def flexure_rssspec(imagefits,fitslist,option=""):
print str(datetime.now())
if option == "filesave": prefix = raw_input("\nFile prefix: ")
pixel = 15. # pixel size in microns
pix_scale=0.125
sexparams = ["X_IMAGE","Y_IMAGE","FLUX_ISO","FLUX_MAX","FLAGS","CLASS_STAR", \
"X2WIN_IMAGE","Y2WIN_IMAGE","XYWIN_IMAGE","ERRX2WIN_IMAGE"]
np.savetxt("qred_thrufoc.param",sexparams,fmt="%s")
fmaxcol,flagcol,xvarcol,yvarcol,xerrcol = (3,4,6,7,9) # column nos (from 0) of data in sextractor
imagestooclosefactor = 3.0 # too close if factor*sep < sqrt(var)
gaptooclose = 1.25 # arcsec
edgetooclose = 1.25 # arcsec
rattolerance = 0.25
toofaint = 250. # FMAX counts
galaxydelta = 0.4 # arcsec
MOSimagelimit = 1. # arcsec
deblend = .005 # default
imagehdr = pyfits.getheader(imagefits)
if imagehdr["GR-STATE"][1] == "4":
print "First fits file "+imagefits+" is not image of mask"
exit()
flexposns = len(fitslist)
obsdict=obslog(fitslist)
image_f = [fitslist[fpos].split(".")[0][-12:] for fpos in range(flexposns)]
dateobs = obsdict["DATE-OBS"][0].replace("-","")
if int(dateobs) > 20110928: rho_f = np.array(obsdict["TRKRHO"]).astype(float)
else: rho_f = np.array(obsdict["TELRHO"]).astype(float)
catpos = np.argmin(np.abs(rho_f))
cbin,rbin = np.array(obsdict["CCDSUM"][catpos].split(" ")).astype(int)
maskid = obsdict["MASKID"][catpos].strip()
filter = obsdict["FILTER"][catpos].strip()
grating = obsdict["GRATING"][catpos].strip()
rows,cols = pyfits.getdata(fitslist[catpos]).shape
isspec = (obsdict["GR-STATE"][catpos][1] =="4")
if not isspec:
print "Use flexure_rssimage for image flexure analysis"
exit()
grang = float(obsdict["GRTILT"][catpos])
artic = float(obsdict["CAMANG"][catpos])
lamp = obsdict["LAMPID"][catpos].strip()
print "\nMask: ", maskid
print "Filter: ", filter
print "Grating: ", grating
print "Artic (deg): ", artic
print "Gr Angle (deg): ", grang
print "Lamp: ", lamp
# map the mask spots _m using the imaging fits file
sex_js = sextract(imagefits,deblend=deblend)
flux_s = sex_js[2]
fluxmedian = np.median(np.sort(flux_s)[-10:])
okm_s = (flux_s > fluxmedian/10) # cull bogus spots
maskholes = okm_s.sum()
r_m = sex_js[1,okm_s]
c_m = sex_js[0,okm_s]
# find mask rows _R, tabulate
histr_b, binr_b = np.histogram(r_m,bins=rows/10,range=(0,rows))
bin0_R = np.where((histr_b[1:]>0) & (histr_b[:-1]==0))[0]
bin1_R = np.where((histr_b[1:]==0) & (histr_b[:-1]>0))[0]
maskRows = bin0_R.shape[0]
bin_m = np.digitize(r_m,binr_b) - 1
R_m = np.array([np.where((bin_m[m] >= bin0_R) & (bin_m[m] <= bin1_R))[0][0] \
for m in range(maskholes)])
# find mask cols _C, tabulate
histc_b, binc_b = np.histogram(c_m,bins=cols/10,range=(0,cols))
bin0_C = np.where((histc_b[1:]>0) & (histc_b[:-1]==0))[0]
bin1_C = np.where((histc_b[1:]==0) & (histc_b[:-1]>0))[0]
maskCols = bin0_C.shape[0]
bin_m = np.digitize(c_m,binc_b) - 1
C_m = np.array([np.where((bin_m[m] >= bin0_C) & (bin_m[m] <= bin1_C))[0][0] \
for m in range(maskholes)])
# identify mask center = optical axis
if maskid == 'P000000N99': # symmetric mask
Raxis = maskRows/2
Caxis = maskCols/2
elif maskid == 'P000000N03': # mask with centered cross
Raxis = np.where((np.argmax(histr_b) >= bin0_R) & (np.argmax(histr_b) <= bin1_R))[0][0]
Caxis = np.where((np.argmax(histc_b) >= bin0_C) & (np.argmax(histc_b) <= bin1_C))[0][0]
else:
print "Not a valid flexure mask"
exit()
maxis = np.where((R_m==Raxis)&(C_m==Caxis))[0][0]
raxis = r_m[maxis]
caxis = c_m[maxis]
print "\nMask_Holes Rows Cols r axis c axis \n pixels pixels"
print " %5i %5i %5i %8.1f %8.1f" % (maskholes,maskRows,maskCols,raxis*rbin,caxis*cbin)
# np.savetxt(dateobs+'_'+"mask.txt",np.vstack((r_m,c_m,sex_js[2,okm_s],R_m)).T,fmt="%10.2f")
# get linelist, predict spots in spectral image
wavcent = rsslam(grating, grang, artic, 0.,dateobs)
specfile = datedfile(datadir+"spectrograph/spec_yyyymmdd.txt",dateobs)
FCampoly=np.loadtxt(specfile,usecols=(1,))[5:11]
fcam = np.polyval(FCampoly,(wavcent/1000. - 4.))
lampfile=iraf.osfn("pysalt$data/linelists/"+lamp+".salt")
wav_l,int_l = np.loadtxt(lampfile,unpack=True)
maxdalpha = -np.degrees((cols/2)*cbin*pixel/(1000.*fcam))
maxgamma = np.degrees((rows/2)*rbin*pixel/(1000.*fcam))
maxwav = rsslam(grating,grang,artic, cols*cbin/2,dateobs,-maxdalpha,0)
minwav = rsslam(grating,grang,artic,-cols*cbin/2,dateobs, maxdalpha,maxgamma)
ok_l = (wav_l >= minwav) & (wav_l <= maxwav)
wav_l = wav_l[ok_l]
int_l = int_l[ok_l]
lines = wav_l.shape[0]
col_ml = np.zeros((maskholes,lines))
dcol_c = np.arange(-(cols*cbin/2),(cols*cbin/2))
for m in range(maskholes):
dalpha = -np.degrees((c_m[m]-caxis)*cbin*pixel/(1000.*fcam))
gamma = np.degrees((r_m[m]-raxis)*rbin*pixel/(1000.*fcam))
wav0,wav1 = rsslam(grating,grang,artic,dcol_c[[0,-1]],dateobs,dalpha,gamma=gamma)
ok_l = ((wav_l > wav0) & (wav_l < wav1))
colwav = interp1d(rsslam(grating,grang,artic,dcol_c, \
dateobs,dalpha=dalpha,gamma=gamma), dcol_c)
col_ml[m,ok_l] = colwav(wav_l[ok_l]) + caxis*cbin
# np.savetxt(dateobs+"_col_ml.txt",np.vstack((R_m,C_m,col_ml.T)),fmt="%8.1f")
# identify mask hole and wavelength for spots in spec image closest to rho=0
os.remove("sexwt.fits")
sex_js = sextract(fitslist[catpos],"",deblend=deblend)
r_s = sex_js[1]
c_s = sex_js[0]
flux_s = sex_js[2]
spots = r_s.shape[0]
fluxmedian = np.median(np.sort(sex_js[2])[-10:])
ok_s = (flux_s > fluxmedian/30) # cull bogus spots
# find spectral bin rows RR in candidates R0, cull non-spectra
histr_b, binr_b = np.histogram(r_s[ok_s],bins=rows/10,range=(0,rows))
histr_b[[0,-1]] = 0
bin0_R0 = np.where((histr_b[1:]>0) & (histr_b[:-1]==0))[0] + 1
bin1_R0 = np.where((histr_b[1:]==0) & (histr_b[:-1]>0))[0]
bin_s = np.digitize(r_s,binr_b) - 1
maxcount_R0 = np.array([(histr_b[bin0_R0[R0]:bin1_R0[R0]+1]).max() \
for R0 in range(bin0_R0.shape[0])])
ok_R0 = (maxcount_R0 > 3)
specrows = ok_R0.sum() # cull down to spectra RR
bin0_RR = bin0_R0[ok_R0]
bin1_RR = bin1_R0[ok_R0]
ok_s &= ((bin_s >= bin0_RR[:,None]) & (bin_s <= bin1_RR[:,None])).any(axis=0)
RR_s = -np.ones(spots)
r_RR = np.zeros(specrows)
for RR in range(specrows):
isRR_s = ok_s & np.in1d(bin_s,np.arange(bin0_RR[RR],bin1_RR[RR]+1))
RR_s[isRR_s] = RR
r_RR[RR] = r_s[isRR_s].mean()
count_RR = (RR_s[:,None]==range(specrows)).sum(axis=0)
if maskid == 'P000000N99':
RRaxis = np.argmin((raxis-r_RR)**2)
elif maskid == 'P000000N03':
RRaxis = np.argmax(count_RR)
# cull weak lines
ptile = 100.*min(1.,5.*maskCols/count_RR.max()) # want like 5 brightest lines
for RR in range(specrows):
isRR_s = ok_s & np.in1d(bin_s,np.arange(bin0_RR[RR],bin1_RR[RR]+1))
fluxmin = np.percentile(sex_js[2,isRR_s],100.-ptile)
ok_s[isRR_s] &= (sex_js[2,isRR_s] > fluxmin)
# identify with mask rows R (assuming no gaps)
RR_m = R_m + RRaxis - Raxis
# find approximate grating shift in dispersion direction by looking for most common id error
histc_b = np.zeros(60)
for RR in range(specrows):
isRR_s = ((RR_s==RR) & ok_s)
cerr_MS = (c_s[None,isRR_s] - col_ml[RR_m==RR].ravel()[:,None])
histc_b += np.histogram(cerr_MS.ravel(),bins=60,range=(-150,150))[0]
cshift = 5*np.argmax(histc_b) - 150
col_ml += cshift
# identify wavelength and mask column with spots in each spectrum
isfound_s = np.zeros((spots),dtype=bool)
bintol = 16/cbin # 2 arcsec tolerance for line ID
R_s = -np.ones(spots,dtype=int)
C_s = -np.ones(spots,dtype=int)
l_s = -np.ones(spots,dtype=int)
m_s = -np.ones(spots,dtype=int)
cerr_s = np.zeros(spots)
rmscol = 0.
for RR in range(specrows): # _S spot in spectrum, _P (mask column, line)
isRR_m = (RR_m==RR)
isRR_s = ((RR_s==RR) & ok_s)
cerr_PS = (c_s[None,isRR_s] - col_ml[isRR_m].ravel()[:,None])
Spots = isRR_s.sum()
Possibles = col_ml[isRR_m].size
Cols = Possibles/lines
P_S = np.argmin(np.abs(cerr_PS),axis=0)
cerr_S = cerr_PS[P_S,range(isRR_s.sum())]
isfound_S = (np.abs(cerr_S) < bintol)
M_P,l_P = np.unravel_index(np.arange(Possibles),(Cols,lines))
m_P = np.where(isRR_m)[0][M_P]
m_S = m_P[P_S]
C_P = C_m[m_P]
C_S = C_P[P_S]
l_S = l_P[P_S]
s_S = np.where(isRR_s)[0]
R_s[isRR_s] = RR + Raxis-RRaxis
cerr_s[s_S] = cerr_S
C_s[s_S[isfound_S]] = C_S[isfound_S]
l_s[s_S[isfound_S]] = l_S[isfound_S]
m_s[s_S[isfound_S]] = m_S[isfound_S]
isfound_s[s_S] |= isfound_S
rmscol += (cerr_S[isfound_S]**2).sum()
# cull wavelengths to _L with < 1/2 Mask Rows or Cols
ok_s &= isfound_s
ok_l = np.zeros((lines),dtype=bool)
for line in range(lines):
lRows = np.unique(R_s[l_s==line]).shape[0]
lCols = np.unique(C_s[l_s==line]).shape[0]
ok_l[line] = ((lRows>=maskRows/2) & (lCols>=maskCols/2))
l_L = np.where(ok_l)[0]
wav_L = wav_l[l_L]
Lines = l_L.shape[0]
ok_s &= np.in1d(l_s,l_L)
# tabulate good catalog spots (final _S)
s_S = np.where(ok_s)[0]
r_S = r_s[s_S]
c_S = c_s[s_S]
cerr_S = cerr_s[s_S]
R_S = R_s[s_S]
C_S = C_s[s_S]
l_S = l_s[s_S]
Spots = ok_s.sum()
rshift = r_S[R_S==Raxis].mean() - raxis
cshift += (c_S - col_ml[m_s[s_S],l_S]).mean()
rmscol = np.sqrt(rmscol/Spots)
np.savetxt("cat_S.txt",np.vstack((s_S,r_S,c_S,R_S,C_S,l_S,cerr_S)).T, \
fmt="%5i %8.2f %8.2f %5i %5i %5i %8.2f")
print "\nSpec_Spots Lines rshift cshift rms\n pixels pixels pixels"
print " %5i %5i %8.1f %8.1f %8.1f" % (Spots,np.unique(l_S).shape[0],rshift,cshift,rmscol)
print "\nLineno Wavel spots Rows Cols"
for L in range(Lines):
line = l_L[L]
lRows = np.unique(R_S[l_S==line]).shape[0]
lCols = np.unique(C_S[l_S==line]).shape[0]
lspots = (l_S==line).sum()
print " %5i %8.2f %5i %5i %5i" % (line,wav_l[line],lspots,lRows,lCols)
sexcols = sex_js.shape[0]
sexdata_jfS = np.zeros((sexcols,flexposns,Spots))
sexdata_jfS[:,catpos] = sex_js[:,ok_s]
xcenter_L = col_ml[maxis,l_L]
ycenter = raxis + rshift
if option == "filesave":
np.savetxt(prefix+"Spots.txt",sexdata_jfS[:,catpos].T, \
fmt=2*"%9.2f "+"%9.0f "+"%9.1f "+"%4i "+"%6.2f "+3*"%7.2f "+"%11.3e")
# find spots in flexure series, in order of increasing abs(rho), and store sextractor output
row_fLd = np.zeros((flexposns,Lines,2))
col_fLd = np.zeros((flexposns,Lines,2))
print "\n fits rho line spots rshift cshift rslope cslope rmserr "
print " deg Ang arcsec arcsec arcmin arcmin bins"
for dirn in (1,-1):
refpos = catpos
posdirlist = np.argsort(dirn*rho_f)
poslist = posdirlist[dirn*rho_f[posdirlist] > rho_f[refpos]]
for fpos in poslist:
col_S,row_S = sexdata_jfS[0:2,refpos,:]
sex_js = sextract(fitslist[fpos],"sexwt.fits",deblend=deblend)
binsqerr_sS = (sex_js[1,:,None] - row_S[None,:])**2 + (sex_js[0,:,None] - col_S[None,:])**2
S_s = np.argmin(binsqerr_sS,axis=1)
# First compute image shift by averaging small errors
rowerr_s = sex_js[1] - row_S[S_s]
colerr_s = sex_js[0] - col_S[S_s]
hist_r,bin_r = np.histogram(rowerr_s,bins=32,range=(-2*bintol,2*bintol))
drow = rowerr_s[(rowerr_s > bin_r[np.argmax(hist_r)]-bintol) & \
(rowerr_s < bin_r[np.argmax(hist_r)]+bintol)].mean()
hist_c,bin_c = np.histogram(colerr_s,bins=32,range=(-2*bintol,2*bintol))
dcol = colerr_s[(colerr_s > bin_c[np.argmax(hist_c)]-bintol) & \
(colerr_s < bin_c[np.argmax(hist_c)]+bintol)].mean()
# Now refind the closest ID
binsqerr_sS = (sex_js[1,:,None] - row_S[None,:] -drow)**2 + \
(sex_js[0,:,None] - col_S[None,:] -dcol)**2
binsqerr_s = binsqerr_sS.min(axis=1)
isfound_s = binsqerr_s < bintol**2
S_s = np.argmin(binsqerr_sS,axis=1)
isfound_s &= (binsqerr_s == binsqerr_sS[:,S_s].min(axis=0))
isfound_S = np.array([S in S_s[isfound_s] for S in range(Spots)])
sexdata_jfS[:,fpos,S_s[isfound_s]] = sex_js[:,isfound_s]
drow_S = sexdata_jfS[1,fpos]-sexdata_jfS[1,catpos]
dcol_S = sexdata_jfS[0,fpos]-sexdata_jfS[0,catpos]
# np.savetxt("motion_"+str(fpos)+".txt",np.vstack((isfound_S,l_S,drow_S,dcol_S)).T,fmt="%3i %3i %8.2f %8.2f")
# Compute flexure image motion parameters for each line
for L in range(Lines):
ok_S = ((l_S == l_L[L]) & isfound_S)
row_fLd[fpos,L],rowchi,d,d,d = \
np.polyfit(sexdata_jfS[0,catpos,ok_S]-xcenter_L[L],drow_S[ok_S],deg=1,full=True)
col_fLd[fpos,L],colchi,d,d,d = \
np.polyfit(sexdata_jfS[1,catpos,ok_S]-ycenter,dcol_S[ok_S],deg=1,full=True)
rms = np.sqrt((rowchi+colchi)/(2*ok_S.sum()))
print ("%12s %5.0f %5i %5i "+5*"%7.2f ") % (image_f[fpos], rho_f[fpos], wav_L[L], \
ok_S.sum(),row_fLd[fpos,L,1]*rbin*pix_scale, col_fLd[fpos,L,1]*cbin*pix_scale, \
60.*np.degrees(row_fLd[fpos,L,0]),-60.*np.degrees(col_fLd[fpos,L,0]), rms)
if option == "filesave":
np.savetxt(prefix+"flex_"+str(fpos)+".txt",np.vstack((isfound_S,drow_S,dcol_S)).T, \
fmt = "%2i %8.3f %8.3f")
np.savetxt(prefix+"sextr_"+str(fpos)+".txt",sexdata_jfS[:,fpos].T)
print
# make plots
fig,plot_s = plt.subplots(2,1,sharex=True)
plt.xlabel('Rho (deg)')
plt.xlim(-120,120)
plt.xticks(range(-120,120,30))
fig.set_size_inches((8.5,11))
fig.subplots_adjust(left=0.175)
plot_s[0].set_title(str(dateobs)+[" Imaging"," Spectral"][isspec]+" Flexure")
plot_s[0].set_ylabel('Mean Position (arcsec)')
plot_s[0].set_ylim(-0.5,4.)
plot_s[1].set_ylabel('Rotation (arcmin ccw)')
plot_s[1].set_ylim(-10.,6.)
lbl_L = [("%5.0f") % (wav_L[L]) for L in range(Lines)]
color_L = 'bgrcmykw'
for L in range(Lines):
plot_s[0].plot(rho_f,row_fLd[:,L,1]*rbin*pix_scale, \
color=color_L[L],marker='D',markersize=8,label='row '+lbl_L[L])
plot_s[1].plot(rho_f,60.*np.degrees(row_fLd[:,L,0]),
color=color_L[L],marker='D',markersize=8,label='row '+lbl_L[L])
collbl = 'col'+lbl_L[0]
for L in range(Lines):
plot_s[0].plot(rho_f,col_fLd[:,L,1]*cbin*pix_scale, \
color=color_L[L],marker='s',markersize=8,label=collbl)
plot_s[1].plot(rho_f,-60.*np.degrees(col_fLd[:,L,0]), \
color=color_L[L],marker='s',markersize=8,label=collbl)
collbl = ''
plot_s[0].legend(fontsize='medium',loc='upper center')
plotfile = str(dateobs)+['_imflex.pdf','_grflex.pdf'][isspec]
plt.savefig(plotfile,orientation='portrait')
if os.name=='posix':
if os.popen('ps -C evince -f').read().count(plotfile)==0: os.system('evince '+plotfile+' &')
os.remove("out.txt")
os.remove("qred_thrufoc.param")
os.remove("sexwt.fits")
return
def flexure_rss(fitslist, option=""):
global sex_js, rd, B, pixarcsec, gridsize, niter
global rho_f, rc0_dgg, usespots_gg, fwhminterp_g # for cube analysis
if option == "filesave": prefix = raw_input("file prefix: ")
pixel = 15. # pixel size in microns
pix_scale = 0.125
sexparams = ["X_IMAGE","Y_IMAGE","FLUX_ISO","FLUX_MAX","FLAGS","CLASS_STAR", \
"X2WIN_IMAGE","Y2WIN_IMAGE","XYWIN_IMAGE","ERRX2WIN_IMAGE"]
np.savetxt("qred_thrufoc.param", sexparams, fmt="%s")
fmaxcol, flagcol, xvarcol, yvarcol, xerrcol = (
3, 4, 6, 7, 9) # column nos (from 0) of data in sextractor
imagestooclosefactor = 3.0 # too close if factor*sep < sqrt(var)
gaptooclose = 1.25 # arcsec
edgetooclose = 1.25 # arcsec
rattolerance = 0.25
toofaint = 250. # FMAX counts
galaxydelta = 0.4 # arcsec
MOSimagelimit = 1. # arcsec
deblend = .005 # default
flexposns = len(fitslist)
obsdict = obslog(fitslist)
image_f = [fitslist[fpos].split(".")[0][-12:] for fpos in range(flexposns)]
dateobs = int(image_f[0][:8])
if dateobs > 20110928: rho_f = np.array(obsdict["TRKRHO"]).astype(float)
else: rho_f = np.array(obsdict["TELRHO"]).astype(float)
catpos = np.argmin(np.abs(rho_f))
cbin, rbin = np.array(obsdict["CCDSUM"][catpos].split(" ")).astype(int)
maskid = obsdict["MASKID"][catpos].strip()
filter = obsdict["FILTER"][catpos].strip()
grating = obsdict["GRATING"][catpos].strip()
rows, cols = pyfits.getdata(fitslist[catpos]).shape
isspec = (obsdict["GR-STATE"][catpos][1] == "4")
print str(datetime.now()), "\n"
print "Mask: ", maskid
print "Filter: ", filter
print "Grating: ", grating
# make catalog of stars using image closest to rho=0 (capital S)
sex_js = sextract(fitslist[catpos], deblend=deblend)
fluxisomedian = np.median(np.sort(
sex_js[2])[-10:]) # median of 10 brightest
ok_s = sex_js[2] > fluxisomedian / 100. # get rid of bogus stars
sexcols = sex_js.shape[0]
Stars = ok_s.sum()
sexdata_jfS = np.zeros((sexcols, flexposns, Stars))
sexdata_jfS[:, catpos] = sex_js[:, ok_s]
xcenter = 0.5 * (sexdata_jfS[0, catpos].min() +
sexdata_jfS[0, catpos].max())
ycenter = 0.5 * (sexdata_jfS[1, catpos].min() +
sexdata_jfS[1, catpos].max())
print "\n fits rho stars rshift cshift rslope cslope rmserr "
print " deg arcsec arcsec arcmin arcmin bins"
print ("%12s %5.1f %5i "+5*"%7.2f ") % \
(image_f[catpos], rho_f[catpos], Stars, 0., 0., 0., 0., 0.)
if option == "filesave":
np.savetxt(prefix+"Stars.txt",sexdata_jfS[:,catpos].T, \
fmt=2*"%9.2f "+"%9.0f "+"%9.1f "+"%4i "+"%6.2f "+3*"%7.2f "+"%11.3e")
# find stars in flexure series, in order of increasing abs(rho), and store sextractor output
row_fd = np.zeros((flexposns, 2))
col_fd = np.zeros((flexposns, 2))
for dirn in (1, -1):
refpos = catpos
posdirlist = np.argsort(dirn * rho_f)
poslist = posdirlist[dirn * rho_f[posdirlist] > rho_f[refpos]]
for fpos in poslist:
col_S, row_S = sexdata_jfS[0:2, refpos, :]
sex_js = sextract(fitslist[fpos], "sexwt.fits", deblend=deblend)
bintol = 16 / cbin # 2 arcsec tolerance for finding star
binsqerr_sS = (sex_js[1, :, None] - row_S[None, :])**2 + (
sex_js[0, :, None] - col_S[None, :])**2
S_s = np.argmin(binsqerr_sS, axis=1)
# First compute image shift by averaging small errors
rowerr_s = sex_js[1] - row_S[S_s]
colerr_s = sex_js[0] - col_S[S_s]
hist_r, bin_r = np.histogram(rowerr_s,
bins=32,
range=(-2 * bintol, 2 * bintol))
drow = rowerr_s[(rowerr_s > bin_r[np.argmax(hist_r)]-bintol) & \
(rowerr_s < bin_r[np.argmax(hist_r)]+bintol)].mean()
hist_c, bin_c = np.histogram(colerr_s,
bins=32,
range=(-2 * bintol, 2 * bintol))
dcol = colerr_s[(colerr_s > bin_r[np.argmax(hist_r)]-bintol) & \
(colerr_s < bin_r[np.argmax(hist_r)]+bintol)].mean()
# Now refind the closest ID
binsqerr_sS = (sex_js[1,:,None] - row_S[None,:] -drow)**2 + \
(sex_js[0,:,None] - col_S[None,:] -dcol)**2
binsqerr_s = binsqerr_sS.min(axis=1)
isfound_s = binsqerr_s < bintol**2
S_s = np.argmin(binsqerr_sS, axis=1)
isfound_s &= (binsqerr_s == binsqerr_sS[:, S_s].min(axis=0))
isfound_S = np.array([S in S_s[isfound_s] for S in range(Stars)])
sexdata_jfS[:, fpos, S_s[isfound_s]] = sex_js[:, isfound_s]
drow_S = sexdata_jfS[1, fpos] - sexdata_jfS[1, catpos]
dcol_S = sexdata_jfS[0, fpos] - sexdata_jfS[0, catpos]
row_fd[fpos],rowchi,d,d,d = np.polyfit(sexdata_jfS[0,catpos,isfound_S]-xcenter, \
drow_S[isfound_S],deg=1,full=True)
col_fd[fpos],colchi,d,d,d = np.polyfit(sexdata_jfS[1,catpos,isfound_S]-ycenter, \
dcol_S[isfound_S],deg=1,full=True)
rms = np.sqrt((rowchi + colchi) / (2 * isfound_S.sum()))
print ("%12s %5.0f %5i "+5*"%7.2f ") % (image_f[fpos], rho_f[fpos], isfound_S.sum(), \
row_fd[fpos,1]*rbin*pix_scale, col_fd[fpos,1]*cbin*pix_scale, \
60.*np.degrees(row_fd[fpos,0]),-60.*np.degrees(col_fd[fpos,0]), rms)
if option == "filesave":
np.savetxt(prefix+"flex_"+str(fpos)+".txt",np.vstack((isfound_S,drow_S,dcol_S)).T, \
fmt = "%2i %8.3f %8.3f")
np.savetxt(prefix + "sextr_" + str(fpos) + ".txt",
sexdata_jfS[:, fpos].T)
# make plots
fig, plot_s = plt.subplots(2, 1, sharex=True)
plt.xlabel('Rho (deg)')
plt.xlim(-120, 120)
plt.xticks(range(-120, 120, 30))
fig.set_size_inches((8.5, 11))
fig.subplots_adjust(left=0.175)
plot_s[0].set_title(str(dateobs) + " Imaging Flexure")
plot_s[0].set_ylabel('Mean Position (arcsec)')
plot_s[0].set_ylim(-0.5, 2.)
plot_s[1].set_ylabel('Rotation (arcmin ccw)')
plot_s[1].set_ylim(-6., 6.)
plot_s[0].plot(rho_f,
row_fd[:, 1] * rbin * pix_scale,
marker='D',
label='row')
plot_s[0].plot(rho_f,
col_fd[:, 1] * rbin * pix_scale,
marker='D',
label='col')
plot_s[1].plot(rho_f,
60. * np.degrees(row_fd[:, 0]),
marker='D',
label='row')
plot_s[1].plot(rho_f,
-60. * np.degrees(col_fd[:, 0]),
marker='D',
label='col')
plot_s[0].legend(fontsize='medium', loc='upper center')
plotfile = str(dateobs) + '_imflex.pdf'
plt.savefig(plotfile, orientation='portrait')
if os.name == 'posix':
if os.popen('ps -C evince -f').read().count(plotfile) == 0:
os.system('evince ' + plotfile + ' &')
os.remove("out.txt")
os.remove("qred_thrufoc.param")
os.remove("sexwt.fits")
return
def flexure_rssspec(imagefits, fitslist, option=""):
print str(datetime.now())
if option == "filesave": prefix = raw_input("\nFile prefix: ")
pixel = 15. # pixel size in microns
pix_scale = 0.125
sexparams = ["X_IMAGE","Y_IMAGE","FLUX_ISO","FLUX_MAX","FLAGS","CLASS_STAR", \
"X2WIN_IMAGE","Y2WIN_IMAGE","XYWIN_IMAGE","ERRX2WIN_IMAGE"]
np.savetxt("qred_thrufoc.param", sexparams, fmt="%s")
fmaxcol, flagcol, xvarcol, yvarcol, xerrcol = (
3, 4, 6, 7, 9) # column nos (from 0) of data in sextractor
imagestooclosefactor = 3.0 # too close if factor*sep < sqrt(var)
gaptooclose = 1.25 # arcsec
edgetooclose = 1.25 # arcsec
rattolerance = 0.25
toofaint = 250. # FMAX counts
galaxydelta = 0.4 # arcsec
MOSimagelimit = 1. # arcsec
deblend = .005 # default
imagehdr = pyfits.getheader(imagefits)
if imagehdr["GR-STATE"][1] == "4":
print "First fits file " + imagefits + " is not image of mask"
exit()
flexposns = len(fitslist)
obsdict = obslog(fitslist)
image_f = [fitslist[fpos].split(".")[0][-12:] for fpos in range(flexposns)]
dateobs = obsdict["DATE-OBS"][0].replace("-", "")
if int(dateobs) > 20110928:
rho_f = np.array(obsdict["TRKRHO"]).astype(float)
else:
rho_f = np.array(obsdict["TELRHO"]).astype(float)
catpos = np.argmin(np.abs(rho_f))
cbin, rbin = np.array(obsdict["CCDSUM"][catpos].split(" ")).astype(int)
maskid = obsdict["MASKID"][catpos].strip()
filter = obsdict["FILTER"][catpos].strip()
grating = obsdict["GRATING"][catpos].strip()
rows, cols = pyfits.getdata(fitslist[catpos]).shape
isspec = (obsdict["GR-STATE"][catpos][1] == "4")
if not isspec:
print "Use flexure_rssimage for image flexure analysis"
exit()
grang = float(obsdict["GRTILT"][catpos])
artic = float(obsdict["CAMANG"][catpos])
lamp = obsdict["LAMPID"][catpos].strip()
print "\nMask: ", maskid
print "Filter: ", filter
print "Grating: ", grating
print "Artic (deg): ", artic
print "Gr Angle (deg): ", grang
print "Lamp: ", lamp
# map the mask spots _m using the imaging fits file
sex_js = sextract(imagefits, deblend=deblend)
flux_s = sex_js[2]
fluxmedian = np.median(np.sort(flux_s)[-10:])
okm_s = (flux_s > fluxmedian / 10) # cull bogus spots
maskholes = okm_s.sum()
r_m = sex_js[1, okm_s]
c_m = sex_js[0, okm_s]
# find mask rows _R, tabulate
histr_b, binr_b = np.histogram(r_m, bins=rows / 10, range=(0, rows))
bin0_R = np.where((histr_b[1:] > 0) & (histr_b[:-1] == 0))[0]
bin1_R = np.where((histr_b[1:] == 0) & (histr_b[:-1] > 0))[0]
maskRows = bin0_R.shape[0]
bin_m = np.digitize(r_m, binr_b) - 1
R_m = np.array([np.where((bin_m[m] >= bin0_R) & (bin_m[m] <= bin1_R))[0][0] \
for m in range(maskholes)])
# find mask cols _C, tabulate
histc_b, binc_b = np.histogram(c_m, bins=cols / 10, range=(0, cols))
bin0_C = np.where((histc_b[1:] > 0) & (histc_b[:-1] == 0))[0]
bin1_C = np.where((histc_b[1:] == 0) & (histc_b[:-1] > 0))[0]
maskCols = bin0_C.shape[0]
bin_m = np.digitize(c_m, binc_b) - 1
C_m = np.array([np.where((bin_m[m] >= bin0_C) & (bin_m[m] <= bin1_C))[0][0] \
for m in range(maskholes)])
# identify mask center = optical axis
if maskid == 'P000000N99': # symmetric mask
Raxis = maskRows / 2
Caxis = maskCols / 2
elif maskid == 'P000000N03': # mask with centered cross
Raxis = np.where((np.argmax(histr_b) >= bin0_R)
& (np.argmax(histr_b) <= bin1_R))[0][0]
Caxis = np.where((np.argmax(histc_b) >= bin0_C)
& (np.argmax(histc_b) <= bin1_C))[0][0]
else:
print "Not a valid flexure mask"
exit()
maxis = np.where((R_m == Raxis) & (C_m == Caxis))[0][0]
raxis = r_m[maxis]
caxis = c_m[maxis]
print "\nMask_Holes Rows Cols r axis c axis \n pixels pixels"
print " %5i %5i %5i %8.1f %8.1f" % (maskholes, maskRows, maskCols,
raxis * rbin, caxis * cbin)
# np.savetxt(dateobs+'_'+"mask.txt",np.vstack((r_m,c_m,sex_js[2,okm_s],R_m)).T,fmt="%10.2f")
# get linelist, predict spots in spectral image
wavcent = rsslam(grating, grang, artic, 0., dateobs)
specfile = datedfile(datadir + "spectrograph/spec_yyyymmdd.txt", dateobs)
FCampoly = np.loadtxt(specfile, usecols=(1, ))[5:11]
fcam = np.polyval(FCampoly, (wavcent / 1000. - 4.))
lampfile = iraf.osfn("pysalt$data/linelists/" + lamp + ".salt")
wav_l, int_l = np.loadtxt(lampfile, unpack=True)
maxdalpha = -np.degrees((cols / 2) * cbin * pixel / (1000. * fcam))
maxgamma = np.degrees((rows / 2) * rbin * pixel / (1000. * fcam))
maxwav = rsslam(grating, grang, artic, cols * cbin / 2, dateobs,
-maxdalpha, 0)
minwav = rsslam(grating, grang, artic, -cols * cbin / 2, dateobs,
maxdalpha, maxgamma)
ok_l = (wav_l >= minwav) & (wav_l <= maxwav)
wav_l = wav_l[ok_l]
int_l = int_l[ok_l]
lines = wav_l.shape[0]
col_ml = np.zeros((maskholes, lines))
dcol_c = np.arange(-(cols * cbin / 2), (cols * cbin / 2))
for m in range(maskholes):
dalpha = -np.degrees((c_m[m] - caxis) * cbin * pixel / (1000. * fcam))
gamma = np.degrees((r_m[m] - raxis) * rbin * pixel / (1000. * fcam))
wav0, wav1 = rsslam(grating,
grang,
artic,
dcol_c[[0, -1]],
dateobs,
dalpha,
gamma=gamma)
ok_l = ((wav_l > wav0) & (wav_l < wav1))
colwav = interp1d(rsslam(grating,grang,artic,dcol_c, \
dateobs,dalpha=dalpha,gamma=gamma), dcol_c)
col_ml[m, ok_l] = colwav(wav_l[ok_l]) + caxis * cbin
# np.savetxt(dateobs+"_col_ml.txt",np.vstack((R_m,C_m,col_ml.T)),fmt="%8.1f")
# identify mask hole and wavelength for spots in spec image closest to rho=0
os.remove("sexwt.fits")
sex_js = sextract(fitslist[catpos], "", deblend=deblend)
r_s = sex_js[1]
c_s = sex_js[0]
flux_s = sex_js[2]
spots = r_s.shape[0]
fluxmedian = np.median(np.sort(sex_js[2])[-10:])
ok_s = (flux_s > fluxmedian / 30) # cull bogus spots
# find spectral bin rows RR in candidates R0, cull non-spectra
histr_b, binr_b = np.histogram(r_s[ok_s], bins=rows / 10, range=(0, rows))
histr_b[[0, -1]] = 0
bin0_R0 = np.where((histr_b[1:] > 0) & (histr_b[:-1] == 0))[0] + 1
bin1_R0 = np.where((histr_b[1:] == 0) & (histr_b[:-1] > 0))[0]
bin_s = np.digitize(r_s, binr_b) - 1
maxcount_R0 = np.array([(histr_b[bin0_R0[R0]:bin1_R0[R0]+1]).max() \
for R0 in range(bin0_R0.shape[0])])
ok_R0 = (maxcount_R0 > 3)
specrows = ok_R0.sum() # cull down to spectra RR
bin0_RR = bin0_R0[ok_R0]
bin1_RR = bin1_R0[ok_R0]
ok_s &= ((bin_s >= bin0_RR[:, None]) &
(bin_s <= bin1_RR[:, None])).any(axis=0)
RR_s = -np.ones(spots)
r_RR = np.zeros(specrows)
for RR in range(specrows):
isRR_s = ok_s & np.in1d(bin_s, np.arange(bin0_RR[RR], bin1_RR[RR] + 1))
RR_s[isRR_s] = RR
r_RR[RR] = r_s[isRR_s].mean()
count_RR = (RR_s[:, None] == range(specrows)).sum(axis=0)
if maskid == 'P000000N99':
RRaxis = np.argmin((raxis - r_RR)**2)
elif maskid == 'P000000N03':
RRaxis = np.argmax(count_RR)
# cull weak lines
ptile = 100. * min(
1., 5. * maskCols / count_RR.max()) # want like 5 brightest lines
for RR in range(specrows):
isRR_s = ok_s & np.in1d(bin_s, np.arange(bin0_RR[RR], bin1_RR[RR] + 1))
fluxmin = np.percentile(sex_js[2, isRR_s], 100. - ptile)
ok_s[isRR_s] &= (sex_js[2, isRR_s] > fluxmin)
# identify with mask rows R (assuming no gaps)
RR_m = R_m + RRaxis - Raxis
# find approximate grating shift in dispersion direction by looking for most common id error
histc_b = np.zeros(60)
for RR in range(specrows):
isRR_s = ((RR_s == RR) & ok_s)
cerr_MS = (c_s[None, isRR_s] - col_ml[RR_m == RR].ravel()[:, None])
histc_b += np.histogram(cerr_MS.ravel(), bins=60, range=(-150, 150))[0]
cshift = 5 * np.argmax(histc_b) - 150
col_ml += cshift
# identify wavelength and mask column with spots in each spectrum
isfound_s = np.zeros((spots), dtype=bool)
bintol = 16 / cbin # 2 arcsec tolerance for line ID
R_s = -np.ones(spots, dtype=int)
C_s = -np.ones(spots, dtype=int)
l_s = -np.ones(spots, dtype=int)
m_s = -np.ones(spots, dtype=int)
cerr_s = np.zeros(spots)
rmscol = 0.
for RR in range(specrows): # _S spot in spectrum, _P (mask column, line)
isRR_m = (RR_m == RR)
isRR_s = ((RR_s == RR) & ok_s)
cerr_PS = (c_s[None, isRR_s] - col_ml[isRR_m].ravel()[:, None])
Spots = isRR_s.sum()
Possibles = col_ml[isRR_m].size
Cols = Possibles / lines
P_S = np.argmin(np.abs(cerr_PS), axis=0)
cerr_S = cerr_PS[P_S, range(isRR_s.sum())]
isfound_S = (np.abs(cerr_S) < bintol)
M_P, l_P = np.unravel_index(np.arange(Possibles), (Cols, lines))
m_P = np.where(isRR_m)[0][M_P]
m_S = m_P[P_S]
C_P = C_m[m_P]
C_S = C_P[P_S]
l_S = l_P[P_S]
s_S = np.where(isRR_s)[0]
R_s[isRR_s] = RR + Raxis - RRaxis
cerr_s[s_S] = cerr_S
C_s[s_S[isfound_S]] = C_S[isfound_S]
l_s[s_S[isfound_S]] = l_S[isfound_S]
m_s[s_S[isfound_S]] = m_S[isfound_S]
isfound_s[s_S] |= isfound_S
rmscol += (cerr_S[isfound_S]**2).sum()
# cull wavelengths to _L with < 1/2 Mask Rows or Cols
ok_s &= isfound_s
ok_l = np.zeros((lines), dtype=bool)
for line in range(lines):
lRows = np.unique(R_s[l_s == line]).shape[0]
lCols = np.unique(C_s[l_s == line]).shape[0]
ok_l[line] = ((lRows >= maskRows / 2) & (lCols >= maskCols / 2))
l_L = np.where(ok_l)[0]
wav_L = wav_l[l_L]
Lines = l_L.shape[0]
ok_s &= np.in1d(l_s, l_L)
# tabulate good catalog spots (final _S)
s_S = np.where(ok_s)[0]
r_S = r_s[s_S]
c_S = c_s[s_S]
cerr_S = cerr_s[s_S]
R_S = R_s[s_S]
C_S = C_s[s_S]
l_S = l_s[s_S]
Spots = ok_s.sum()
rshift = r_S[R_S == Raxis].mean() - raxis
cshift += (c_S - col_ml[m_s[s_S], l_S]).mean()
rmscol = np.sqrt(rmscol / Spots)
np.savetxt("cat_S.txt",np.vstack((s_S,r_S,c_S,R_S,C_S,l_S,cerr_S)).T, \
fmt="%5i %8.2f %8.2f %5i %5i %5i %8.2f")
print "\nSpec_Spots Lines rshift cshift rms\n pixels pixels pixels"
print " %5i %5i %8.1f %8.1f %8.1f" % (Spots, np.unique(l_S).shape[0],
rshift, cshift, rmscol)
print "\nLineno Wavel spots Rows Cols"
for L in range(Lines):
line = l_L[L]
lRows = np.unique(R_S[l_S == line]).shape[0]
lCols = np.unique(C_S[l_S == line]).shape[0]
lspots = (l_S == line).sum()
print " %5i %8.2f %5i %5i %5i" % (line, wav_l[line], lspots, lRows,
lCols)
sexcols = sex_js.shape[0]
sexdata_jfS = np.zeros((sexcols, flexposns, Spots))
sexdata_jfS[:, catpos] = sex_js[:, ok_s]
xcenter_L = col_ml[maxis, l_L]
ycenter = raxis + rshift
if option == "filesave":
np.savetxt(prefix+"Spots.txt",sexdata_jfS[:,catpos].T, \
fmt=2*"%9.2f "+"%9.0f "+"%9.1f "+"%4i "+"%6.2f "+3*"%7.2f "+"%11.3e")
# find spots in flexure series, in order of increasing abs(rho), and store sextractor output
row_fLd = np.zeros((flexposns, Lines, 2))
col_fLd = np.zeros((flexposns, Lines, 2))
print "\n fits rho line spots rshift cshift rslope cslope rmserr "
print " deg Ang arcsec arcsec arcmin arcmin bins"
for dirn in (1, -1):
refpos = catpos
posdirlist = np.argsort(dirn * rho_f)
poslist = posdirlist[dirn * rho_f[posdirlist] > rho_f[refpos]]
for fpos in poslist:
col_S, row_S = sexdata_jfS[0:2, refpos, :]
sex_js = sextract(fitslist[fpos], "sexwt.fits", deblend=deblend)
binsqerr_sS = (sex_js[1, :, None] - row_S[None, :])**2 + (
sex_js[0, :, None] - col_S[None, :])**2
S_s = np.argmin(binsqerr_sS, axis=1)
# First compute image shift by averaging small errors
rowerr_s = sex_js[1] - row_S[S_s]
colerr_s = sex_js[0] - col_S[S_s]
hist_r, bin_r = np.histogram(rowerr_s,
bins=32,
range=(-2 * bintol, 2 * bintol))
drow = rowerr_s[(rowerr_s > bin_r[np.argmax(hist_r)]-bintol) & \
(rowerr_s < bin_r[np.argmax(hist_r)]+bintol)].mean()
hist_c, bin_c = np.histogram(colerr_s,
bins=32,
range=(-2 * bintol, 2 * bintol))
dcol = colerr_s[(colerr_s > bin_c[np.argmax(hist_c)]-bintol) & \
(colerr_s < bin_c[np.argmax(hist_c)]+bintol)].mean()
# Now refind the closest ID
binsqerr_sS = (sex_js[1,:,None] - row_S[None,:] -drow)**2 + \
(sex_js[0,:,None] - col_S[None,:] -dcol)**2
binsqerr_s = binsqerr_sS.min(axis=1)
isfound_s = binsqerr_s < bintol**2
S_s = np.argmin(binsqerr_sS, axis=1)
isfound_s &= (binsqerr_s == binsqerr_sS[:, S_s].min(axis=0))
isfound_S = np.array([S in S_s[isfound_s] for S in range(Spots)])
sexdata_jfS[:, fpos, S_s[isfound_s]] = sex_js[:, isfound_s]
drow_S = sexdata_jfS[1, fpos] - sexdata_jfS[1, catpos]
dcol_S = sexdata_jfS[0, fpos] - sexdata_jfS[0, catpos]
# np.savetxt("motion_"+str(fpos)+".txt",np.vstack((isfound_S,l_S,drow_S,dcol_S)).T,fmt="%3i %3i %8.2f %8.2f")
# Compute flexure image motion parameters for each line
for L in range(Lines):
ok_S = ((l_S == l_L[L]) & isfound_S)
row_fLd[fpos,L],rowchi,d,d,d = \
np.polyfit(sexdata_jfS[0,catpos,ok_S]-xcenter_L[L],drow_S[ok_S],deg=1,full=True)
col_fLd[fpos,L],colchi,d,d,d = \
np.polyfit(sexdata_jfS[1,catpos,ok_S]-ycenter,dcol_S[ok_S],deg=1,full=True)
rms = np.sqrt((rowchi + colchi) / (2 * ok_S.sum()))
print ("%12s %5.0f %5i %5i "+5*"%7.2f ") % (image_f[fpos], rho_f[fpos], wav_L[L], \
ok_S.sum(),row_fLd[fpos,L,1]*rbin*pix_scale, col_fLd[fpos,L,1]*cbin*pix_scale, \
60.*np.degrees(row_fLd[fpos,L,0]),-60.*np.degrees(col_fLd[fpos,L,0]), rms)
if option == "filesave":
np.savetxt(prefix+"flex_"+str(fpos)+".txt",np.vstack((isfound_S,drow_S,dcol_S)).T, \
fmt = "%2i %8.3f %8.3f")
np.savetxt(prefix + "sextr_" + str(fpos) + ".txt",
sexdata_jfS[:, fpos].T)
print
# make plots
fig, plot_s = plt.subplots(2, 1, sharex=True)
plt.xlabel('Rho (deg)')
plt.xlim(-120, 120)
plt.xticks(range(-120, 120, 30))
fig.set_size_inches((8.5, 11))
fig.subplots_adjust(left=0.175)
plot_s[0].set_title(
str(dateobs) + [" Imaging", " Spectral"][isspec] + " Flexure")
plot_s[0].set_ylabel('Mean Position (arcsec)')
plot_s[0].set_ylim(-0.5, 4.)
plot_s[1].set_ylabel('Rotation (arcmin ccw)')
plot_s[1].set_ylim(-10., 6.)
lbl_L = [("%5.0f") % (wav_L[L]) for L in range(Lines)]
color_L = 'bgrcmykw'
for L in range(Lines):
plot_s[0].plot(rho_f,row_fLd[:,L,1]*rbin*pix_scale, \
color=color_L[L],marker='D',markersize=8,label='row '+lbl_L[L])
plot_s[1].plot(rho_f,
60. * np.degrees(row_fLd[:, L, 0]),
color=color_L[L],
marker='D',
markersize=8,
label='row ' + lbl_L[L])
collbl = 'col' + lbl_L[0]
for L in range(Lines):
plot_s[0].plot(rho_f,col_fLd[:,L,1]*cbin*pix_scale, \
color=color_L[L],marker='s',markersize=8,label=collbl)
plot_s[1].plot(rho_f,-60.*np.degrees(col_fLd[:,L,0]), \
color=color_L[L],marker='s',markersize=8,label=collbl)
collbl = ''
plot_s[0].legend(fontsize='medium', loc='upper center')
plotfile = str(dateobs) + ['_imflex.pdf', '_grflex.pdf'][isspec]
plt.savefig(plotfile, orientation='portrait')
if os.name == 'posix':
if os.popen('ps -C evince -f').read().count(plotfile) == 0:
os.system('evince ' + plotfile + ' &')
os.remove("out.txt")
os.remove("qred_thrufoc.param")
os.remove("sexwt.fits")
return
def flexure_rss(fitslist,option=""):
global sex_js, rd, B, pixarcsec, gridsize, niter
global rho_f, rc0_dgg, usespots_gg, fwhminterp_g # for cube analysis
if option == "filesave": prefix = raw_input("file prefix: ")
pixel = 15. # pixel size in microns
pix_scale=0.125
sexparams = ["X_IMAGE","Y_IMAGE","FLUX_ISO","FLUX_MAX","FLAGS","CLASS_STAR", \
"X2WIN_IMAGE","Y2WIN_IMAGE","XYWIN_IMAGE","ERRX2WIN_IMAGE"]
np.savetxt("qred_thrufoc.param",sexparams,fmt="%s")
fmaxcol,flagcol,xvarcol,yvarcol,xerrcol = (3,4,6,7,9) # column nos (from 0) of data in sextractor
imagestooclosefactor = 3.0 # too close if factor*sep < sqrt(var)
gaptooclose = 1.25 # arcsec
edgetooclose = 1.25 # arcsec
rattolerance = 0.25
toofaint = 250. # FMAX counts
galaxydelta = 0.4 # arcsec
MOSimagelimit = 1. # arcsec
deblend = .005 # default
flexposns = len(fitslist)
obsdict=obslog(fitslist)
image_f = [fitslist[fpos].split(".")[0][-12:] for fpos in range(flexposns)]
dateobs = int(image_f[0][:8])
if dateobs > 20110928: rho_f = np.array(obsdict["TRKRHO"]).astype(float)
else: rho_f = np.array(obsdict["TELRHO"]).astype(float)
catpos = np.argmin(np.abs(rho_f))
cbin,rbin = np.array(obsdict["CCDSUM"][catpos].split(" ")).astype(int)
maskid = obsdict["MASKID"][catpos].strip()
filter = obsdict["FILTER"][catpos].strip()
grating = obsdict["GRATING"][catpos].strip()
rows,cols = pyfits.getdata(fitslist[catpos]).shape
isspec = (obsdict["GR-STATE"][catpos][1] =="4")
print str(datetime.now()), "\n"
print "Mask: ", maskid
print "Filter: ", filter
print "Grating: ", grating
# make catalog of stars using image closest to rho=0 (capital S)
sex_js = sextract(fitslist[catpos],deblend=deblend)
fluxisomedian = np.median(np.sort(sex_js[2])[-10:]) # median of 10 brightest
ok_s = sex_js[2] > fluxisomedian/100. # get rid of bogus stars
sexcols = sex_js.shape[0]
Stars = ok_s.sum()
sexdata_jfS = np.zeros((sexcols,flexposns,Stars))
sexdata_jfS[:,catpos] = sex_js[:,ok_s]
xcenter = 0.5*(sexdata_jfS[0,catpos].min() + sexdata_jfS[0,catpos].max())
ycenter = 0.5*(sexdata_jfS[1,catpos].min() + sexdata_jfS[1,catpos].max())
print "\n fits rho stars rshift cshift rslope cslope rmserr "
print " deg arcsec arcsec arcmin arcmin bins"
print ("%12s %5.1f %5i "+5*"%7.2f ") % \
(image_f[catpos], rho_f[catpos], Stars, 0., 0., 0., 0., 0.)
if option == "filesave":
np.savetxt(prefix+"Stars.txt",sexdata_jfS[:,catpos].T, \
fmt=2*"%9.2f "+"%9.0f "+"%9.1f "+"%4i "+"%6.2f "+3*"%7.2f "+"%11.3e")
# find stars in flexure series, in order of increasing abs(rho), and store sextractor output
row_fd = np.zeros((flexposns,2))
col_fd = np.zeros((flexposns,2))
for dirn in (1,-1):
refpos = catpos
posdirlist = np.argsort(dirn*rho_f)
poslist = posdirlist[dirn*rho_f[posdirlist] > rho_f[refpos]]
for fpos in poslist:
col_S,row_S = sexdata_jfS[0:2,refpos,:]
sex_js = sextract(fitslist[fpos],"sexwt.fits",deblend=deblend)
bintol = 16/cbin # 2 arcsec tolerance for finding star
binsqerr_sS = (sex_js[1,:,None] - row_S[None,:])**2 + (sex_js[0,:,None] - col_S[None,:])**2
S_s = np.argmin(binsqerr_sS,axis=1)
# First compute image shift by averaging small errors
rowerr_s = sex_js[1] - row_S[S_s]
colerr_s = sex_js[0] - col_S[S_s]
hist_r,bin_r = np.histogram(rowerr_s,bins=32,range=(-2*bintol,2*bintol))
drow = rowerr_s[(rowerr_s > bin_r[np.argmax(hist_r)]-bintol) & \
(rowerr_s < bin_r[np.argmax(hist_r)]+bintol)].mean()
hist_c,bin_c = np.histogram(colerr_s,bins=32,range=(-2*bintol,2*bintol))
dcol = colerr_s[(colerr_s > bin_r[np.argmax(hist_r)]-bintol) & \
(colerr_s < bin_r[np.argmax(hist_r)]+bintol)].mean()
# Now refind the closest ID
binsqerr_sS = (sex_js[1,:,None] - row_S[None,:] -drow)**2 + \
(sex_js[0,:,None] - col_S[None,:] -dcol)**2
binsqerr_s = binsqerr_sS.min(axis=1)
isfound_s = binsqerr_s < bintol**2
S_s = np.argmin(binsqerr_sS,axis=1)
isfound_s &= (binsqerr_s == binsqerr_sS[:,S_s].min(axis=0))
isfound_S = np.array([S in S_s[isfound_s] for S in range(Stars)])
sexdata_jfS[:,fpos,S_s[isfound_s]] = sex_js[:,isfound_s]
drow_S = sexdata_jfS[1,fpos]-sexdata_jfS[1,catpos]
dcol_S = sexdata_jfS[0,fpos]-sexdata_jfS[0,catpos]
row_fd[fpos],rowchi,d,d,d = np.polyfit(sexdata_jfS[0,catpos,isfound_S]-xcenter, \
drow_S[isfound_S],deg=1,full=True)
col_fd[fpos],colchi,d,d,d = np.polyfit(sexdata_jfS[1,catpos,isfound_S]-ycenter, \
dcol_S[isfound_S],deg=1,full=True)
rms = np.sqrt((rowchi+colchi)/(2*isfound_S.sum()))
print ("%12s %5.0f %5i "+5*"%7.2f ") % (image_f[fpos], rho_f[fpos], isfound_S.sum(), \
row_fd[fpos,1]*rbin*pix_scale, col_fd[fpos,1]*cbin*pix_scale, \
60.*np.degrees(row_fd[fpos,0]),-60.*np.degrees(col_fd[fpos,0]), rms)
if option == "filesave":
np.savetxt(prefix+"flex_"+str(fpos)+".txt",np.vstack((isfound_S,drow_S,dcol_S)).T, \
fmt = "%2i %8.3f %8.3f")
np.savetxt(prefix+"sextr_"+str(fpos)+".txt",sexdata_jfS[:,fpos].T)
# make plots
fig,plot_s = plt.subplots(2,1,sharex=True)
plt.xlabel('Rho (deg)')
plt.xlim(-120,120)
plt.xticks(range(-120,120,30))
fig.set_size_inches((8.5,11))
fig.subplots_adjust(left=0.175)
plot_s[0].set_title(str(dateobs)+" Imaging Flexure")
plot_s[0].set_ylabel('Mean Position (arcsec)')
plot_s[0].set_ylim(-0.5,2.)
plot_s[1].set_ylabel('Rotation (arcmin ccw)')
plot_s[1].set_ylim(-6.,6.)
plot_s[0].plot(rho_f,row_fd[:,1]*rbin*pix_scale,marker='D',label='row')
plot_s[0].plot(rho_f,col_fd[:,1]*rbin*pix_scale,marker='D',label='col')
plot_s[1].plot(rho_f,60.*np.degrees(row_fd[:,0]),marker='D',label='row')
plot_s[1].plot(rho_f,-60.*np.degrees(col_fd[:,0]),marker='D',label='col')
plot_s[0].legend(fontsize='medium',loc='upper center')
plotfile = str(dateobs)+'_imflex.pdf'
plt.savefig(plotfile,orientation='portrait')
if os.name=='posix':
if os.popen('ps -C evince -f').read().count(plotfile)==0: os.system('evince '+plotfile+' &')
os.remove("out.txt")
os.remove("qred_thrufoc.param")
os.remove("sexwt.fits")
return
