def loadCalibration(self):
coef, domain = np.load(self.calibrationfilename)
self.pixelstowavelengths = Legendre(coef, domain)
self.polynomialdegree = self.pixelstowavelengths.degree()
self.speak("loaded wavelength calibration"
"from {0}".format(self.calibrationfilename))
def subtract_legendre_fit(array2d: np.ndarray,
keep_offset: bool = False,
deg: int = 1) -> Optional[np.ndarray]:
"""
Use a legendre polynomial fit of degree legendre_deg in X and Y direction to correct background.
legendre_deg = 0 ... subtract mean value
legendre_deg = 1 ... subtract mean plane
legendre_deg = 2 ... subtract simple curved mean surface
legendre_deg = 3 ... also corrects "s-shaped" distortion
...
"""
if deg == 0 and keep_offset:
return array2d.copy() # return a copy of input data
n_row = np.linspace(-1, 1, array2d.shape[0])
n_col = np.linspace(-1, 1, array2d.shape[1])
mean_row = array2d.mean(axis=1)
mean_col = array2d.mean(axis=0)
fit_x = Legendre.fit(n_row, mean_row, deg)
fit_y = Legendre.fit(n_col, mean_col, deg)
result = array2d.copy()
result = (result.transpose() -
np.polynomial.legendre.legval(n_row, fit_x.coef)).transpose()
result = result - np.polynomial.legendre.legval(n_col, fit_y.coef)
if keep_offset:
result = result + 2 * array2d.mean(
) # mean was subtracted 2 times (once for fit_x ans once for fit_y)
else:
result = result + array2d.mean()
return result
def basis(self, type, M, X_star=None):
if type is 'monomial':
if X_star is None:
Phi = np.zeros((np.shape(self.X)[0], (M + 1)))
for c in range(0, np.shape(Phi)[1]):
Phi[:, c] = self.X[:, 0].T**c
else:
Phi = np.zeros((np.shape(X_star)[0], (M + 1)))
for c in range(0, np.shape(Phi)[1]):
Phi[:, c] = X_star[:, 0].T**c
if type is 'fourier':
if X_star is None:
Phi = np.zeros((np.shape(self.X)[0], (M + 1) * 2))
for c in range(0, np.shape(Phi)[1], 2):
Phi[:, c] = np.sin(int(c / 2) * np.pi * self.X[:, 0].T)
Phi[:, c + 1] = np.cos(int(c / 2) * np.pi * self.X[:, 0].T)
else:
Phi = np.zeros((np.shape(X_star)[0], (M + 1) * 2))
for c in range(0, np.shape(Phi)[1], 2):
Phi[:, c] = np.sin(int(c / 2) * np.pi * X_star[:, 0].T)
Phi[:, c + 1] = np.cos(int(c / 2) * np.pi * X_star[:, 0].T)
if type is 'legendre':
from numpy.polynomial import Legendre
if X_star is None:
Phi = Legendre.basis(M)(self.X)
else:
Phi = Legendre.basis(M)(X_star)
return Phi
def loadCalibration(self):
'''Load the wavelength calibration.'''
self.speak('trying to load calibration data')
coef, domain = np.load(self.calibrationfilename, allow_pickle=True)
self.pixelstowavelengths = Legendre(coef, domain)
self.polynomialdegree = self.pixelstowavelengths.degree()
self.speak("loaded wavelength calibration"
"from {0}".format(self.calibrationfilename))
def dg_interior_flux_matrix(p):
"""Calculate the interior flux matrix for the standard DG advection equations
There is an analytical expression I could use for this too.
See my thesis page 74.
"""
F = np.zeros((p + 1, p + 1))
for i in range(0, p + 1):
dli = L.basis(i).deriv()
for j in range(0, p + 1):
lj = L.basis(j)
F[i, j] = basis.integrate_legendre_product(dli, lj)
return F
def dg_interior_flux_matrix(p):
"""Calculate the interior flux matrix for the standard DG advection equations
There is an analytical expression I could use for this too.
See my thesis page 74.
"""
F = np.zeros((p + 1, p + 1))
for i in range(0, p + 1):
dli = L.basis(i).deriv()
for j in range(0, p + 1):
lj = L.basis(j)
F[i, j] = basis.integrate_legendre_product(dli, lj)
return F
def wavelength_to_pix(wavelength,
tracepars,
m=1,
frame='dms',
subarray='SUBSTRIP256',
oversample=1):
"""Convert wavelength to pixel coordinates for order m.
:param wavelength: wavelength values in microns.
:param tracepars: the trace polynomial solutions returned by get_tracepars.
:param m: the spectral order.
:param frame: the coordinate frame of the output coordinates (nat, dms or sim).
:param subarray: the subarray of the output coordinates (SUBARRAY256 or SUBARRAY96).
:param oversample: the oversampling factor of the outpur coordinates.
:type wavelength: array[float]
:type tracepars: dict
:type m: int
:type frame: str
:type subarray: str
:type oversample: int
:returns: specpix - the spectral pixel coordinates, spatpix - the spatial pixel coordinates,
mask - an array that is True when the specpix values were within the valid range of the polynomial.
:rtype: Tuple(array[float], array[float], array[bool])
"""
# Convert wavelenght to nat pixel coordinates.
w2spec = Legendre(tracepars[m]['spec_coef'],
domain=tracepars[m]['spec_domain'])
w2spat = Legendre(tracepars[m]['spat_coef'],
domain=tracepars[m]['spat_domain'])
specpix_nat = w2spec(np.log(wavelength))
spatpix_nat = w2spat(np.log(wavelength))
mask = bounds_check(np.log(wavelength), tracepars[m]['spec_domain'][0],
tracepars[m]['spec_domain'][1])
# Convert coordinates to the requested frame.
specpix, spatpix = pix_ref_to_frame(specpix_nat,
spatpix_nat,
frame=frame,
subarray=subarray)
# Oversample the coordinates.
specpix = specpix * oversample
spatpix = spatpix * oversample
return specpix, spatpix, mask
def icb_interface_flux_matrix(p, K, T):
"""The interface flux matrices for the ICB schemes, we use upwinding
to get the fluxes. Denote T the translation necessary to
evaluate the flux from the other cell.
"""
# Enhanced polynomial degree
phat = p + len(K)
G0 = smp.zeros(p + 1, phat + 1)
G1 = smp.zeros(p + 1, phat + 1)
for i in range(0, p + 1):
for j in range(0, phat + 1):
li = L.basis(i)
lj = L.basis(j)
G0[i, j] = leg.legval(-1, li.coef) * \
leg.legval(1, lj.coef) * (T**-1)
G1[i, j] = leg.legval(1, li.coef) * leg.legval(1, lj.coef)
# Enhancement matrix
A, Ainv, B, Binv = enhance.enhancement_matrices(p, K)
# Using the enhanced function in the flux (see notes 21/4/15)
BL = smp.zeros(phat + 1, phat + 1)
BR = smp.zeros(phat + 1, phat + 1)
for i in range(p + 1):
li = L.basis(i)
BL[i, i] = 1 # basis.integrate_legendre_product(li,li)
BR[i, i] = 1 # BL[i,i]
for i, k in enumerate(K):
lk = L.basis(k)
int_lklk = basis.integrate_legendre_product(lk, lk)
BL[i + p + 1, i + p + 1] = T # * int_lklk
BR[i + p + 1, i + p + 1] = (T**(-1)) # * int_lklk
# reduction matrix
R = smp.zeros(phat + 1, p + 1)
for i in range(p + 1):
R[i, i] = 1
for i, k in enumerate(K):
for j in range(p + 1):
R[i + p + 1, j] = auxf.delta(k, j)
# Convert to sympy matrices
G0 = smp.Matrix(G0) * smp.Matrix(Ainv) * BL * R
G1 = smp.Matrix(G1) * smp.Matrix(Ainv) * BL * R
return G0, G1
def icb_interface_flux_matrix(p, K, T):
"""The interface flux matrices for the ICB schemes, we use upwinding
to get the fluxes. Denote T the translation necessary to
evaluate the flux from the other cell.
"""
# Enhanced polynomial degree
phat = p + len(K)
G0 = smp.zeros(p + 1, phat + 1)
G1 = smp.zeros(p + 1, phat + 1)
for i in range(0, p + 1):
for j in range(0, phat + 1):
li = L.basis(i)
lj = L.basis(j)
G0[i, j] = leg.legval(-1, li.coef) * \
leg.legval(1, lj.coef) * (T**-1)
G1[i, j] = leg.legval(1, li.coef) * leg.legval(1, lj.coef)
# Enhancement matrix
A, Ainv, B, Binv = enhance.enhancement_matrices(p, K)
# Using the enhanced function in the flux (see notes 21/4/15)
BL = smp.zeros(phat + 1, phat + 1)
BR = smp.zeros(phat + 1, phat + 1)
for i in range(p + 1):
li = L.basis(i)
BL[i, i] = 1 # basis.integrate_legendre_product(li,li)
BR[i, i] = 1 # BL[i,i]
for i, k in enumerate(K):
lk = L.basis(k)
int_lklk = basis.integrate_legendre_product(lk, lk)
BL[i + p + 1, i + p + 1] = T # * int_lklk
BR[i + p + 1, i + p + 1] = (T**(-1)) # * int_lklk
# reduction matrix
R = smp.zeros(phat + 1, p + 1)
for i in range(p + 1):
R[i, i] = 1
for i, k in enumerate(K):
for j in range(p + 1):
R[i + p + 1, j] = auxf.delta(k, j)
# Convert to sympy matrices
G0 = smp.Matrix(G0) * smp.Matrix(Ainv) * BL * R
G1 = smp.Matrix(G1) * smp.Matrix(Ainv) * BL * R
return G0, G1
def trend(p, t):
"""
Fit the local transit shape with the following function.
"""
domain = [t.min(), t.max()]
return Legendre(p, domain=domain)(t)
def evaluate_basis_gauss(self):
"""Evaluate the basis at the Gaussian quadrature nodes.
phi will be used to transform Legendre solution coefficients
to the solution evaluated at the Gaussian quadrature nodes.
dphi_w will be used for the interior flux integral
"""
phi = np.zeros((len(self.x), self.N_s))
dphi_w = np.zeros((len(self.x), self.N_s))
for n in range(self.N_s):
# Get the Legendre polynomial of order n and its gradient
l = L.basis(n)
dl = l.deriv()
# Evaluate the basis at the Gaussian nodes
phi[:, n] = leg.legval(self.x, l.coef)
# Evaluate the gradient at the Gaussian nodes and multiply by the
# weights
dphi_w[n, :] = leg.legval(self.x, dl.coef) * self.w
return phi, dphi_w
def evaluate_basis_gauss(self):
"""Evaluate the basis at the Gaussian quadrature nodes.
phi will be used to transform Legendre solution coefficients
to the solution evaluated at the Gaussian quadrature nodes.
dphi_w will be used for the interior flux integral
"""
phi = np.zeros((len(self.x), self.N_s))
dphi_w = np.zeros((len(self.x), self.N_s))
for n in range(self.N_s):
# Get the Legendre polynomial of order n and its gradient
l = L.basis(n)
dl = l.deriv()
# Evaluate the basis at the Gaussian nodes
phi[:, n] = leg.legval(self.x, l.coef)
# Evaluate the gradient at the Gaussian nodes and multiply by the
# weights
dphi_w[n, :] = leg.legval(self.x, dl.coef) * self.w
return phi, dphi_w
def regress_poly(degree, data, remove_mean=True, axis=-1):
''' returns data with degree polynomial regressed out.
Be default it is calculated along the last axis (usu. time).
If remove_mean is True (default), the data is demeaned (i.e. degree 0).
If remove_mean is false, the data is not.
'''
IFLOG.debug('Performing polynomial regression on data of shape ' +
str(data.shape))
datashape = data.shape
timepoints = datashape[axis]
# Rearrange all voxel-wise time-series in rows
data = data.reshape((-1, timepoints))
# Generate design matrix
X = np.ones((timepoints, 1)) # quick way to calc degree 0
for i in range(degree):
polynomial_func = Legendre.basis(i + 1)
value_array = np.linspace(-1, 1, timepoints)
X = np.hstack((X, polynomial_func(value_array)[:, np.newaxis]))
# Calculate coefficients
betas = np.linalg.pinv(X).dot(data.T)
# Estimation
if remove_mean:
datahat = X.dot(betas).T
else: # disregard the first layer of X, which is degree 0
datahat = X[:, 1:].dot(betas[1:, ...]).T
regressed_data = data - datahat
# Back to original shape
return regressed_data.reshape(datashape)
def regress_poly(degree, data, remove_mean=True, axis=-1):
''' returns data with degree polynomial regressed out.
Be default it is calculated along the last axis (usu. time).
If remove_mean is True (default), the data is demeaned (i.e. degree 0).
If remove_mean is false, the data is not.
'''
IFLOG.debug('Performing polynomial regression on data of shape ' + str(data.shape))
datashape = data.shape
timepoints = datashape[axis]
# Rearrange all voxel-wise time-series in rows
data = data.reshape((-1, timepoints))
# Generate design matrix
X = np.ones((timepoints, 1)) # quick way to calc degree 0
for i in range(degree):
polynomial_func = Legendre.basis(i + 1)
value_array = np.linspace(-1, 1, timepoints)
X = np.hstack((X, polynomial_func(value_array)[:, np.newaxis]))
# Calculate coefficients
betas = np.linalg.pinv(X).dot(data.T)
# Estimation
if remove_mean:
datahat = X.dot(betas).T
else: # disregard the first layer of X, which is degree 0
datahat = X[:, 1:].dot(betas[1:, ...]).T
regressed_data = data - datahat
# Back to original shape
return regressed_data.reshape(datashape)
def enhancement_matrices(solution_order, modes):
"""Returns the enhancement matrices (and their inverse)
Returns A and inv(A) where A \hat{u} = [uL;some_modes_of(uR)]
B and inv(B) where B \hat{u} = [uR;some_modes_of(uL)]
Note: this is slightly different than what I do in
icb_functions.py (called by advection.py) where the right hand
side contains the normalization factors (i.e A x = b where b =
uL_i \int \phi_i \phi_i dx). Here I put \int \phi_i \phi_i dx into
A and B (denoted norm down in the code below).
"""
# Enhanced solution order
order = solution_order + len(modes)
# Submatrices to build the main matrix later
a = np.diag(np.ones(solution_order + 1))
b = np.zeros((solution_order + 1, len(modes)))
cl = np.zeros((len(modes), order + 1))
cr = np.zeros((len(modes), order + 1))
# Loop on the modes we are keeping in the neighboring cell
# (the right cell)
for i, mode in enumerate(modes):
# Loop on the enhancement basis
for j in range(order + 1):
# Basis function in the right cell
l1 = L.basis(mode)
# Enhanced basis function extending into the right cell (or left
# cell)
ll = basis.shift_legendre_polynomial(L.basis(j), 2)
lr = basis.shift_legendre_polynomial(L.basis(j), -2)
# Inner product for the left and right enhancements
norm = basis.integrate_legendre_product(l1, l1)
cl[i, j] = basis.integrate_legendre_product(l1, ll) / norm
cr[i, j] = basis.integrate_legendre_product(l1, lr) / norm
# Put the matrices together
A = np.vstack((np.hstack((a, b)), cl))
B = np.vstack((np.hstack((a, b)), cr))
return A, np.linalg.inv(A), B, np.linalg.inv(B)
def enhancement_matrices(solution_order, modes):
"""Returns the enhancement matrices (and their inverse)
Returns A and inv(A) where A \hat{u} = [uL;some_modes_of(uR)]
B and inv(B) where B \hat{u} = [uR;some_modes_of(uL)]
Note: this is slightly different than what I do in
icb_functions.py (called by advection.py) where the right hand
side contains the normalization factors (i.e A x = b where b =
uL_i \int \phi_i \phi_i dx). Here I put \int \phi_i \phi_i dx into
A and B (denoted norm down in the code below).
"""
# Enhanced solution order
order = solution_order + len(modes)
# Submatrices to build the main matrix later
a = np.diag(np.ones(solution_order + 1))
b = np.zeros((solution_order + 1, len(modes)))
cl = np.zeros((len(modes), order + 1))
cr = np.zeros((len(modes), order + 1))
# Loop on the modes we are keeping in the neighboring cell
# (the right cell)
for i, mode in enumerate(modes):
# Loop on the enhancement basis
for j in range(order + 1):
# Basis function in the right cell
l1 = L.basis(mode)
# Enhanced basis function extending into the right cell (or left
# cell)
ll = basis.shift_legendre_polynomial(L.basis(j), 2)
lr = basis.shift_legendre_polynomial(L.basis(j), -2)
# Inner product for the left and right enhancements
norm = basis.integrate_legendre_product(l1, l1)
cl[i, j] = basis.integrate_legendre_product(l1, ll) / norm
cr[i, j] = basis.integrate_legendre_product(l1, lr) / norm
# Put the matrices together
A = np.vstack((np.hstack((a, b)), cl))
B = np.vstack((np.hstack((a, b)), cr))
return A, np.linalg.inv(A), B, np.linalg.inv(B)
def create(self, remake=False):
'''
Populate the wavelength calibration for this aperture,
using input from the user to try to link up lines between
the measured arc spectra and the known line wavelengths.
'''
self.speak("populating wavelength calibration")
# loop through until we've decided we're converged
self.notconverged = True
while (self.notconverged):
# set an initial guess matching wavelengths to my pixles
#self.guessMatches()
# do a fit
if self.justloaded:
self.justloaded = False
else:
# do an initial fit
self.pixelstowavelengths = Legendre.fit(
x=self.pixel,
y=self.wavelength,
deg=self.polynomialdegree,
w=self.weights)
# identify outliers
limit = 1.48 * craftroom.oned.mad(self.residuals[self.good]) * 4
limit = np.maximum(limit, 1.0)
# outliers get reset each time (so don't lose at edges)
outlier = np.abs(self.residuals) > limit
# keep track of which are outlier
for i, m in enumerate(self.matches):
self.matches[i]['mine']['outlier'] = outlier[i]
self.speak('points beyond {0} are outliers ({1})'.format(limit, i))
self.pixelstowavelengths = Legendre.fit(x=self.pixel,
y=self.wavelength,
deg=self.polynomialdegree,
w=self.weights)
self.plotWavelengthFit()
def dg_interface_flux_matrix(p, T):
"""The interface flux matrices, we use upwinding to get the fluxes
Denote T the translation necessary to evaluate the flux from the
other cell.
There are also analytical expressions for these. See my thesis page 73.
"""
G0 = smp.zeros(p + 1)
G1 = smp.zeros(p + 1)
for i in range(0, p + 1):
for j in range(0, p + 1):
li = L.basis(i)
lj = L.basis(j)
G0[i, j] = leg.legval(-1, li.coef) * \
leg.legval(1, lj.coef) * (T**-1)
G1[i, j] = leg.legval(1, li.coef) * leg.legval(1, lj.coef)
return G0, G1
def dg_interface_flux_matrix(p, T):
"""The interface flux matrices, we use upwinding to get the fluxes
Denote T the translation necessary to evaluate the flux from the
other cell.
There are also analytical expressions for these. See my thesis page 73.
"""
G0 = smp.zeros(p + 1)
G1 = smp.zeros(p + 1)
for i in range(0, p + 1):
for j in range(0, p + 1):
li = L.basis(i)
lj = L.basis(j)
G0[i, j] = leg.legval(-1, li.coef) * \
leg.legval(1, lj.coef) * (T**-1)
G1[i, j] = leg.legval(1, li.coef) * leg.legval(1, lj.coef)
return G0, G1
def shift_legendre_polynomial(l, shift):
"""Returns the Legendre polynomial shifted by a certain amount
Basically, given a Legendre polynomial l(x), return l(x+shift)
"""
# Shift the window of the polynomial to get the new coefficients
ls = L.cast(l, window=l.window - shift)
# Create a Legendre polynomial with these new coefficients and the
# default window
return L(ls.coef)
def test_shift_legendre_polynomial(self):
"""Is the shifting of Legendre polynomials correct"""
# Given a Legendre Polynomial: L(x) = 0.5*(3x^2 -1)
l = L.basis(2)
# Evaluate L(x+2)
ls = basis.shift_legendre_polynomial(l, 2)
# This should be equal to 5.5 + 6x + 1.5 x^2
npt.assert_array_almost_equal(ls.convert(
kind=P).coef, np.array([5.5, 6, 1.5]), decimal=13)
def test_integrate_legendre_product(self):
"""Is the integral of the product of two Legendre polynomials correct
"""
# Given a Legendre Polynomial: L(x) = 0.5*(3x^2 -1)
l1 = L.basis(2)
# Evaluate L(x+2)
l2 = basis.shift_legendre_polynomial(l1, 2)
# The integral of l1*l2 over [-1,1] = 0.4
self.assertAlmostEqual(basis.integrate_legendre_product(l1, l2), 0.4)
def shift_legendre_polynomial(l, shift):
"""Returns the Legendre polynomial shifted by a certain amount
Basically, given a Legendre polynomial l(x), return l(x+shift)
"""
# Shift the window of the polynomial to get the new coefficients
ls = L.cast(l, window=l.window - shift)
# Create a Legendre polynomial with these new coefficients and the
# default window
return L(ls.coef)
def test_integrate_legendre_product(self):
"""Is the integral of the product of two Legendre polynomials correct
"""
# Given a Legendre Polynomial: L(x) = 0.5*(3x^2 -1)
l1 = L.basis(2)
# Evaluate L(x+2)
l2 = basis.shift_legendre_polynomial(l1, 2)
# The integral of l1*l2 over [-1,1] = 0.4
self.assertAlmostEqual(basis.integrate_legendre_product(l1, l2), 0.4)
def nodal_flux_builder_edge(coefficient, diff1, diff2):
flux_legendre1 = Legendre(coefficient[:5])
flux_legendre2 = Legendre(coefficient[5:])
flux_legendre1 = flux_legendre1.deriv(1)
flux_legendre2 = flux_legendre2.deriv(1)
flux1 = flux_legendre1.linspace(65, [0, 10])
flux2 = flux_legendre2.linspace(65, [10, 20])
total_flux = np.concatenate((diff1 * flux1[1], diff2 * flux2[1]))
flux_position = np.concatenate((flux1[0], flux2[0]))
return total_flux, flux_position
def genNoise(self,grid,maxNoiseOrder):
"""Noise is a matrix of Legendre polynomials of 0<order<maxNoiseOrder
Additionally 60Hz sine and cosine waves are added to account for the DC component of EEG
grid -- Grid to be used for timing information
maxNoiseOrder--Maximum order of noise to be considered"""
if self.grid() is not None and self.noiseOrders() is not None:
logger.info( 'Generating noise matrix')
legpoly = np.array([Legendre.basis(i)(np.arange(len(grid.times()))) for i in self.noiseOrders()]).T #Polynomials
sw = np.sin(60 * np.arange(len(grid.times())) * 2 * np.pi / float(grid.fs())) # Sine for AC component
cw = np.cos(60 * np.arange(len(grid.times())) * 2 * np.pi / float(grid.fs())) # Cosine for AC component
legpoly = np.column_stack((legpoly, sw, cw))
return pd.DataFrame(legpoly,index=self.grid().times())
def test_shift_legendre_polynomial(self):
"""Is the shifting of Legendre polynomials correct"""
# Given a Legendre Polynomial: L(x) = 0.5*(3x^2 -1)
l = L.basis(2)
# Evaluate L(x+2)
ls = basis.shift_legendre_polynomial(l, 2)
# This should be equal to 5.5 + 6x + 1.5 x^2
npt.assert_array_almost_equal(ls.convert(kind=P).coef,
np.array([5.5, 6, 1.5]),
decimal=13)
def LDT(epoch,tdur,t,f,pad=0.2,deg=1):
"""
Local detrending
A simple function that subtracts a polynomial trend from the
lightcurve excluding a region around the transit.
pad : Extra number of days to notch out of the of the transit region.
"""
bcont = abs(t - epoch) > tdur/2 + pad
fcont = f[bcont]
tcont = t[bcont]
legtrend = Legendre.fit(tcont,fcont,deg,domain=[t.min(),t.max()])
trend = legtrend(t)
return trend
def nodal_flux_builder_cell(coefficient):
flux_legendre1 = Legendre(coefficient[:5])
flux_legendre2 = Legendre(coefficient[5:])
flux1 = flux_legendre1.linspace(64, [0, 10])
flux2 = flux_legendre2.linspace(64, [10, 20])
total_flux = np.concatenate((flux1[1], flux2[1]))
flux_position = np.concatenate((flux1[0], flux2[0]))
return total_flux, flux_position
def genNoise(self, grid, maxNoiseOrder):
"""Noise is a matrix of Legendre polynomials of 0<order<maxNoiseOrder
Additionally 60Hz sine and cosine waves are added to account for the DC component of EEG
grid -- Grid to be used for timing information
maxNoiseOrder--Maximum order of noise to be considered"""
if self.grid() is not None and self.noiseOrders() is not None:
logger.info('Generating noise matrix')
legpoly = np.array([
Legendre.basis(i)(np.arange(len(grid.times())))
for i in self.noiseOrders()
]).T #Polynomials
sw = np.sin(60 * np.arange(len(grid.times())) * 2 * np.pi /
float(grid.fs())) # Sine for AC component
cw = np.cos(60 * np.arange(len(grid.times())) * 2 * np.pi /
float(grid.fs())) # Cosine for AC component
legpoly = np.column_stack((legpoly, sw, cw))
return pd.DataFrame(legpoly, index=self.grid().times())
def regress_poly(degree, data, remove_mean=True, axis=-1):
"""
Returns data with degree polynomial regressed out.
:param bool remove_mean: whether or not demean data (i.e. degree 0),
:param int axis: numpy array axes along which regression is performed
"""
timepoints = data.shape[0]
# Generate design matrix
X = np.ones((timepoints, 1)) # quick way to calc degree 0
for i in range(degree):
polynomial_func = Legendre.basis(i + 1)
value_array = np.linspace(-1, 1, timepoints)
X = np.hstack((X, polynomial_func(value_array)[:, np.newaxis]))
non_constant_regressors = X[:, :-1] if X.shape[1] > 1 else np.array([])
betas = np.linalg.pinv(X).dot(data)
if remove_mean:
datahat = X.dot(betas)
else: # disregard the first layer of X, which is degree 0
datahat = X[:, 1:].dot(betas[1:, ...])
regressed_data = data - datahat
return regressed_data, non_constant_regressors
def approx_legendre_poly(Moments):
n_moments = Moments.shape[0]-1
exp_coef = (np.zeros((1)))
# For method description see, for instance:
# Chapter 3 of "The Problem of Moments", James Alexander Shohat, Jacob David Tamarkin
for i in range(n_moments+1):
p = Legendre.basis(i).convert(window = [0.0,1.0], kind=Polynomial)
q = (2*i+1)*np.sum(Moments[0:(i+1)]*p.coef)
pq = (p.coef*q)
exp_coef = polynomial.polyadd(exp_coef, pq)
expansion = Polynomial(exp_coef)
return expansion
def regress_poly(degree, data, remove_mean=True, axis=-1):
"""
Returns data with degree polynomial regressed out.
:param bool remove_mean: whether or not demean data (i.e. degree 0),
:param int axis: numpy array axes along which regression is performed
"""
IFLOGGER.debug('Performing polynomial regression on data of shape %s',
str(data.shape))
datashape = data.shape
timepoints = datashape[axis]
# Rearrange all voxel-wise time-series in rows
data = data.reshape((-1, timepoints))
# Generate design matrix
X = np.ones((timepoints, 1)) # quick way to calc degree 0
for i in range(degree):
polynomial_func = Legendre.basis(i + 1)
value_array = np.linspace(-1, 1, timepoints)
X = np.hstack((X, polynomial_func(value_array)[:, np.newaxis]))
non_constant_regressors = X[:, :-1] if X.shape[1] > 1 else np.array([])
# Calculate coefficients
betas = np.linalg.pinv(X).dot(data.T)
# Estimation
if remove_mean:
datahat = X.dot(betas).T
else: # disregard the first layer of X, which is degree 0
datahat = X[:, 1:].dot(betas[1:, ...]).T
regressed_data = data - datahat
# Back to original shape
return regressed_data.reshape(datashape), non_constant_regressors
def regress_poly(degree, data, remove_mean=True, axis=-1):
"""
Returns data with degree polynomial regressed out.
:param bool remove_mean: whether or not demean data (i.e. degree 0),
:param int axis: numpy array axes along which regression is performed
"""
IFLOGGER.debug('Performing polynomial regression on data of shape %s',
str(data.shape))
datashape = data.shape
timepoints = datashape[axis]
# Rearrange all voxel-wise time-series in rows
data = data.reshape((-1, timepoints))
# Generate design matrix
X = np.ones((timepoints, 1)) # quick way to calc degree 0
for i in range(degree):
polynomial_func = Legendre.basis(i + 1)
value_array = np.linspace(-1, 1, timepoints)
X = np.hstack((X, polynomial_func(value_array)[:, np.newaxis]))
non_constant_regressors = X[:, :-1] if X.shape[1] > 1 else np.array([])
# Calculate coefficients
betas = np.linalg.pinv(X).dot(data.T)
# Estimation
if remove_mean:
datahat = X.dot(betas).T
else: # disregard the first layer of X, which is degree 0
datahat = X[:, 1:].dot(betas[1:, ...]).T
regressed_data = data - datahat
# Back to original shape
return regressed_data.reshape(datashape), non_constant_regressors
def specpix_to_wavelength(specpix, tracepars, m=1, frame='dms', oversample=1):
"""Convert the spectral pixel coordinate to wavelength for order m.
:param specpix: the pixel values.
:param tracepars: the trace polynomial solutions returned by get_tracepars.
:param m: the spectral order.
:param frame: the coordinate frame of the input coordinates (nat, dms or sim).
:param oversample: the oversampling factor of the input coordinates.
:type specpix: array[float]
:type tracepars: dict
:type m: int
:type frame: str
:type oversample: int
:returns: wavelength - an array containing the wavelengths corresponding to specpix,
mask - an array that is True when the specpix values were within the valid range of the polynomial.
:rtype: Tuple(array[float], array[bool])
"""
# Remove any oversampling.
specpix = specpix / oversample
# Convert the input coordinates to nat coordinates.
specpix_nat = specpix_frame_to_ref(specpix, frame=frame)
# Convert the specpix coordinates to wavelength.
spec2w = Legendre(tracepars[m]['wave_coef'],
domain=tracepars[m]['wave_domain'])
with np.errstate(over='ignore'):
wavelength = np.exp(spec2w(specpix_nat))
mask = bounds_check(specpix_nat, tracepars[m]['wave_domain'][0],
tracepars[m]['wave_domain'][1])
return wavelength, mask
#starmaster = np.load('/h/mulan0/data/working/GJ1132b_ut160421_22/multipleapertures/aperture_587_1049/extracted'+mastern+'.npy')[()]
#oef, domain = np.load('/h/mulan0/data/working/GJ1132b_ut160421_22/multipleapertures/aperture_587_1049/aperture_587_1049_wavelengthcalibration.npy')[()]
# this is a check to make sure you have commented in the correct star (this is obviously a hack that could be done better...)
response = raw_input(
'Is this the correct starmaster for this dataset? [y/n] \n ' +
starmasterstr + '\n')
if response == 'n':
sys.exit(
'Please edit wavelength_recalibration.py to point to the correct starmaster'
)
elif response == 'y':
starmaster = np.load(starmasterstr)[()]
# reacreate the wavelength solution (way to go from pixel space to wavelength space) from mosasaurus
pxtowavemaster = Legendre(coef, domain)
# apertures form a given night (Need to have run mosasaurus in ipython or similar and then copy and paste this monster code in. This is not ideal.)
apertures = r.mask.apertures
makeplot = True
UV_poses, O2_poses, Ca1_poses, Ca2_poses, Ca3_poses, H2O_poses = [], [], [], [], [], []
x_poses = []
#shifts_1132 = []
# Fixed alignment ranges for each prominent freature
align_UV = (6870, 6900)
align_O2 = (7580, 7650)
#align_H2Osmall = (8220, 8260)
align_Ca1 = (8490, 8525)
align_Ca2 = (8535, 8580)
align_Ca3 = (8650, 8700)
align_H2O = (9300, 9700)
for n in r.obs.nScience:
def trace_polynomial(trace, m=1, maxorder=15):
"""Fit a polynomial to the trace of order m and return a
dictionary containing the parameters and validity intervals.
:param trace: astropy table containing modelled trace points.
:param m: spectral order for which to fit a polynomial.
:param maxorder: maximum polynomial order to use.
:type trace: astropy.table.Table
:type m: int
:type maxorder: int
:returns: pars - dictionary containg the polynomial solution.
:rtype: dict
"""
# TODO added arbitrary maxorder to deal with poor exrapolatian, revisit when extrapolation fixed.
# Select the data for order m.
mask = (trace['order'] == m)
wave = trace['Wavelength'][mask]
spatpix_ref = trace['xpos'][mask]
specpix_ref = trace['ypos'][mask]
# Find the edges of the domain.
wavemin = np.amin(wave)
wavemax = np.amax(wave)
specmin = np.amin(specpix_ref)
specmax = np.amax(specpix_ref)
# Compute the polynomial parameters for x and y.
order = 0
while order <= maxorder:
spatpol = Legendre.fit(np.log(wave), spatpix_ref, order)
specpol = Legendre.fit(np.log(wave), specpix_ref, order)
spatpixp_nat = spatpol(np.log(wave))
specpixp_nat = specpol(np.log(wave))
if np.all(np.abs(spatpix_ref - spatpixp_nat) < 0.5) & np.all(
np.abs(specpix_ref - specpixp_nat) < 0.5):
break
order += 1
# Compute the transform back to wavelength.
wavegrid = wavemin + (wavemax - wavemin) * np.linspace(0., 1., 501)
specgrid = specpol(np.log(wavegrid))
wavepol = Legendre.fit(specgrid, np.log(wavegrid), order)
# Add the parameters to a dictionary.
pars = dict()
pars['spat_coef'] = spatpol.coef
pars['spat_domain'] = spatpol.domain
pars['spec_coef'] = specpol.coef
pars['spec_domain'] = specpol.domain
pars['wave_coef'] = wavepol.coef
pars['wave_domain'] = wavepol.domain
return pars
class WavelengthCalibrator(Talker):
def __init__(self, aperture,
elements=['He', 'Ne','Ar'],
polynomialdegree=3,
matchdistance=100):
Talker.__init__(self)
# keep track of the aperture this belongs to
self.aperture = aperture
# set up the some defaults
self.elements = elements
self.polynomialdegree = polynomialdegree
self.matchdistance = matchdistance
# either load a previous wavecal, or create a new one
self.populate()
@property
def wavelengthprefix(self):
return self.aperture.directory + '{0}_'.format(self.aperture.name)
@property
def calibrationfilename(self):
return self.wavelengthprefix + 'wavelengthcalibration.npy'
@property
def waveidfilename(self):
return self.wavelengthprefix + 'waveids.txt'
def loadWavelengthIdentifications(self, restart=False):
''' Try loading a custom stored wavelength ID file
from the aperture's directory, and if that
doesn't work, load the default one for this grism.'''
try:
# try to load a custom wavelength id file
assert(restart == False)
self.speak('checking for a custom wavelength-to-pixel file')
self.waveids = astropy.io.ascii.read(self.waveidfilename)
# keep track of which waveid file is used
self.whichwaveid = 'Aperture-Specific ({0})'.format(
self.waveidfilename.split('/')[-1])
except (IOError,AssertionError):
self.speak('no custom wavelength-to-pixel files found')
# load the default for this grism, as set in obs. file
d = astropy.io.ascii.read(self.aperture.obs.wavelength2pixelsFile)
self.rawwaveids = d[['pixel', 'wavelength', 'name']]
self.rawwaveids['pixel'] /= self.aperture.obs.binning
# use a cross-corrlation to find the rough offset
# (the function call will define waveids)
# pull out the peaks from extracted arc spectra
self.findPeaks()
# perform a first rough alignment, to make pixels
self.findRoughShift()
# keep track of whcih waveid file is used
self.whichwaveid = 'Default ({0})'.format(
self.aperture.obs.wavelength2pixelsFile.split('/')[-1])
self.speak('loaded {0}:'.format(self.whichwaveid))
self.findKnownWavelengths()
self.guessMatches()
def saveWavelengthIdentification(self):
'''store the wavelength-to-pixel identifications'''
self.waveids.write( self.waveidfilename,
format='ascii.fixed_width',
delimiter='|',
bookend=False)
self.speak('saved wavelength-to-pixel matches to '+self.waveidfilename)
def save(self):
self.speak('saving all aspects of the wavelength calibration')
# the wavelength identifications
self.saveWavelengthIdentification()
# the actual matches used (with outliers, etc...)
self.saveMatches()
# save the coefficients (and domain) of the calibration
self.saveCalibration()
# save the figure
self.figcal.savefig(self.wavelengthprefix + 'calibration.pdf')
def loadCalibration(self):
coef, domain = np.load(self.calibrationfilename)
self.pixelstowavelengths = Legendre(coef, domain)
self.polynomialdegree = self.pixelstowavelengths.degree()
self.speak("loaded wavelength calibration"
"from {0}".format(self.calibrationfilename))
def saveCalibration(self):
np.save(self.calibrationfilename, (self.pixelstowavelengths.coef, self.pixelstowavelengths.domain))
self.speak("saved wavelength calibration coefficients to {0}".format(self.calibrationfilename))
def populate(self, restart=False):
# populate the wavelength identifications
self.loadWavelengthIdentifications(restart=restart)
try:
# populate the wavelength calibration polynomial
assert(restart==False)
self.loadCalibration()
self.justloaded = True
self.loadMatches()
#self.plotWavelengthFit(interactive=False)
#unhappy = ('n' in self.input('Are you happy with the wavelength calibration? [Y,n]').lower())
#assert(unhappy == False)
except (IOError, AssertionError):
self.justloaded = False
self.create()
def findRoughShift(self, blob=2.0):
'''using a list of pixels matched to wavelengths, find the rough
offset of the this arc relative to the standard slits/setting'''
self.speak("cross correlating arcs with known wavelengths")
# create a plot showing how well the lines match
figure_waverough = plt.figure( 'wavelength rough offset',
figsize=(6,4), dpi=100)
gs = plt.matplotlib.gridspec.GridSpec( 3,2,
bottom=0.15, top=0.85,
hspace=0.1,wspace=0,
width_ratios=[1, .5])
# make the axes for checking the rough alignment
self.ax_waverough, self.ax_wavecor = {}, {}
sharer, sharec = None, None
for i, e in enumerate(self.elements):
self.ax_waverough[e] = plt.subplot(gs[i,0],sharex=sharer)
self.ax_wavecor[e] = plt.subplot(gs[i,1], sharex=sharec)
sharer, sharec = self.ax_waverough[e], self.ax_wavecor[e]
# calculate correlation functions
self.corre = {}
for count, element in enumerate(self.elements):
# pull out the peaks
xPeak = [p['w'] for p in self.peaks[element]]
yPeak = [p['intensity'] for p in self.peaks[element]]
# create fake spectra using the line positions (reference + new)
x = np.arange(-self.aperture.obs.ysize,self.aperture.obs.ysize)
myPeaks, theirPeaks = np.zeros(len(x)), np.zeros(len(x))
# create fake spectrum of their peaks
for i in range(len(self.rawwaveids)):
if element in self.rawwaveids['name'][i]:
center = self.rawwaveids['pixel'][i]
theirPeaks += np.exp(-0.5*((x-center)/blob)**2)
# create fake spectrum of my peaks
for i in range(len(xPeak)):
center = xPeak[i]
myPeaks += np.exp(-0.5*((x-center)/blob)**2)*np.log(yPeak[i])
# calculate the correlation function for this element
self.corre[element] = np.correlate(myPeaks, theirPeaks, 'full')
# plot the rough shift and identifications
self.ax_waverough[element].plot(x, myPeaks/myPeaks.max(),
label='extracted', alpha=0.5,
color=colors[element])
self.ax_waverough[element].set_ylim(0, 2)
# plot the correlation functions
normcor = self.corre[element]
assert(np.isfinite(self.corre[element]).any())
normcor /= np.nanmax(self.corre[element])
self.ax_wavecor[element].plot(normcor, label=element, alpha=0.5,
color=colors[element])
# tidy up the plots
for a in [self.ax_wavecor, self.ax_waverough]:
plt.setp(a[element].get_xticklabels(), visible=False)
plt.setp(a[element].get_yticklabels(), visible=False)
# multiply the correlation functions together
assert(np.isfinite(self.corre[element]).any())
if count == 0:
self.corre['combined'] = np.ones_like(self.corre[element])
self.corre['combined'] *= self.corre[element]
# find the peak of the combined correlation function
self.peakoffset = -1024 # KLUDGE KLUDGE KLUDGE! np.where(self.corre['combined'] == self.corre['combined'].max())[0][0] - len(x)
# (old?) to convert: len(x) - xPeak = x + peakoffset
# define the new, shifted, waveids array
self.waveids = copy.deepcopy(self.rawwaveids)
self.waveids['pixel'] += self.peakoffset
# plot the shifted wavelength ids, and combined corfuncs
for element in self.elements:
for i in range(len(self.rawwaveids)):
if element in self.rawwaveids['name'][i]:
center = self.waveids['pixel'][i]
self.ax_waverough[element].axvline(center, alpha=0.25, color='black')
# plot the combined correlation function
normedcombined = self.corre['combined']/np.max(self.corre['combined'])
self.ax_wavecor[element].plot(normedcombined,
label='combined', alpha=0.25, color='black')
# tidy up the plots
self.ax_waverough[element].set_ylabel(element)
ax = self.ax_waverough[self.elements[-1]]
plt.setp(ax.get_xticklabels(), visible=True)
ax.set_xlabel('Pixel Position')
fontsize = 8
self.ax_wavecor[self.elements[0]].set_title('cross correlation peaks at \n{0} pixels ({1}x{1} binned pixels)'.format(self.peakoffset, self.aperture.obs.binning), fontsize=fontsize)
self.ax_waverough[self.elements[0]].set_title(
'Coarse Wavelength Alignment\nfor ({0:0.1f},{1:0.1f})'.format(
self.aperture.x, self.aperture.y),fontsize=fontsize)
# save the figure
figure_waverough.savefig(
self.aperture.directory + 'roughWavelengthAlignment_{0}.pdf'.format(
self.aperture.name))
@property
def peaks(self):
try:
return self._peaks
except AttributeError:
self.findPeaks()
return self._peaks
def findPeaks(self):
'''identify peaks in the extracted arc spectrum'''
# extract a spectrum from the master image for each lamp
self.aperture.arcs = {}
for element in self.elements:
self.aperture.arcs[element] = self.aperture.extract(
n=element,
image=self.aperture.images[element],
arc=True)
# find the peaks in my spectra
self._peaks = {}
for count, element in enumerate(self.elements):
# the pixel spectrum self.aperture.extracted from this arc lamp
width = np.min(self.aperture.trace.extractionwidths)
flux = self.aperture.arcs[element][width]['raw_counts']
# identify my peaks
xPeak, yPeak, xfiltered, yfiltered = zachopy.oned.peaks(
self.aperture.waxis,
flux,
plot=False,
xsmooth=30,
threshold=100,
edgebuffer=10,
widthguess=1,
maskwidth=3,
returnfiltered=True)
self.aperture.arcs[element]['filtered'] = yfiltered
# for some reason, need to trim peaks outside range
pad = 25
toright = xPeak > (np.min(self.aperture.waxis) + pad)
toleft = xPeak < (np.max(self.aperture.waxis) - pad)
ok = toleft*toright
xPeak, yPeak = xPeak[ok], yPeak[ok]
n = len(xPeak)
# store those peaks
self._peaks[element] = []
for i in range(n):
peak = {
'w':xPeak[i],
'intensity':yPeak[i],
'handpicked':False,
'outlier':False
}
self._peaks[element].append(peak)
def findKnownWavelengths(self):
# create a temporary calibration to match reference wavelengths to reference pixels (so we can extrapolate to additional wavelengths not recorded in the dispersion solution file)
self.knownwavelengths = {}
# treat the arc lamps separately
for count, element in enumerate(self.elements):
# pull out the wavelengths from the complete file
self.knownwavelengths[element] = []
for i in range(len(self.waveids)):
if element in self.waveids['name'][i]:
wave = self.waveids['wavelength'][i]
pixel = self.waveids['pixel'][i]
known = {
'element':element,
'wavelength':wave,
'pixelguess':pixel
}
self.knownwavelengths[element].append(known)
@property
def matchesfilename(self):
return self.wavelengthprefix + 'wavelengthmatches.npy'
def saveMatches(self):
self.speak('saving wavelength dictionaries to {}'.format(self.matchesfilename))
np.save(self.matchesfilename, (self.matches, self.knownwavelengths))
def loadMatches(self):
(self.matches, self.knownwavelengths) = np.load(self.matchesfilename)
self.speak('loaded wavelength matches from {0}'.format(self.matchesfilename))
def guessMatches(self):
self.matches = []
# do identification with one arc at a time
for element in self.elements:
# pull out my peaks and theirs
myPeaks = np.array([p['w'] for p in self.peaks[element]])
theirPeaksOnMyPixels = np.array([known['pixelguess']
for known
in self.knownwavelengths[element]])
thiselement = []
# loop over my peaks
for i in range(len(myPeaks)):
# find my closest peak to theirs
distance = myPeaks[i] - theirPeaksOnMyPixels
closest = np.sort(np.nonzero(np.abs(distance) == np.min(np.abs(distance)))[0])[0]
if distance[closest] < self.matchdistance:
# call this a match (no matter how far)
match = {
'mine':self.peaks[element][i],
'theirs':self.knownwavelengths[element][closest],
'distance':distance[closest]
}
thiselement.append(match)
self.matches.extend(thiselement)
'''for theirs in theirPeaksOnMyPixels:
relevant = [m for m in thiselement if
m['theirs']['pixelguess'] == theirs]
if len(relevant) > 0:
print "{0} of my peaks match to their {1}".format(len(relevant), theirs)
distances = np.abs([m['theirs']['pixelguess'] - m['mine']['w'] for m in relevant])
best = distances.argmin()
print "the closests one is {0}".format(relevant[best]['mine']['w'])
self.matches.append(relevant[best])
'''
# add this to the list
#self.matches.extend(thiselement)
def create(self, remake=False):
'''Populate the wavelength calibration for this aperture.'''
self.speak("populating wavelength calibration")
# loop through
self.notconverged = True
while(self.notconverged):
# set an initial guess matching wavelengths to my pixles
#self.guessMatches()
# do a fit
if self.justloaded:
self.justloaded = False
else:
# do an initial fit
self.pixelstowavelengths = Legendre.fit(
x=self.pixel,
y=self.wavelength,
deg=self.polynomialdegree,
w=self.weights
)
# identify outliers
limit = 1.48*zachopy.oned.mad(self.residuals[self.good])*4
limit = np.maximum(limit, 1.0)
# outliers get reset each time (so don't lose at edges)
outlier = np.abs(self.residuals) > limit
# keep track of which are outlier
for i, m in enumerate(self.matches):
self.matches[i]['mine']['outlier'] = outlier[i]
self.speak('points beyond {0} are outliers ({1})'.format(limit, i))
self.pixelstowavelengths = Legendre.fit(
x=self.pixel,
y=self.wavelength,
deg=self.polynomialdegree,
w=self.weights)
self.plotWavelengthFit()
#self.updatew2p()
@property
def weights(self):
return self.good+self.handpicked*10
@property
def residuals(self):
return self.wavelength - self.pixelstowavelengths(self.pixel)
@property
def good(self):
isntoutlier = np.array([m['mine']['outlier'] == False for m in self.matches])
ishandpicked = np.array([m['mine']['handpicked'] for m in self.matches])
return (isntoutlier + ishandpicked) > 0
@property
def pixel(self):
return np.array([m['mine']['w'] for m in self.matches])
@property
def pixelelement(self):
return np.array([m['theirs']['element'] for m in self.matches])
@property
def intensity(self):
return np.array([m['mine']['intensity'] for m in self.matches])
@property
def wavelength(self):
return np.array([m['theirs']['wavelength'] for m in self.matches])
@property
def emissioncolor(self):
return np.array([colors[m['theirs']['element']] for m in self.matches])
def plotWavelengthFit(self, interactive=True):
# plot to make sure the wavelength calibration makes sense
self.speak('{}, {}'.format(self.pixelstowavelengths, self.pixelstowavelengths.domain, self.pixelstowavelengths.window))
self.figcal = plt.figure('wavelength calibration',
figsize=(15,6), dpi=72)
self.interactivewave = zachopy.iplot.iplot(4,1,
height_ratios=[0.1, 0.4, 0.2, .2], hspace=0.1,
bottom=0.15)
self.ax_header = self.interactivewave.subplot(0)
# do the lamp spectra overlap?
self.ax_walign = self.interactivewave.subplot(1)
# what is the actual wavelength calibration
self.ax_wcal = self.interactivewave.subplot(2, sharex=self.ax_walign)
# what are the residuals from the fit
self.ax_wres = self.interactivewave.subplot(3, sharex=self.ax_walign)
for ax in [self.ax_header, self.ax_walign, self.ax_wcal]:
plt.setp(ax.get_xticklabels(), visible=False)
self.ax_header.set_title("Wavelength Calib. for Aperture"
" (%0.1f,%0.1f)" % (self.aperture.x, self.aperture.y))
# print information about the wavelength calibration
self.ax_header.set_xlim(0,1)
self.ax_header.set_ylim(0,1)
plt.setp(self.ax_header.get_yticklabels(), visible=False)
self.ax_header.patch.set_visible(False)
text = 'Wavelength-to-pixel guesses are {0}\n'.format(self.whichwaveid)
text += 'Hand-picked matches are from {0}.\n'.format(self.matchesfilename.split('/')[-1])
text += 'Pixel-to-wavelength calibration is '
if self.justloaded:
text += 'from ' + self.calibrationfilename.split('/')[-1]
else:
text += '[new!]'
self.ax_header.text(0.025, 0.5, text,
va='center', ha='left', fontsize=10)
for i, e in enumerate(self.elements):
self.ax_header.text(0.98,
1.0-(i+1.0)/(len(self.elements)+1),
e,
ha='right', va='center',
color=colors[e],
fontsize=6)
# plot the backwards calibration
for e in self.elements:
ok = np.nonzero([e in self.waveids['name'][i] for i in range(len(self.waveids))])[0]
xvals = self.waveids['pixel'][ok]
yvals = self.waveids['wavelength'][ok]
self.ax_header.scatter(xvals, yvals,
marker='o', color=colors[e], alpha=0.3)
xfine = np.linspace(min(self.waveids['wavelength']), max(self.waveids['wavelength']))
#self.ax_header.plot(self.waveids, xfine,
# alpha=0.3)
# plot the overlap of the lamp spectra
scatterkw = dict( marker='o', linewidth=0,
alpha=0.5, color=self.emissioncolor)
for element in self.elements:
self.ax_walign.plot(self.aperture.waxis,
self.aperture.arcs[element]['filtered'],
color=colors[element], alpha=0.5)
self.ax_walign.scatter(self.pixel, self.intensity, **scatterkw)
self.ax_walign.set_yscale('log')
self.ax_walign.set_ylim(1, None)
# plot tick marks for the known wavelengths
for e in self.elements:
for tw in self.knownwavelengths[e]:
pix = tw['pixelguess']
self.ax_walign.axvline(pix,
ymin=0.9,
color=colors[e],
alpha=0.5)
# plot the calibration
x = np.linspace(*zachopy.oned.minmax(self.aperture.waxis),num=200)
self.ax_wcal.plot(x, self.pixelstowavelengths(x), alpha=0.5, color='black')
self.ax_wcal.set_ylabel('Wavelength (angstroms)')
self.ax_wres.set_ylabel('Residuals')
scatterkw['color'] = self.emissioncolor[self.good]
self.ax_wcal.scatter( self.pixel[self.good],
self.wavelength[self.good],
**scatterkw)
self.ax_wres.scatter( self.pixel[self.good],
self.residuals[self.good],
**scatterkw)
scatterkw['marker'] = 'x'
bad = self.good == False
scatterkw['color'] = self.emissioncolor[bad]
self.ax_wcal.scatter(self.pixel[bad], self.wavelength[bad], **scatterkw)
self.ax_wres.scatter(self.pixel[bad], self.residuals[bad], **scatterkw)
self.ax_wres.set_xlabel('Pixel # (by python rules)')
self.ax_wres.set_xlim(*zachopy.oned.minmax(self.aperture.waxis))
self.ax_walign.scatter( self.pixel[self.handpicked],
self.intensity[self.handpicked],
marker='+', color='black')
self.ax_wcal.scatter( self.pixel[self.handpicked],
self.wavelength[self.handpicked],
marker='+', color='black')
self.ax_wres.scatter( self.pixel[self.handpicked],
self.residuals[self.handpicked],
marker='+', color='black')
self.ax_wres.axhline(0, linestyle='--', color='gray', zorder=-100)
rms = np.std(self.residuals[self.good])
performance = 'using a {}-degree {}'.format(self.polynomialdegree,
self.pixelstowavelengths.__class__.__name__)
performance += ' the calibration has an RMS of {0:.2f}A'.format(rms)
performance += ' with {0:.0f} good points'.format(np.sum(self.good))
performance += ' from {:.0f} to {:.0f}'.format(*zachopy.oned.minmax(self.pixel[self.good]))
self.ax_wres.text(0.98, 0.05, performance,
fontsize=10,
ha='right', va='bottom',
transform=self.ax_wres.transAxes)
self.ax_wres.set_ylim(*np.array([-1,1])*np.maximum(zachopy.oned.mad(self.residuals[self.good])*10, 1))
plt.draw()
self.speak('check out the wavelength calibration')
if interactive == False:
return
options = {}
options['q'] = dict(description='[q]uit without writing',
function=self.quit,
requiresposition=False)
options['w'] = dict(description='[w]rite the new calibration',
function=self.saveandquit,
requiresposition=False)
options['h'] = dict(description='match up a [h]elium line',
function=self.see,
requiresposition=True)
options['n'] = dict(description='match up a [n]eon line',
function=self.see,
requiresposition=True)
options['a'] = dict(description='match up a [a]rgon line',
function=self.see,
requiresposition=True)
options['d'] = dict(description='change the polynomial [d]egree',
function=self.changedegree,
requiresposition=True)
options['z'] = dict(description='[z]ap a previously matched line',
function=self.zap,
requiresposition=True)
options['u'] = dict(description='[u]ndo all changes made this session',
function=self.undo,
requiresposition=False)
options['r'] = dict(description='[r]estart from all defaults',
function=self.restart,
requiresposition=False)
options['!'] = dict(description='raise an error[!]',
function=self.freakout,
requiresposition=False)
self.speak('your options include:')
for v in options.values():
self.speak(' ' + v['description'])
pressed = self.interactivewave.getKeyboard()
try:
# figure out which option we're on
thing = options[pressed.key.lower()]
# check that it's a valid position, if need be
if thing['requiresposition']:
assert(pressed.inaxes is not None)
# execute the function associated with this option
thing['function'](pressed)
except KeyError:
self.speak("nothing yet defined for [{}]".format(pressed.key))
return
except AssertionError:
self.speak("that didn't seem to be at a valid position!")
return
def freakout(self, *args):
raise RuntimeError('Breaking here, for debugginng')
def undo(self, *args):
self.populate()
def restart(self, *args):
self.populate(restart=True)
def changedegree(self, pressed):
'''change the degree of the polynomial'''
self.speak('please enter a new polynomial degree (1-9)')
new = self.interactivewave.getKeyboard()
self.polynomialdegree = int(new.key)
def quit(self, *args):
'''quit without saving'''
self.speak('finished!')
self.notconverged = False
def saveandquit(self, *args):
'''save and quit'''
self.speak('saving the wavelength calibration')
self.save()
self.quit()
@property
def handpicked(self):
return np.array([m['mine']['handpicked'] for m in self.matches])
def selectpeak(self, pressed):
#print pressed.xdata, pressed.ydata
element = shortcuts[pressed.key]
self.speak('matching for the closest {0} line'.format(element))
valid = np.where([e == element for e in self.pixelelement])[0]
closest = valid[np.argmin(np.abs(self.pixel[valid] - pressed.xdata))]
return closest
def selectwavelength(self, pressed):
#print pressed.xdata, pressed.ydata
element = shortcuts[pressed.key.lower()]
pixguess = [w['pixelguess'] for w in self.knownwavelengths[element]]
closest = ((pixguess - pressed.xdata)**2).argmin()
return self.knownwavelengths[element][closest]
def checkvalid(self, pressed):
if pressed.xdata is None:
self.speak('''you typed "{0}", but the window didn't return a coordinate, please try again!'''.format(pressed.key))
return False
else:
return True
def zap(self, pressed):
if self.checkvalid(pressed) == False:
return
closest = ((self.pixel - pressed.xdata)**2).argmin()
self.matches[closest]['mine']['handpicked'] = False
def see(self, pressed):
if self.checkvalid(pressed) == False:
return
closest = self.selectpeak(pressed)
if closest is None:
return
match = self.matches[closest]
pixel = match['mine']['w']
self.speak('you are e[X]cited about an indentifiable emission line at {0:.1}'.format(pixel))
self.ax_walign.axvline(pixel, alpha=0.5, color=colors[match['theirs']['element']])
self.speak(' now, please click a [H]e, [N]e, or [A]r line')
# get a keyboard input at mouse position
secondpressed = self.interactivewave.getKeyboard()
if secondpressed.key.lower() in ['h', 'n', 'a']:
# pull out the closest wavelength match
wavematch = self.selectwavelength(secondpressed)
else:
self.speak("hmmm...ignoring")
return
self.matches[closest]['mine']['handpicked'] = True
self.speak('updating the guess for {0}A from {1} to {2}'.format(
wavematch['wavelength'],
wavematch['pixelguess'],
pixel
))
self.matches[closest]['theirs']['pixelguess'] = pixel
self.matches[closest]['mine']['wavelength'] = wavematch['wavelength']
def exclude(self, pressed):
self.speak('excluding a point')
if pressed.inaxes == self.ax_wres:
closest = self.select(pressed)
self.matches[closest]['mine']['handpicked'] = False
def fit_legendres_images(images,
centers,
lg_inds,
rad_inds,
maxPixel,
rotate=0,
image_stds=None,
image_counts=None,
image_nanMaps=None,
image_weights=None,
chiSq_fit=False,
rad_range=None):
"""
Fits legendre polynomials to an array of single images (3d) or a list/array of
an array of scan images, possible dimensionality:
1) [NtimeSteps, image_rows, image_cols]
2) [NtimeSteps (list), Nscans, image_rows, image_cols]
"""
if image_counts is None:
image_counts = []
for im in range(len(images)):
image_counts.append(np.ones_like(images[im]))
image_counts[im][np.isnan(images[im])] = 0
if chiSq_fit and (image_stds is None):
print("If using the chiSq fit you must supply image_stds")
return None
if image_stds is None:
image_stds = []
for im in range(len(images)):
image_stds.append(np.ones_like(images[im]))
image_stds[im][np.isnan(images[im])] = 0
with_scans = len(images[0].shape) + 1 >= 4
img_fits = [[] for x in range(len(images))]
img_covs = [[] for x in range(len(images))]
for rad in range(maxPixel):
if rad_range is not None:
if rad < rad_range[0] or rad >= rad_range[1]:
continue
if rad % 25 == 0:
print("Fitting radius {}".format(rad))
pixels, nans, angles = [], [], []
all_angles = np.arctan2(rad_inds[rad][1].astype(float),
rad_inds[rad][0].astype(float))
all_angles[all_angles < 0] += 2 * np.pi
all_angles = np.mod(all_angles + rotate, 2 * np.pi)
all_angles[all_angles > np.pi] -= 2 * np.pi
if np.sum(np.mod(lg_inds, 2)) == 0:
all_angles[np.abs(all_angles) > np.pi / 2.] -= np.pi * np.sign(
all_angles[np.abs(all_angles) > np.pi / 2.])
angles = np.unique(np.abs(all_angles))
ang_sort_inds = np.argsort(angles)
angles = angles[ang_sort_inds]
Nangles = angles.shape[0]
if len(angles) == len(all_angles):
do_merge = False
else:
do_merge = True
mi_rows, mi_cols, mi_data = [], [], []
pr, pc, pv = [], [], []
for ia, ang in enumerate(angles):
inds = np.where(np.abs(all_angles) == ang)[0]
mi_rows.append(np.ones_like(inds) * ia)
mi_cols.append(inds)
mi_rows, mi_cols = np.concatenate(mi_rows), np.concatenate(mi_cols)
merge_indices = csr_matrix(
(np.ones_like(mi_rows), (mi_rows, mi_cols)),
shape=(len(angles), len(all_angles)))
for im in range(len(images)):
if with_scans:
angs_tile = np.tile(angles, (images[im].shape[0], 1))
scn_inds, row_inds, col_inds = [], [], []
for isc in range(images[im].shape[0]):
scn_inds.append(
np.ones(rad_inds[rad][0].shape[0], dtype=int) * isc)
row_inds.append(rad_inds[rad][0] + centers[im][isc, 0])
col_inds.append(rad_inds[rad][1] + centers[im][isc, 1])
scn_inds = np.concatenate(scn_inds)
row_inds = np.concatenate(row_inds)
col_inds = np.concatenate(col_inds)
img_pixels = np.reshape(
copy(images[im][scn_inds, row_inds, col_inds]),
(images[im].shape[0], -1))
img_counts = np.reshape(
copy(image_counts[im][scn_inds, row_inds, col_inds]),
(images[im].shape[0], -1))
img_stds = np.reshape(
copy(image_stds[im][scn_inds, row_inds, col_inds]),
(images[im].shape[0], -1))
if image_nanMaps is not None:
img_pixels[np.reshape(
image_nanMaps[im][scn_inds, row_inds, col_inds],
(images[im].shape[0], -1)).astype(bool)] = np.nan
img_counts[np.reshape(
image_nanMaps[im][scn_inds, row_inds, col_inds],
(images[im].shape[0], -1)).astype(bool)] = 0
if image_weights is not None:
img_weights = np.reshape(
copy(image_weights[im][scn_inds, row_inds, col_inds]),
(images[im].shape[0], -1))
else:
angs_tile = np.expand_dims(angles, 0)
row_inds = rad_inds[rad][0] + centers[im, 0]
col_inds = rad_inds[rad][1] + centers[im, 1]
img_pixels = np.reshape(copy(images[im][row_inds, col_inds]),
(1, -1))
img_counts = np.reshape(
copy(image_counts[im][row_inds, col_inds]), (1, -1))
img_stds = np.reshape(copy(image_stds[im][row_inds, col_inds]),
(1, -1))
if image_nanMaps is not None:
img_pixels[np.reshape(
image_nanMaps[im][row_inds, col_inds],
(1, -1)).astype(bool)] = np.nan
img_counts[np.reshape(
image_nanMaps[im][row_inds, col_inds],
(1, -1)).astype(bool)] = 0
if image_weights is not None:
img_weights = np.reshape(
copy(image_weights[im][row_inds, col_inds]), (1, -1))
img_pix = img_pixels * img_counts
img_var = img_counts * (img_stds**2)
img_pix[np.isnan(img_pixels)] = 0
img_var[np.isnan(img_pixels)] = 0
if do_merge:
img_pixels[np.isnan(img_pixels)] = 0
img_pix = np.transpose(merge_indices.dot(
np.transpose(img_pix)))
img_var = np.transpose(merge_indices.dot(
np.transpose(img_var)))
img_counts = np.transpose(
merge_indices.dot(np.transpose(img_counts)))
if image_weights is not None:
print("Must fill this in, don't forget std option")
sys.exit(0)
else:
img_pix = img_pix[:, ang_sort_inds]
img_var = img_var[:, ang_sort_inds]
img_counts = img_counts[:, ang_sort_inds]
img_pix /= img_counts
img_var /= img_counts
Nnans = np.sum(np.isnan(img_pix), axis=-1)
ang_inds = np.where(img_counts > 0)
arr_inds = np.concatenate(
[np.arange(Nangles - Nn) for Nn in Nnans])
img_pixels = np.zeros_like(img_pix)
img_vars = np.zeros_like(img_var)
img_angs = np.zeros_like(img_pix)
img_dang = np.zeros_like(img_pix)
img_pixels[ang_inds[0][:-1], arr_inds[:-1]] =\
(img_pix[ang_inds[0][:-1], ang_inds[1][:-1]] + img_pix[ang_inds[0][1:], ang_inds[1][1:]])/2.
img_vars[ang_inds[0][:-1], arr_inds[:-1]] =\
(img_var[ang_inds[0][:-1], ang_inds[1][:-1]] + img_var[ang_inds[0][1:], ang_inds[1][1:]])/2.
img_angs[ang_inds[0][:-1], arr_inds[:-1]] =\
(angs_tile[ang_inds[0][:-1], ang_inds[1][:-1]] + angs_tile[ang_inds[0][1:], ang_inds[1][1:]])/2.
img_dang[ang_inds[0][:-1], arr_inds[:-1]] =\
(angs_tile[ang_inds[0][1:], ang_inds[1][1:]] - angs_tile[ang_inds[0][:-1], ang_inds[1][:-1]])
for isc in range(Nnans.shape[0]):
# Using angle midpoint => one less angle => Nnans[isc]+1
img_pixels[isc, -1 * (Nnans[isc] + 1):] = 0
img_vars[isc, -1 * (Nnans[isc] + 1):] = 0
img_angs[isc, -1 * (Nnans[isc] + 1):] = 0
img_dang[isc, -1 * (Nnans[isc] + 1):] = 0
if image_weights is not None:
print("Must fill this in and check below")
sys.exit(0)
elif chiSq_fit:
img_weights = 1. / img_vars
img_weights[img_vars == 0] = 0
else:
img_weights = np.ones_like(img_pixels)
img_weights *= np.sin(img_angs) * img_dang
lgndrs = []
for lg in lg_inds:
lgndrs.append(Legendre.basis(lg)(np.cos(img_angs)))
lgndrs = np.transpose(np.array(lgndrs), (1, 0, 2))
empty_scan = np.sum(img_weights.astype(bool), -1) < 2
overlap = np.einsum('bai,bi,bci->bac',
lgndrs[np.invert(empty_scan)],
img_weights[np.invert(empty_scan)],
lgndrs[np.invert(empty_scan)],
optimize='greedy')
empty_scan[np.invert(empty_scan)] = (np.linalg.det(overlap) == 0.0)
if np.any(empty_scan):
fit = np.ones((img_pixels.shape[0], len(lg_inds))) * np.nan
cov = np.ones(
(img_pixels.shape[0], len(lg_inds), len(lg_inds))) * np.nan
if np.any(np.invert(empty_scan)):
img_pixels = img_pixels[np.invert(empty_scan)]
img_weights = img_weights[np.invert(empty_scan)]
img_vars = img_vars[np.invert(empty_scan)]
lgndrs = lgndrs[np.invert(empty_scan)]
fit[np.invert(empty_scan)], cov[np.invert(empty_scan)] =\
normal_eqn_vects(lgndrs, img_pixels, img_weights, img_vars)
else:
fit, cov = normal_eqn_vects(lgndrs, img_pixels, img_weights,
img_vars)
img_fits[im].append(np.expand_dims(fit, 1))
img_covs[im].append(np.expand_dims(cov, 1))
Nscans = None
for im in range(len(img_fits)):
img_fits[im] = np.concatenate(img_fits[im], 1)
img_covs[im] = np.concatenate(img_covs[im], 1)
if Nscans is None:
Nscans = img_fits[im].shape[0]
elif Nscans != img_fits[im].shape[0]:
Nscans = -1
if Nscans > 0:
img_fits = np.array(img_fits)
img_covs = np.array(img_covs)
if with_scans:
return img_fits, img_covs
else:
return img_fits[:, 0, :, :], img_covs[:, 0, :, :]
