatype:", self.dtype)
# control outputs
self.msgid = 0
# visualization paramters
self.num_viz_paths = 40
if self.K < self.num_viz_paths:
self.num_viz_paths = self.K
# We will publish control messages and a way to visualize a subset of our
# rollouts, much like the particle filter
self.ctrl_pub = rospy.Publisher(rospy.get_param(
"~ctrl_topic", "/vesc/high_level/ackermann_cmd_mux/input/nav0"),
AckermannDriveStamped,
queue_size=2)
self.path_pub = rospy.Publisher("/mppi/paths",
Path,
queue_size=self.num_viz_paths)
# Use the 'static_map' service (launched by MapServer.launch) to get the map
map_service_name = rospy.get_param("~static_map", "static_map")
print("Getting map from service: ", map_service_name)
rospy.wait_for_service(map_service_name)
map_msg = rospy.ServiceProxy(
map_service_name,
GetMap)().map # The map, will get passed to init of sensor model
self.map_info = map_msg.info # Save info about map for later use
print("Map Information:\n", self.map_info)
# Create numpy array representing map for later use
self.map_height = map_msg.info.height
self.map_width = map_msg.info.width
array_255 = np.array(map_msg.data).reshape(
(map_msg.info.height, map_msg.info.width))
self.permissible_region = np.zeros_like(array_255, dtype=bool)
self.permissible_region[
array_255 ==
0] = 1 # Numpy array of dimension (map_msg.info.height, map_msg.info.width),
# With values 0: not permissible, 1: permissible
self.permissible_region = np.negative(
self.permissible_region) # 0 is permissible, 1 is not
print("Making callbacks")
self.goal_sub = rospy.Subscriber("/move_base_simple/goal",
PoseStamped,
self.clicked_goal_cb,
queue_size=1)
self.pose_sub = rospy.Subscriber("/pf/ta/viz/inferred_pose",
PoseStamped,
self.mppi_cb,
queue_size=1)
def predict_image(self,
inRaster,
outRaster,
model=None,
inMask=None,
confidenceMap=None,
confidenceMapPerClass=None,
NODATA=0,
SCALE=None,
classifier='GMM',
feedback=None):
"""!@brief The function classify the whole raster image, using per block image analysis.
The classifier is given in classifier and options in kwargs
Input :
inRaster : Filtered image name ('sample_filtered.tif',str)
outRaster :Raster image name ('outputraster.tif',str)
model : model file got from precedent step ('model', str)
inMask : mask to
confidenceMap : map of confidence per pixel
NODATA : Default set to 0 (int)
SCALE : Default set to None
classifier = Default 'GMM'
Output :
nothing but save a raster image and a confidence map if asked
"""
# Open Raster and get additionnal information
raster = gdal.Open(inRaster, gdal.GA_ReadOnly)
if raster is None:
# fix_print_with_import
print('Impossible to open ' + inRaster)
exit()
if inMask is None:
mask = None
else:
mask = gdal.Open(inMask, gdal.GA_ReadOnly)
if mask is None:
# fix_print_with_import
print('Impossible to open ' + inMask)
exit()
# Check size
if (raster.RasterXSize != mask.RasterXSize) or (
raster.RasterYSize != mask.RasterYSize):
# fix_print_with_import
print('Image and mask should be of the same size')
exit()
if SCALE is not None:
M, m = np.asarray(SCALE[0]), np.asarray(SCALE[1])
# Get the size of the image
d = raster.RasterCount
nc = raster.RasterXSize
nl = raster.RasterYSize
# Get the geoinformation
GeoTransform = raster.GetGeoTransform()
Projection = raster.GetProjection()
# Get block size
band = raster.GetRasterBand(1)
block_sizes = band.GetBlockSize()
x_block_size = block_sizes[0]
y_block_size = block_sizes[1]
del band
# Initialize the output
if not os.path.exists(os.path.dirname(outRaster)):
os.makedirs(os.path.dirname(outRaster))
driver = gdal.GetDriverByName('GTiff')
if np.amax(model.classes_) > 255:
dtype = gdal.GDT_UInt16
else:
dtype = gdal.GDT_Byte
dst_ds = driver.Create(outRaster, nc, nl, 1, dtype)
dst_ds.SetGeoTransform(GeoTransform)
dst_ds.SetProjection(Projection)
out = dst_ds.GetRasterBand(1)
if classifier != 'GMM':
nClass = len(model.classes_)
if confidenceMap:
dst_confidenceMap = driver.Create(confidenceMap, nc, nl, 1,
gdal.GDT_Int16)
dst_confidenceMap.SetGeoTransform(GeoTransform)
dst_confidenceMap.SetProjection(Projection)
out_confidenceMap = dst_confidenceMap.GetRasterBand(1)
if confidenceMapPerClass:
dst_confidenceMapPerClass = driver.Create(confidenceMapPerClass,
nc, nl, nClass,
gdal.GDT_Int16)
dst_confidenceMapPerClass.SetGeoTransform(GeoTransform)
dst_confidenceMapPerClass.SetProjection(Projection)
# Perform the classification
total = nl * y_block_size
pushFeedback('Predicting model...')
if feedback == 'gui':
progress = pB.progressBar('Predicting model...', total / 10)
for i in range(0, nl, y_block_size):
if not 'lastBlock' in locals():
lastBlock = i
if int(lastBlock / total * 100) != int(i / total * 100):
lastBlock = i
pushFeedback(int(i / total * 100))
if feedback == 'gui':
progress.addStep()
if i + y_block_size < nl: # Check for size consistency in Y
lines = y_block_size
else:
lines = nl - i
for j in range(0, nc,
x_block_size): # Check for size consistency in X
if j + x_block_size < nc:
cols = x_block_size
else:
cols = nc - j
# Load the data and Do the prediction
X = np.empty((cols * lines, d))
for ind in range(d):
X[:, ind] = raster.GetRasterBand(int(ind + 1)).ReadAsArray(
j, i, cols, lines).reshape(cols * lines)
# Do the prediction
band_temp = raster.GetRasterBand(1)
nodata_temp = band_temp.GetNoDataValue()
if nodata_temp is None:
nodata_temp = -9999
if mask is None:
band_temp = raster.GetRasterBand(1)
mask_temp = band_temp.ReadAsArray(
j, i, cols, lines).reshape(cols * lines)
temp_nodata = np.where(mask_temp != nodata_temp)[0]
#t = np.where((mask_temp!=0) & (X[:,0]!=NODATA))[0]
t = np.where(X[:, 0] != nodata_temp)[0]
yp = np.zeros((cols * lines, ))
#K = np.zeros((cols*lines,))
if confidenceMapPerClass or confidenceMap and classifier != 'GMM':
K = np.zeros((cols * lines, nClass))
K[:, :] = -1
else:
K = np.zeros((cols * lines))
K[:] = -1
else:
mask_temp = mask.GetRasterBand(1).ReadAsArray(
j, i, cols, lines).reshape(cols * lines)
t = np.where((mask_temp != 0)
& (X[:, 0] != nodata_temp))[0]
yp = np.zeros((cols * lines, ))
yp[:] = NODATA
#K = np.zeros((cols*lines,))
if confidenceMapPerClass or confidenceMap and classifier != 'GMM':
K = np.ones((cols * lines, nClass))
K = np.negative(K)
else:
K = np.zeros((cols * lines))
K = np.negative(K)
# TODO: Change this part accorindgly ...
if t.size > 0:
if confidenceMap and classifier == 'GMM':
yp[t], K[t] = model.predict(
self.scale(X[t, :], M=M, m=m), None, confidenceMap)
elif confidenceMap or confidenceMapPerClass and classifier != 'GMM':
yp[t] = model.predict(self.scale(X[t, :], M=M, m=m))
K[t, :] = model.predict_proba(
self.scale(X[t, :], M=M, m=m)) * 100
else:
yp[t] = model.predict(self.scale(X[t, :], M=M, m=m))
#QgsMessageLog.logMessage('amax from predict proba is : '+str(sp.amax(model.predict.proba(self.scale(X[t,:],M=M,m=m)),axis=1)))
# Write the data
out.WriteArray(yp.reshape(lines, cols), j, i)
out.SetNoDataValue(NODATA)
out.FlushCache()
if confidenceMap and classifier == 'GMM':
K *= 100
out_confidenceMap.WriteArray(K.reshape(lines, cols), j, i)
out_confidenceMap.SetNoDataValue(-1)
out_confidenceMap.FlushCache()
if confidenceMap and classifier != 'GMM':
Kconf = np.amax(K, axis=1)
out_confidenceMap.WriteArray(Kconf.reshape(lines, cols), j,
i)
out_confidenceMap.SetNoDataValue(-1)
out_confidenceMap.FlushCache()
if confidenceMapPerClass and classifier != 'GMM':
for band in range(nClass):
gdalBand = band + 1
out_confidenceMapPerClass = dst_confidenceMapPerClass.GetRasterBand(
gdalBand)
out_confidenceMapPerClass.SetNoDataValue(-1)
out_confidenceMapPerClass.WriteArray(
K[:, band].reshape(lines, cols), j, i)
out_confidenceMapPerClass.FlushCache()
del X, yp
# Clean/Close variables
if feedback == 'gui':
progress.reset()
raster = None
dst_ds = None
return outRaster
def molly_parameter(directory, objecto, flux_corrected=True, suffix=''):
'''
Create two molly files with the information about each spectrum.
Telescope options:
Palomar 200in
Wilson
Campanas
Lemmon
WHT
INT
JKT
UKIRT
Kitt Peak
AAT
CTIO
McDonald
MMT
VLT
ANU 2.3m
SAAO 1.9m
NTT
'elsewhere'
Parameters
-------------
directory: Directory where all the science files are located.
objecto: Name of the files to process
flux_corrected: Plot the files that are flux corrected?
suffix: Optional subset of files to plot, i.e. different standards.
If not specified, all the files will be plotted.
Output
-------------
headerfile and listfile of molly files
'''
# Import all the wavelength and flux corrected files
if flux_corrected:
fits_files = glob.glob("%s/%s*WaveStd*%s*.fits" %
(directory, objecto, suffix))
# If not, import the files that are wavelength corrected, but not flux corrected.
else:
fits_files = glob.glob("%s/%s*Wave*.fits" % (directory, objecto))
# Text Files
optimal_files = glob.glob("%s/*%s*%s*_optimal_final.txt" %
(directory, objecto, suffix))
raw_files = glob.glob("%s/*%s*%s*_raw_final.txt" %
(directory, objecto, suffix))
# Header of header output file
header_data = "Object UTC Date RA DEC Dwell Airmass Equinox JD Day Month Year \n"
header_data += " C D C D D D D D D I I I \n"
# Empty list for list output file
list_data = ""
for i in range(len(optimal_files)):
# Edges of the filename to find the filename
a = optimal_files[i].find('/')
# Name of the row name that molly needs
filename = optimal_files[i][a + 1:]
molly_name = "lasc %s %s 1 2 3 A M 0.05 \n" % (filename, i + 1)
# Add to list of all files
list_data += molly_name
##### Header output file #####
# For every variable and every file, extract the value listed here
input_file = fits.open(fits_files[i])
objecto = input_file[0].header['OBJECT']
ut_time = input_file[0].header['UT']
date_obs = input_file[0].header['DATE-OBS']
ra_raw = input_file[0].header['RA']
dec_raw = input_file[0].header['DEC']
exptime = input_file[0].header['EXPTIME']
airmass = input_file[0].header['AIRMASS']
equinox = 2000.0
juliand = Time(date_obs + " " + ut_time).jd
# Get the variables into the right format
# Time
hr = float(ut_time[0:2])
mi = float(ut_time[3:5])
se = float(ut_time[6:])
ut_out = hr + mi / 60.0 + se / 3600.0
# Date
year = date_obs[0:4]
month = date_obs[5:7]
day = date_obs[8:10]
date_out = str(day) + "/" + str(month) + "/" + str(year)
# Right Ascension
if ra_raw[0] == '+':
rahr = float(ra_raw[1:3])
rami = float(ra_raw[4:6])
rase = float(ra_raw[7:])
ra_out = rahr + rami / 60.0 + rase / 3600.0
if ra_raw[0] == '-':
rahr = float(ra_raw[1:3])
rami = float(ra_raw[4:6])
rase = float(ra_raw[7:])
ra_out = np.negative(rahr + rami / 60.0 + rase / 3600.0)
# Declination
if dec_raw[0] == '+':
dechr = float(dec_raw[1:3])
decmi = float(dec_raw[4:6])
decse = float(dec_raw[7:])
dec_out = dechr + decmi / 60.0 + decse / 3600.0
if dec_raw[0] == '-':
dechr = float(dec_raw[1:3])
decmi = float(dec_raw[4:6])
decse = float(dec_raw[7:])
dec_out = np.negative(dechr + decmi / 60.0 + decse / 3600.0)
# One line for each file with all the variables in the right format
molly_header = str(objecto) + " " + str(ut_out) + " " + str(
date_out) + " " + str(ra_out) + " " + str(dec_out) + " " + str(
exptime) + " " + str(airmass) + " " + str(equinox) + " " + str(
juliand) + " " + str(day) + " " + str(month) + " " + str(
year) + "\n"
# Append to the big list of all headers
header_data += molly_header
# Save header output file
headerfile_name = 'headerfile_%s%s.txt' % (objecto, suffix)
headerfile = open(directory + '/' + headerfile_name, 'w')
headerfile.write(header_data)
print('Saved %s/headerfile_%s%s.txt' % (directory, objecto, suffix))
# Save list output file
listfile_name = "listfile_" + objecto + suffix + ".txt"
listfile = open(directory + '/' + listfile_name, "w")
listfile.write(list_data)
print("Saved " + "listfile_" + objecto + suffix + ".txt")
# Create molly instructions file
size = len(optimal_files)
instructions = 'mxpix 4500 sure\n\n@%s.txt\n\nedit 1 %s\nfile\n%s\nq\n\nhfix 1 %s MMT\n\nvbin 1 %s 101\n\n\n\n\n' % (
listfile_name, size, headerfile_name, size, size)
for i in range(len(optimal_files)):
a = optimal_files[i].find('/')
# Name of the row name that molly needs
filename = optimal_files[i][a + 1:-9] + '_molly.txt'
number = i + 101
instructions += 'wasc %s %s ANGSTROMS MJY\n' % (filename, number)
instructions_name = "instructions_" + objecto + suffix + ".txt"
instructions_file = open(directory + '/' + instructions_name, "w")
instructions_file.write(instructions)
print("Saved " + "instructions_" + objecto + suffix + ".txt")
# Hypothesis hTheta = sigmoid(theta * X)
max_iter = 10000 # change the iteration value
# max_iter = 2
cost = np.zeros((max_iter, 1))
for dummyCounter in range(max_iter):
z = np.matmul(X, theta1)
hypothesis = sigmoid(z) # 3163 X thetaSize
# J(θ)= (−yTlog(h)−(1−y)Tlog(1−h))/m
firstPart = np.matmul(np.transpose(yTrainingData), np.log(hypothesis))
secondPart = np.matmul(np.transpose(np.subtract(1, yTrainingData)),
np.log(np.subtract(1, hypothesis)))
cost[dummyCounter] = np.divide(np.negative(np.add(firstPart, secondPart)),
dataToRead)
print(cost[dummyCounter].reshape(-1))
gradient = np.matmul(np.transpose(X), np.subtract(hypothesis,
yTrainingData))
theta1 = np.subtract(theta1,
np.divide(np.multiply(alpha, gradient), dataToRead))
testDataStarts = 3168 # simply hardcoded
testingData = np.array(df.iloc[testDataStarts:, 2])
yTestingData = np.array(df.iloc[testDataStarts:, 3])
totalTestData = 2006 # calculated manually
yTestingData = yTestingData.reshape(totalTestData, 1)
def run(self, x, y=None):
"""
Runs the model for a batch of examples.
The correct outputs `y` are known during training, but not at test time.
If correct outputs `y` are provided, this method must construct and
return a nn.Graph for computing the training loss. If `y` is None, this
method must instead return predicted y-values.
Inputs:
x: a (batch_size x 1) numpy array
y: a (batch_size x 1) numpy array, or None
Output:
(if y is not None) A nn.Graph instance, where the last added node is
the loss
(if y is None) A (batch_size x 1) numpy array of predicted y-values
Note: DO NOT call backprop() or step() inside this method!
"""
"*** YOUR CODE HERE ***"
#function nodes are, multiply, add vector, relu, matrix multiply, add vector
#variables are w1, w2, b1, b2
#size of the input vector
i = x.shape[1]
#to test and modify
h = 100
if not self.w1:
self.w1 = nn.Variable(i, h)
if not self.w2:
self.w2 = nn.Variable(h, i)
if not self.b1:
self.b1 = nn.Variable(h)
if not self.b2:
self.b2 = nn.Variable(i)
graph = nn.Graph([self.w1, self.w2, self.b1, self.b2])
input_nodeX = nn.Input(graph, x)
neg_X = np.negative(x)
neg_inputX = nn.Input(graph, neg_X)
# print x.shape
# xm = MatrixMultiply(graph, input_x, m)
# xm_plus_b = MatrixVectorAdd(graph, xm, b)
multiply1 = nn.MatrixMultiply(graph, input_nodeX, self.w1)
add1 = nn.MatrixVectorAdd(graph, multiply1, self.b1)
relu = nn.ReLU(graph, add1)
multiply2 = nn.MatrixMultiply(graph, relu, self.w2)
add2 = nn.MatrixVectorAdd(graph, multiply2, self.b2)
#for the f(-x)
neg_multiply1 = nn.MatrixMultiply(graph, neg_inputX, self.w1)
neg_add1 = nn.MatrixVectorAdd(graph, neg_multiply1, self.b1)
neg_relu = nn.ReLU(graph, neg_add1)
neg_multiply2 = nn.MatrixMultiply(graph, neg_relu, self.w2)
neg_add2 = nn.MatrixVectorAdd(graph, neg_multiply2, self.b2)
ones = np.ones((1, 1))
ones = np.negative(ones)
neg_one = nn.Input(graph, ones)
neg_negate = nn.MatrixMultiply(graph, neg_add2, neg_one)
final_add = nn.Add(graph, neg_negate, add2)
if y is not None:
# At training time, the correct output `y` is known.
# Here, you should construct a loss node, and return the nn.Graph
# that the node belongs to. The loss node must be the last node
# added to the graph.
input_nodeY = nn.Input(graph, y)
loss_node = nn.SquareLoss(graph, final_add, input_nodeY)
graph.add(loss_node)
return graph
"*** YOUR CODE HERE ***"
else:
# print graph.get_output(add2).shape
return graph.get_output(final_add)
# At test time, the correct output is unknown.
# You should instead return your model's prediction as a numpy array
"*** YOUR CODE HERE ***"