def measure_flexure_y(fine, HDUlist, profwidth=5, plot=False, outname=None):
dat = HDUlist[0].data
exptime = HDUlist[0].header["EXPTIME"]
profs = []
xx = np.arange(profwidth * 2)
for ix in np.arange(500, 1200, 10):
f = fine[ix]
profile = np.zeros(profwidth * 2)
# bad fit
if not f.ok:
continue
# noisy fit
if f.lamnrms > 1:
continue
# short spectrum
if f.xrange[1] - f.xrange[0] < 200:
continue
yfun = np.poly1d(f.poly)
for xpos in np.arange(f.xrange[0], f.xrange[1]):
try:
ypos = int(np.round(yfun(xpos)))
except:
continue
try:
profile += dat[ypos - profwidth : ypos + profwidth, xpos]
except:
continue
profs.append(profile - np.min(profile))
if plot:
pl.figure(1)
ffun = FF.mpfit_residuals(FF.gaussian4)
parinfo = [{"value": 1}, {"value": profwidth}, {"value": 2}, {"value": 0}, {"value": 0}]
profposys = []
for prof in profs:
if plot:
pl.step(xx, prof)
parinfo[0]["value"] = np.max(prof)
fit = FF.mpfit_do(ffun, xx, prof, parinfo)
if plot:
pl.plot(xx, FF.gaussian4(fit.params, xx))
profposys.append(fit.params[1] - profwidth - 1)
if plot:
pl.show()
profposys = np.array(profposys)
mn = np.mean(profposys)
sd = np.std(profposys)
ok = np.abs(profposys - mn) / sd < 3
required_shift = np.mean(profposys[ok])
print "dY = %3.2f pixel shift" % required_shift
return required_shift
def measure_flexure_y(fine, HDUlist, profwidth=5, plot=False, outname=None):
dat = HDUlist[0].data
exptime = HDUlist[0].header['EXPTIME']
profs = []
xx = np.arange(profwidth*2)
for ix in np.arange(500, 1200, 10):
f = fine[ix]
profile = np.zeros(profwidth*2)
if not f.ok: continue
if f.lamrms > 1: continue
if f.xrange[1] - f.xrange[0] < 200: continue
yfun = np.poly1d(f.poly)
for xpos in np.arange(f.xrange[0], f.xrange[1]):
try: ypos = int(np.round(yfun(xpos)))
except: continue
try: profile += dat[ypos-profwidth:ypos+profwidth, xpos]
except: continue
profs.append(profile - np.min(profile))
if plot: pl.figure(1)
ffun = FF.mpfit_residuals(FF.gaussian4)
parinfo= [{'value': 1}, {'value': profwidth}, {'value': 2},
{'value': 0}, {'value': 0}]
profposys = []
for prof in profs:
if plot: pl.step(xx, prof)
parinfo[0]['value'] = np.max(prof)
fit = FF.mpfit_do(ffun, xx, prof, parinfo)
if plot: pl.plot(xx, FF.gaussian4(fit.params, xx))
profposys.append(fit.params[1] - profwidth-1)
if plot: pl.show()
profposys = np.array(profposys)
mn = np.mean(profposys)
sd = np.std(profposys)
ok = np.abs(profposys - mn)/sd < 3
required_shift = np.mean(profposys[ok])
print "dY = %3.2f pixel shift" % required_shift
return required_shift
def measure_flexure_y(cube, hdulist, profwidth=5, plot=False):
# get image to measure
dat = hdulist[0].data
# store profiles
profs = []
# x values of profiles
xx = np.arange(profwidth*2)
# get a sample (70) of spaxels
for ix in np.arange(500, 1200, 10):
# get specific spaxel
f = cube[ix]
# set up profile vector for this spaxel
profile = np.zeros(profwidth*2)
# weed out baddies
# bad fit
if not f.ok:
continue
# noisy fit
if f.lamnrms > 1:
continue
# short spectrum
if f.xrange[1] - f.xrange[0] < 200:
continue
# get trace of spaxel on image in y
yfun = np.poly1d(f.poly)
# loop over x positions of this spaxel
for xpos in np.arange(f.xrange[0], f.xrange[1]):
# get the y position at the given xpos
try:
ypos = int(np.round(yfun(xpos)))
except:
continue
# sum profile over all xpos
try:
profile += dat[ypos-profwidth:ypos+profwidth, xpos]
except:
continue
# append profile after subtracting minimum
profs.append(profile - np.min(profile))
if plot:
pl.figure(1)
# set up Gaussian fit to profiles
ffun = NFit.mpfit_residuals(NFit.gaussian4)
# initial guess
parinfo = [{'value': 1}, {'value': profwidth}, {'value': 2},
{'value': 0}, {'value': 0}]
# y positions of profiles
profposys = []
profwidys = []
for prof in profs:
# plot input profile
if plot:
pl.plot(xx, prof, 'ro')
# update initial guess
parinfo[0]['value'] = np.max(prof)
# fit Gaussian
fit = NFit.mpfit_do(ffun, xx, prof, parinfo)
# overplot fit
if plot:
pl.plot(xx, NFit.gaussian4(fit.params, xx))
# x offset between nominal trace position (profwidth-1) and
# Gaussian fit position (fit.params[1])
profposys.append(fit.params[1] - profwidth-1)
profwidys.append(fit.params[2])
profposys = np.array(profposys)
profwidys = np.array(profwidys)
# get statistics
mn = np.mean(profposys)
sd = np.std(profposys)
# clean 3 sigma outliers
ok = np.abs(profposys - mn)/sd < 3
# final shift
required_shift = np.mean(profposys[ok])
# now do width
mn = np.mean(profwidys)
sd = np.std(profwidys)
# clean 3 sigma outliers
ok = np.abs(profwidys - mn)/sd < 3
average_width = np.mean(profwidys[ok]) * 2.354
print "yFWHM = %5.2f pixels" % average_width
print "dY = %3.2f pixel shift" % required_shift
if plot:
px = (profwidth+1) + required_shift
pl.plot([px, px], [0, np.max(profs)], '--')
pl.show()
return required_shift, average_width
def measure_flexure_x(cube, hdulist, drow=0., skylines=(557.0, 589.0),
lamstart=1000.0, lamratio=239./240., lamlen=250,
extract_width=3, skywidth=9, outfile='dX', plot=False):
"""Measures flexure in X direction, returns pixel offset
Args:
cube (extraction array): List of Extraction object, the fine loc +
wave solution for each spectrum
hdulist (pyfits obj): Pyfits object for the spectrum to measure
drow (float): offset in rows for flexure (y-axis)
skylines (float, float): The night skylines to centroid on in nm
- See Wavelength.fiducial spectrum for following:
lamstart (float): Wavelength to start the grid on, default 1000 nm
lamratio (float): Resolution of sed machine
lamlen (int): Length of spectrum
extract_width(int): Number of pixels to extract spectrum around
skywidth(float): Fit gaussian to the ROI of skyline+-skywidth in nm.
outfile (string): output pdf plot showing fits to skyline(s)
plot (bool): set to True to show plot, else plot to pdf file
Returns:
Offset number of pixels in X direction.
"""
# read in image
dat = hdulist[0].data
# select 70 representative spaxels
spec_ixs = np.arange(500, 1200, 10)
# fiducial wavelength grid
lamgrid = Wavelength.fiducial_spectrum(lamstart=lamstart,
lamratio=lamratio, len=lamlen)
# initialize grid of spaxel spectra
specgrid = np.zeros((len(lamgrid), len(spec_ixs)))
# loop over selected spaxels
for i, ix in enumerate(spec_ixs):
# get spaxel
f = cube[ix]
# skip baddies
# bad fit
if not f.ok:
continue
# noisy fit
if f.lamnrms > 1:
continue
# short spectrum
if f.xrange[1] - f.xrange[0] < 200:
continue
# set up spectral vector for this spaxel
spec = np.zeros(f.xrange[1] - f.xrange[0])
yfun = np.poly1d(f.poly)
# loop over x positions in spaxel
for jx, xpos in enumerate(np.arange(f.xrange[0], f.xrange[1])):
# get y position on image
ypos = yfun(xpos)
# extract spectrum from image
try:
spec[jx] = np.sum(dat[ypos-extract_width:ypos+extract_width,
xpos])
except:
continue
# get wavelengths of spaxel
try:
ll = f.get_lambda_nm()
except:
continue
# resample spectrum on fiducial wavelength grid
specfun = interp1d(ll, spec, bounds_error=False)
# insert into grid of spaxel spectra
specgrid[:, i] = specfun(lamgrid)
# create a median spectrum from spaxel grid
# taking a median minimizes impact of objects in sample
skyspec = np.median(specgrid, axis=1)
# plot resulting sky spectrum
pl.step(lamgrid, skyspec, where='mid')
pl.xlabel("Wavelength [nm]")
pl.ylabel("Spec Irr [ph/10 m/nm]")
legend = ["Sky"]
# accumulate average offsets from known sky lines
sumoff = 0.
sumscale = 0.
minsig = 10000.
# loop over input sky lines
for skyline in skylines:
# extract a wavelength window around sky line
roi = (lamgrid > skyline-skywidth) & (lamgrid < skyline+skywidth)
# prepare for Gaussian fit
ffun = NFit.mpfit_residuals(NFit.gaussian4)
# initial setup of fit
parinfo = [
{'value': np.max(skyspec[roi]), 'limited': [1, 0],
'limits': [0, 0]},
{'value': skyline},
{'value': 3},
{'value': np.min(skyspec[roi]), 'limited': [1, 0],
'limits': [0, 0]}]
# do the fit
fit = NFit.mpfit_do(ffun, lamgrid[roi], skyspec[roi], parinfo)
# did the fit succeed?
if fit.status == 1 and fit.params[2] > 0.:
off = fit.params[1] - skyline
sumoff += off * fit.params[0]
sumscale += fit.params[0]
if fit.params[2] < minsig:
minsig = fit.params[2]
pl.plot(lamgrid[roi], NFit.gaussian4(fit.params, lamgrid[roi]))
dxnm = fit.params[1] - skyline
legend.append("%.1f, %.2f" % (skyline, off))
else:
dxnm = 0.
print("line = %6.1f (%6.1f), FWHM = %.2f nm, status = %d, dX = %3.2f nm shift" %
(skyline, fit.params[0], fit.params[2]*2.354, fit.status, dxnm))
if sumscale > 0.:
dxnm = sumoff / sumscale
else:
print "Warning: no good skylines to fit! Setting X offset to 0.0 nm"
dxnm = 0.
minsig = 0.
print "dX = %3.2f nm shift" % dxnm
pl.title("dX = %3.2f nm shift, dY = %3.2f px shift" % (dxnm, drow))
pl.legend(legend)
if plot:
pl.show()
else:
pl.savefig(outfile + ".pdf")
# return offset and best FWHM
return dxnm, minsig*2.354
def measure_flexure_y(cube, hdulist, profwidth=5, plot=False):
# get image to measure
dat = hdulist[0].data
# store profiles
profs = []
# x values of profiles
xx = np.arange(profwidth*2)
# get a sample (70) of spaxels
for ix in np.arange(500, 1200, 10):
# get specific spaxel
f = cube[ix]
# set up profile vector for this spaxel
profile = np.zeros(profwidth*2)
# weed out baddies
# bad fit
if not f.ok:
continue
# noisy fit
if f.lamnrms > 1:
continue
# short spectrum
if f.xrange[1] - f.xrange[0] < 200:
continue
# get trace of spaxel on image in y
yfun = np.poly1d(f.poly)
# loop over x positions of this spaxel
for xpos in np.arange(f.xrange[0], f.xrange[1]):
# get the y position at the given xpos
try:
ypos = int(np.round(yfun(xpos)))
except:
continue
# sum profile over all xpos
try:
profile += dat[ypos-profwidth:ypos+profwidth, xpos]
except:
continue
# append profile after subtracting minimum
profs.append(profile - np.min(profile))
if plot:
pl.figure(1)
# set up Gaussian fit to profiles
ffun = NFit.mpfit_residuals(NFit.gaussian4)
# initial guess
parinfo = [{'value': 1}, {'value': profwidth}, {'value': 2},
{'value': 0}, {'value': 0}]
# y positions of profiles
profposys = []
profwidys = []
for prof in profs:
# plot input profile
if plot:
pl.plot(xx, prof, 'ro')
# update initial guess
parinfo[0]['value'] = np.max(prof)
# fit Gaussian
fit = NFit.mpfit_do(ffun, xx, prof, parinfo)
# overplot fit
if plot:
pl.plot(xx, NFit.gaussian4(fit.params, xx))
# x offset between nominal trace position (profwidth-1) and
# Gaussian fit position (fit.params[1])
profposys.append(fit.params[1] - profwidth-1)
profwidys.append(fit.params[2])
profposys = np.array(profposys)
profwidys = np.array(profwidys)
# get statistics
mn = np.mean(profposys)
sd = np.std(profposys)
# clean 3 sigma outliers
ok = np.abs(profposys - mn)/sd < 3
# final shift
required_shift = np.mean(profposys[ok])
# now do width
mn = np.mean(profwidys)
sd = np.std(profwidys)
# clean 3 sigma outliers
ok = np.abs(profwidys - mn)/sd < 3
average_width = np.mean(profwidys[ok]) * 2.354
print("yFWHM = %5.2f pixels" % float(average_width))
print("dY = %3.2f pixel shift" % required_shift)
if plot:
px = (profwidth+1) + required_shift
pl.plot([px, px], [0, np.max(profs)], '--')
pl.show()
return required_shift, average_width
def measure_flexure_x(cube, hdulist, drow=0., skylines=(557.0, 589.0),
extract_width=3, skywidth=9, outfile='dX', plot=False):
"""Measures flexure in X direction, returns pixel offset
Args:
cube (extraction array): List of Extraction object, the fine loc +
wave solution for each spectrum
hdulist (astropy.io.fits obj): Pyfits object for the spectrum to measure
drow (float): offset in rows for flexure (y-axis)
skylines (float, float): The night skylines to centroid on in nm
extract_width(int): Number of pixels to extract spectrum around
skywidth(float): Fit gaussian to the ROI of skyline+-skywidth in nm.
outfile (string): output pdf plot showing fits to skyline(s)
plot (bool): set to True to show plot, else plot to pdf file
Returns:
Offset number of pixels in X direction.
"""
# read in image
dat = hdulist[0].data
# select 70 representative spaxels
spec_ixs = np.arange(500, 1200, 10)
# fiducial wavelength grid
lamgrid = None
# loop over selected spaxels
for i, ix in enumerate(spec_ixs):
# get spaxel
f = cube[ix]
# skip baddies
# bad fit
if not f.ok:
continue
# noisy fit
if f.lamnrms > 1:
continue
# short spectrum
if f.xrange[1] - f.xrange[0] < 200:
continue
# set up spectral vector for this spaxel
spec = np.zeros(f.xrange[1] - f.xrange[0])
yfun = np.poly1d(f.poly)
# loop over x positions in spaxel
for jx, xpos in enumerate(np.arange(f.xrange[0], f.xrange[1])):
# get y position on image
ypos = int(yfun(xpos))
# extract spectrum from image
try:
spec[jx] = np.sum(dat[ypos-extract_width:ypos+extract_width,
xpos])
except:
print("Warning: no sum for sky spectrum %d at %d" % (jx, xpos))
continue
# get wavelengths of spaxel
try:
ll = f.get_lambda_nm()
except:
continue
if lamgrid is None:
lamgrid = ll
# initialize grid of spaxel spectra
specgrid = np.zeros((len(lamgrid), len(spec_ixs)))
# resample spectrum on fiducial wavelength grid
specfun = interp1d(ll, spec, bounds_error=False)
# insert into grid of spaxel spectra
specgrid[:, i] = specfun(lamgrid)
# create a median spectrum from spaxel grid
# taking a median minimizes impact of objects in sample
skyspec = np.nanmedian(specgrid, axis=1)
# plot resulting sky spectrum
pl.step(lamgrid, skyspec, where='mid')
pl.xlabel("Wavelength [nm]")
pl.ylabel("Spec Irr [ph/10 m/nm]")
legend = ["Sky"]
# accumulate average offsets from known sky lines
sumoff = 0.
sumscale = 0.
minsig = 10000.
# loop over input sky lines
for skyline in skylines:
# extract a wavelength window around sky line
roi = (lamgrid > skyline-skywidth) & (lamgrid < skyline+skywidth)
# prepare for Gaussian fit
ffun = NFit.mpfit_residuals(NFit.gaussian4)
# initial setup of fit
parinfo = [
{'value': np.max(skyspec[roi]), 'limited': [1, 0],
'limits': [0, 0]},
{'value': skyline},
{'value': 3},
{'value': np.min(skyspec[roi]), 'limited': [1, 0],
'limits': [0, 0]}]
# do the fit
fit = NFit.mpfit_do(ffun, lamgrid[roi], skyspec[roi], parinfo)
# did the fit succeed?
if fit.status == 1 and 0. < fit.params[2] < 12.:
off = fit.params[1] - skyline
sumoff += off * fit.params[0]
sumscale += fit.params[0]
if fit.params[2] < minsig:
minsig = fit.params[2]
pl.plot(lamgrid[roi], NFit.gaussian4(fit.params, lamgrid[roi]))
dxnm = fit.params[1] - skyline
legend.append("%.1f, %.2f" % (skyline, off))
print("line = %6.1f (%6.1f), FWHM = %.2f nm, status = %d,"
" dX = %3.2f nm shift" %
(skyline, fit.params[0], fit.params[2]*2.354, fit.status,
dxnm))
if sumscale > 0.:
dxnm = sumoff / sumscale
else:
print("Warning: no good skylines to fit! Setting X offset to 0.0 nm")
dxnm = 0.
minsig = 0.
print("dX = %3.2f nm shift" % dxnm)
pl.title("dX = %3.2f nm shift, dY = %3.2f px shift" % (dxnm, drow))
pl.legend(legend)
ax = pl.gca()
ax.annotate('DRP: ' + drp_ver, xy=(0.0, 0.01), xytext=(0, 0),
xycoords=('axes fraction', 'figure fraction'),
textcoords='offset points', size=6,
ha='center', va='bottom')
if plot:
pl.show()
else:
pl.savefig(outfile + ".pdf")
# return offset and best FWHM
return dxnm, minsig*2.354
def measure_flexure_x(
fine,
HDUlist,
plot=True,
dY=0,
skyline=589.0,
lamstart=1000.0,
lamratio=239.0 / 240.0,
lamlen=250,
extract_width=3,
skywidth=9,
outfile="dX",
):
"""Measures flexure in X direction, returns pixel offset
Args:
fine: List of Extraction object, the fine loc + wave solution for
each spectrum
HDUlist: Pyfits object for the spectrum to measure
plot: Plot + save results to a file
dY: the measured pixel flexure in Y direction to account for
skyline(float): The night skyline to centroid on in nm
skywidth(float): Fit gaussian to the ROI of (skyline-skywidth to
skyline+skywidth) in nm.
extract_width(int): Number of pixels to extract spectrum around
- See Wavelength.fiducial spectrum for following:
lamstart: Wavelength to start the grid on, default 1000 nm
lamratio: Resolution of sed machine
lamlen: Length of spectrum
Returns:
Offset number of pixels in X direction.
"""
dat = HDUlist[0].data
exptime = HDUlist[0].header["EXPTIME"]
spec_ixs = np.arange(500, 1200, 10)
lamgrid = Wavelength.fiducial_spectrum(lamstart=lamstart, lamratio=lamratio, len=lamlen)
specgrid = np.zeros((len(lamgrid), len(spec_ixs)))
for i, ix in enumerate(spec_ixs):
f = fine[ix]
# bad fit
if not f.ok:
continue
# noisy fit
if f.lamnrms > 1:
continue
# short spectrum
if f.xrange[1] - f.xrange[0] < 200:
continue
spec = np.zeros(f.xrange[1] - f.xrange[0])
yfun = np.poly1d(f.poly)
for jx, xpos in enumerate(np.arange(f.xrange[0], f.xrange[1])):
ypos = yfun(xpos)
try:
spec[jx] = np.sum(dat[ypos - extract_width : ypos + extract_width, xpos])
except:
continue
try:
ll = f.get_lambda_nm()
except:
continue
specfun = interp1d(ll, spec, bounds_error=False)
specgrid[:, i] = specfun(lamgrid)
skyspec = np.median(specgrid, axis=1)
pl.step(lamgrid, skyspec, where="mid")
roi = (lamgrid > skyline - skywidth) & (lamgrid < skyline + skywidth)
ffun = FF.mpfit_residuals(FF.gaussian4)
parinfo = [
{"value": np.max(skyspec[roi]), "limited": [1, 0], "limits": [0, 0]},
{"value": skyline},
{"value": 3},
{"value": np.min(skyspec[roi]), "limited": [1, 0], "limits": [0, 0]},
]
fit = FF.mpfit_do(ffun, lamgrid[roi], skyspec[roi], parinfo)
pl.plot(lamgrid, FF.gaussian4(fit.params, lamgrid))
pl.savefig(outfile + ".pdf")
dXnm = fit.params[1] - skyline
print "dX = %3.2f nm shift" % dXnm
return dXnm
def measure_flexure_x(fine, HDUlist, plot=True, dY=0,
skyline=589.0, lamstart=1000.0, lamratio=239./240., lamlen=250,
extract_width=3, skywidth=9, outfile='dX'):
'''Measures flexure in X direction, returns pixel offset
Args:
fine: List of Extraction object, the fine loc + wave solution for
each spectrum
HDUlist: Pyfits object for the spectrum to measure
plot: Plot + save results to a file
dY: the measured pixel flexure in Y direction to account for
skyline(float): The night skyline to centroid on in nm
skywidth(float): Fit gaussian to the ROI of (skyline-skywidth to
skyline+skywidth) in nm.
extract_width(int): Number of pixels to extract spectrum around
- See Wavelength.fiducial spectrum for following:
lamstart: Wavelength to start the grid on, default 1000 nm
lamratio: Resolution of sed machine
lamlen: Length of spectrum
Returns:
Offset number of pixels in X direction.
'''
dat = HDUlist[0].data
exptime = HDUlist[0].header['EXPTIME']
spec_ixs = np.arange(500, 1200, 10)
lamgrid = Wavelength.fiducial_spectrum(lamstart=lamstart,
lamratio=lamratio, len=lamlen)
specgrid = np.zeros((len(lamgrid), len(spec_ixs)))
for i,ix in enumerate(spec_ixs):
f = fine[ix]
if not f.ok: continue
if f.lamrms > 1: continue
if f.xrange[1] - f.xrange[0] < 200: continue
spec = np.zeros(f.xrange[1] - f.xrange[0])
yfun = np.poly1d(f.poly)
for jx,xpos in enumerate(np.arange(f.xrange[0], f.xrange[1])):
ypos = yfun(xpos)
try:spec[jx] = np.sum(dat[ypos-extract_width:ypos+extract_width,
xpos])
except: continue
try:ll = f.get_lambda_nm()
except: continue
specfun = interp1d(ll, spec, bounds_error=False)
specgrid[:,i] = specfun(lamgrid)
skyspec = np.median(specgrid, axis=1)
pl.step(lamgrid, skyspec, where='mid')
roi = (lamgrid>skyline-skywidth) & (lamgrid<skyline+skywidth)
ffun = FF.mpfit_residuals(FF.gaussian4)
parinfo= [
{'value': np.max(skyspec[roi]), 'limited': [1,0],
'limits': [0, 0]},
{'value': skyline},
{'value': 3},
{'value': np.min(skyspec[roi]), 'limited': [1,0],
'limits': [0,0]}]
fit = FF.mpfit_do(ffun, lamgrid[roi], skyspec[roi], parinfo)
pl.plot(lamgrid, FF.gaussian4(fit.params, lamgrid))
pl.savefig(outfile + ".pdf")
dXnm = fit.params[1] - skyline
print "dX = %3.2f nm shift" % dXnm
return dXnm
def measure_flexure_x(cube,
hdulist,
drow=0.,
skylines=(557.0, 589.0),
lamstart=1000.0,
lamratio=239. / 240.,
lamlen=250,
extract_width=3,
skywidth=9,
outfile='dX'):
"""Measures flexure in X direction, returns pixel offset
Args:
cube (extraction array): List of Extraction object, the fine loc +
wave solution for each spectrum
hdulist (pyfits obj): Pyfits object for the spectrum to measure
drow (float): offset in rows for flexure (y-axis)
skylines (float, float): The night skylines to centroid on in nm
skywidth(float): Fit gaussian to the ROI of (skyline-skywidth to
skyline+skywidth) in nm.
extract_width(int): Number of pixels to extract spectrum around
- See Wavelength.fiducial spectrum for following:
lamstart (float): Wavelength to start the grid on, default 1000 nm
lamratio (float): Resolution of sed machine
lamlen (int): Length of spectrum
outfile (string): output pdf plot showing fits to skyline(s)
Returns:
Offset number of pixels in X direction.
"""
dat = hdulist[0].data
spec_ixs = np.arange(500, 1200, 10)
lamgrid = Wavelength.fiducial_spectrum(lamstart=lamstart,
lamratio=lamratio,
len=lamlen)
specgrid = np.zeros((len(lamgrid), len(spec_ixs)))
for i, ix in enumerate(spec_ixs):
f = cube[ix]
# bad fit
if not f.ok:
continue
# noisy fit
if f.lamnrms > 1:
continue
# short spectrum
if f.xrange[1] - f.xrange[0] < 200:
continue
spec = np.zeros(f.xrange[1] - f.xrange[0])
yfun = np.poly1d(f.poly)
for jx, xpos in enumerate(np.arange(f.xrange[0], f.xrange[1])):
ypos = yfun(xpos)
try:
spec[jx] = np.sum(dat[ypos - extract_width:ypos +
extract_width, xpos])
except:
continue
try:
ll = f.get_lambda_nm()
except:
continue
specfun = interp1d(ll, spec, bounds_error=False)
specgrid[:, i] = specfun(lamgrid)
skyspec = np.median(specgrid, axis=1)
pl.step(lamgrid, skyspec, where='mid')
pl.xlabel("Wavelength [nm]")
pl.ylabel("Spec Irr [ph/10 m/nm]")
sumoff = 0.
sumscale = 0.
legend = ["Sky"]
for skyline in skylines:
roi = (lamgrid > skyline - skywidth) & (lamgrid < skyline + skywidth)
ffun = NFit.mpfit_residuals(NFit.gaussian4)
parinfo = [{
'value': np.max(skyspec[roi]),
'limited': [1, 0],
'limits': [0, 0]
}, {
'value': skyline
}, {
'value': 3
}, {
'value': np.min(skyspec[roi]),
'limited': [1, 0],
'limits': [0, 0]
}]
fit = NFit.mpfit_do(ffun, lamgrid[roi], skyspec[roi], parinfo)
if fit.status == 1:
off = fit.params[1] - skyline
sumoff += off * fit.params[0]
sumscale += fit.params[0]
pl.plot(lamgrid[roi], NFit.gaussian4(fit.params, lamgrid[roi]))
dxnm = fit.params[1] - skyline
legend.append("%.1f, %.2f" % (skyline, off))
else:
dxnm = 0.
print("line = %6.1f (%6.1f), status = %d, dX = %3.2f nm shift" %
(skyline, fit.params[0], fit.status, dxnm))
if sumscale > 0.:
dxnm = sumoff / sumscale
else:
print "Warning: no good skylines to fit! Setting X offset to 0.0 nm"
dxnm = 0.
print "dX = %3.2f nm shift" % dxnm
pl.title("dX = %3.2f nm shift, dY = %3.2f px shift" % (dxnm, drow))
pl.legend(legend)
pl.savefig(outfile + ".pdf")
return dxnm