def run():
xall,yall = np.loadtxt(gpr.get_com_file(0),skiprows=1,usecols=(0,1),unpack=True) # 2*[rcore]
# calculate 2D radius on the skyplane
r = np.sqrt(xall**2+yall**2) #[rcore]
# set number and size of (linearly spaced) bins
rmin = 0. #[rcore]
rmax = max(r) if gpr.rprior<0 else 1.0*gpr.rprior #[rcore]
print 'rmax [rcore] = ', rmax
r = r[(r<rmax)]
# determine radius once and for all
if gp.lograd:
print gpr.nbins,' bins in log spacings'
binmin, binmax, rbin = bin_r_log(rmax/gpr.nbins, rmax, gpr.nbins)
elif gp.consttr:
print len(r)/gpr.nbins,' particles per bin'
binmin, binmax, rbin = bin_r_const_tracers(r, len(r)/gpr.nbins)
else:
print gpr.nbins, ' bins in linear spacings'
binmin, binmax, rbin = bin_r_linear(rmin, rmax, gpr.nbins)
# volume of a circular ring from binmin to binmax
vol = np.zeros(gpr.nbins)
for k in range(gpr.nbins):
vol[k] = np.pi*(binmax[k]**2-binmin[k]**2) # [rcore^2]
for comp in range(gpr.ncomp):
print 'comp = ',comp
print 'input: ',gpr.get_com_file(comp)
# start from data centered on COM already:
if gfile.bufcount(gpr.get_com_file(comp))<2: continue
x,y,vlos = np.loadtxt(gpr.get_com_file(comp),\
skiprows=1,usecols=(0,1,2),unpack=True) #[rcore], [rcore], [km/s]
# calculate 2D radius on the skyplane
r = np.sqrt(x**2+y**2) #[rcore]
# set maximum radius (if gpr.rprior is set)
rmax = max(r) if gpr.rprior<0 else 1.0*gpr.rprior #[rcore]
print 'rmax [rcore] = ', rmax
sel = (r<=rmax)
x = x[sel]; y = y[sel]; vlos = vlos[sel]; r = r[sel] #[rcore]
totmass = 1.*len(x) #[munit], munit = 1/star
rs = r
# no offset from the start!
# rs = gpr.rerror*np.random.randn(len(r))+r #[rcore]
# vlos = gpr.vrerror*np.random.randn(len(vlos))+vlos #[km/s]
print 'output: ',gpr.get_siglos_file(comp)
vfil = open(gpr.get_siglos_file(comp),'w')
print >> vfil,'r','sigma_r(r)','error'
# 30 iterations for drawing a given radius in bin
dispvelocity = np.zeros((gpr.nbins,gpr.n))
a = np.zeros((gpr.nbins,gpr.n))
p_dvlos = np.zeros(gpr.nbins)
p_edvlos = np.zeros(gpr.nbins)
for k in range(gpr.n):
rsi = gpr.rerror*np.random.randn(len(rs))+rs #[rcore]
vlosi = gpr.vrerror*np.random.randn(len(vlos))+vlos #[km/s]
for i in range(gpr.nbins):
ind1 = np.argwhere(np.logical_and(rsi>binmin[i],rsi<binmax[i])).flatten()
a[i][k] = len(ind1) #[1]
vlos1 = vlosi[ind1] #[km/s]
if(len(ind1)<=1):
dispvelocity[i][k] = dispvelocity[i-1][k]
# attention! should be 0, uses last value
else:
dispvelocity[i][k] = meanbiweight(vlos1,ci_perc=68.4,\
ci_mean=True,ci_std=True)[1]
# [km/s], see BiWeight.py
for i in range(gpr.nbins):
dispvel = np.sum(dispvelocity[i])/gpr.n #[km/s]
ab = np.sum(a[i])/(1.*gpr.n) #[1]
if ab == 0:
dispvelerror = p_edvlos[i-1] #[km/s]
# attention! uses last error
else:
dispvelerror = dispvel/np.sqrt(ab) #[km/s]
p_dvlos[i] = dispvel #[km/s]
p_edvlos[i]= dispvelerror #[km/s]
maxvlos = max(p_dvlos) #[km/s]
print 'maxvlos = ',maxvlos,'[km/s]'
fpars = open(gpr.get_params_file(comp),'a')
print >> fpars,maxvlos #[km/s]
fpars.close()
for i in range(gpr.nbins):
# [rcore] [maxvlos] [maxvlos]
print >> vfil,rbin[i], np.abs(p_dvlos[i]/maxvlos),np.abs(p_edvlos[i]/maxvlos) #/np.sqrt(n))
vfil.close()
if not gp.testplot_read: continue
ion(); subplot(111)
print 'rbin = ',rbin,' rcore'
print 'p_dvlos = ',p_dvlos,' km/s'
print 'p_edvlos = ',p_edvlos, 'km/s'
plot(rbin,p_dvlos,'b',lw=1)
fill_between(rbin,p_dvlos-p_edvlos,p_dvlos+p_edvlos,alpha=0.5,color='r') #[rcore],[km/s],[km/s]
xlabel(r'$r [\mathrm{rcore}]$')
ylabel(r'$\langle\sigma_{\mathrm{LOS}}\rangle [\mathrm{km/s}]$')
ylim([-5,30])
# xscale('log')
xlim([np.min(rbin),np.max(rbin)])
#plt.legend(['\rho','\rho'],'lower left'); #title(dwarf)
savefig(gpr.get_siglos_png(comp))
if gpr.showplots:
ioff();show();clf()
def run():
xall,yall = np.loadtxt(gpr.get_com_file(0), skiprows=1, usecols=(0,1), unpack=True) # 2*[Rscale]
# calculate 2D radius on the skyplane
R = np.sqrt(xall**2+yall**2) # [Rscale]
# set number and size of (linearly spaced) bins
Rmin = 0. #[rscale]
Rmax = max(R) if gpr.rprior<0 else 1.0*gpr.rprior # [Rscale]
print('Rmax [Rscale] = ', Rmax)
R = R[(R<Rmax)]
# determine radius once and for all
# this must not be changed between readout and gravlite run
# if you wish to change: set gp.getnewdata = True in gl_params.py
if gp.lograd:
print(gpr.nbins,' bins in log spacings')
Binmin, Binmax, Rbin = bin_r_log(Rmax/gpr.nbins, Rmax, gpr.nbins)
elif gp.consttr:
print(len(R)/gpr.nbins,' particles per bin')
Binmin, Binmax, Rbin = bin_r_const_tracers(R, len(R)/gpr.nbins)
else:
print(gpr.nbins, ' bins in linear spacings')
Binmin, Binmax, Rbin = bin_r_linear(Rmin, Rmax, gpr.nbins)
# volume of a circular ring from binmin to binmax
Vol = np.zeros(gpr.nbins)
for k in range(gpr.nbins):
Vol[k] = np.pi*(Binmax[k]**2-Binmin[k]**2) # [Rscale^2]
for comp in range(gpr.ncomp):
print('####### working on component ',comp)
print('input: ',gpr.get_com_file(comp))
# start from data centered on COM already:
if gfile.bufcount(gpr.get_com_file(comp))<2: continue
x,y,v = np.loadtxt(gpr.get_com_file(comp),\
skiprows=1,usecols=(0,1,2),unpack=True) #[rscale], [rscale], [km/s]
# calculate 2D radius on the skyplane
R = np.sqrt(x**2+y**2) #[rscale]
# set maximum radius (if gpr.rprior is set)
Rmax = max(R) if gpr.rprior<0 else 1.0*gpr.rprior # [Rscale]
print('Rmax [Rscale] = ', Rmax)
sel = (R<=Rmax)
x = x[sel]; y = y[sel]; v = v[sel]; R = R[sel] # [Rscale]
totmass = 1.*len(x) # [munit], munit = 1/star
Rs = R # + possible starting offset, [Rscale]
vlos = v # + possible starting offset, [km/s]
print('output density: ')
print(gpr.get_ntracer_file(comp))
tr = open(gpr.get_ntracer_file(comp),'w')
print(totmass, file=tr)
tr.close()
print(gpr.get_dens_file(comp))
de = open(gpr.get_dens_file(comp),'w')
print('Rbin [Rscale]','Binmin [Rscale]','Binmax [Rscale]','Nu(R)/Nu(0) [1]','error [1]', file=de)
print(gpr.get_enc_mass_file(comp))
em = open(gpr.get_enc_mass_file(comp),'w')
print('R [Rscale]','Binmin [Rscale]','Binmax [Rscale]','M(<Binmax) [Msun]','error [Msun]', file=em)
print('output siglos: ',gpr.get_siglos_file(comp))
sigfil = open(gpr.get_siglos_file(comp),'w')
print('R [Rscale]','Binmin [Rscale]','Binmax [Rscale]','sigma_r(R) [km/s]','error [km/s]', file=sigfil)
print('output kurtosis: ',gpr.get_kurtosis_file(comp))
kappafil = open(gpr.get_kurtosis_file(comp),'w')
print('R [Rscale]','Binmin [Rscale]','Binmax [Rscale]','kappa_los(R) [1]','error [1]', file=kappafil)
# gpr.n=30 iterations for getting random picked radius values
Density = np.zeros((gpr.nbins,gpr.n))
dispvelocity = np.zeros((gpr.nbins,gpr.n))
mom4 = np.zeros((gpr.nbins,gpr.n))
a = np.zeros((gpr.nbins,gpr.n)) # shared by density, siglos, kappa calcs
for k in range(gpr.n):
Rsi = gpr.Rerror * np.random.randn(len(Rs)) + Rs # [Rscale]
vlosi = gpr.vrerror * np.random.randn(len(vlos)) + vlos # [km/s]
for i in range(gpr.nbins):
ind1 = np.argwhere(np.logical_and(Rsi >= Binmin[i],Rsi<Binmax[i])).flatten() # [1]
Density[i][k] = (1.*len(ind1))/Vol[i]*totmass # [munit/rscale**2]
vlos1 = vlosi[ind1] # [km/s]
if(len(ind1)<=1):
dispvelocity[i][k] = dispvelocity[i-1][k]
mom4[i][k] = mom4[i-1][k]
# attention! should be 0, uses last value
else:
dispvelocity[i][k] = meanbiweight(vlos1,ci_perc=68.4,ci_mean=True,ci_std=True)[1]
# [km/s], see BiWeight.py
mom4[i][k] = kurtosis(vlos1, axis=0, fisher=False, bias=False) # [1]
a[i][k] = 1.*len(ind1) #[1]
# output density
Dens0 = np.sum(Density[0])/(1.*gpr.n) # [munit/Rscale^2]
print('Dens0 = ', Dens0, '[munit/Rscale^2]')
crscale = open(gpr.get_params_file(comp),'r')
Rscale = np.loadtxt(crscale, comments='#', skiprows=1, unpack=False)
crscale.close()
cdens = open(gpr.get_params_file(comp),'a')
print(Dens0, file=cdens) # [munit/Rscale^2]
Dens0pc = Dens0/Rscale**2 # [munis/pc^2]
print(Dens0pc, file=cdens) # [munit/pc^2]
print(totmass, file=cdens) # [munit]
cdens.close()
ab0 = np.sum(a[0])/(1.*gpr.n) # [1]
Denserr0 = Dens0/np.sqrt(ab0) # [munit/Rscale^2]
P_dens = np.zeros(gpr.nbins); P_edens = np.zeros(gpr.nbins)
for b in range(gpr.nbins):
Dens = np.sum(Density[b])/(1.*gpr.n) # [munit/Rscale^2]
ab = np.sum(a[b])/(1.*gpr.n) # [1]
Denserr = Dens/np.sqrt(ab) # [munit/Rscale^2]
# TODO: too small? offset in nu?
Denserror = np.sqrt((Denserr/Dens0)**2+(Dens*Denserr0/(Dens0**2))**2) # [1]
if(math.isnan(Denserror)):
Denserror = 0. # [1]
P_dens[b] = P_dens[b-1] # [1]
P_edens[b]= P_edens[b-1] # [1]
else:
P_dens[b] = Dens/Dens0 # [1]
P_edens[b]= Denserror # [1] #100/rbin would be artificial guess
print(Rbin[b], Binmin[b], Binmax[b], P_dens[b], P_edens[b], file=de)
# 3*[rscale], [dens0], [dens0]
indr = (R<Binmax[b])
Menclosed = 1.0*np.sum(indr)/totmass # for normalization to 1 #[totmass]
Merror = Menclosed/np.sqrt(ab) # or artificial Menclosed/10 #[totmass]
print(Rbin[b], Binmin[b], Binmax[b], Menclosed, Merror, file=em) # [Rscale], 2* [totmass]
# TODO: check: take rbinmax for MCMC?
de.close()
em.close()
# output siglos
p_dvlos = np.zeros(gpr.nbins); p_edvlos = np.zeros(gpr.nbins)
for b in range(gpr.nbins):
dispvel = np.sum(dispvelocity[b])/gpr.n #[km/s]
ab = np.sum(a[b])/(1.*gpr.n) #[1]
if ab == 0:
dispvelerror = p_edvlos[b-1] #[km/s]
# attention! uses last error
else:
dispvelerror = dispvel/np.sqrt(ab) #[km/s]
p_dvlos[b] = dispvel #[km/s]
p_edvlos[b]= dispvelerror #[km/s]
maxvlos = max(p_dvlos) #[km/s]
print('maxvlos = ', maxvlos, '[km/s]')
fpars = open(gpr.get_params_file(comp),'a')
print(maxvlos, file=fpars) #[km/s]
fpars.close()
for b in range(gpr.nbins):
print(Rbin[b], Binmin[b], Binmax[b], np.abs(p_dvlos[b]/maxvlos),np.abs(p_edvlos[b]/maxvlos), file=sigfil)
# 3*[rscale], 2*[maxvlos]
# TODO: check uncommented /np.sqrt(n))
sigfil.close()
# output kurtosis kappa
p_kappa = np.zeros(gpr.nbins) # needed for plotting later
p_ekappa = np.zeros(gpr.nbins)
for b in range(gpr.nbins):
kappavel = np.sum(mom4[b])/gpr.n #[1]
ab = np.sum(a[b])/(1.*gpr.n) #[1]
if ab == 0:
kappavelerror = p_edvlos[b-1] #[1]
# attention! uses last error
else:
kappavelerror = np.abs(kappavel/np.sqrt(ab)) #[1]
p_kappa[b] = kappavel
p_ekappa[b] = kappavelerror
print(Rbin[b],Binmin[b],Binmax[b], kappavel, kappavelerror, file=kappafil) # [rscale], 2*[1]
# TODO: /np.sqrt(n))
kappafil.close()
if not gpr.showplots: continue
# plot density
ion(); subplot(111)
print('Rbin = ', Rbin)
print('P_dens = ', P_dens)
print('P_edens = ', P_edens)
plot(Rbin, P_dens*Dens0pc, 'b', lw=1)
lbound = (P_dens-P_edens)*Dens0pc; lbound[lbound<1e-6] = 1e-6
ubound = (P_dens+P_edens)*Dens0pc
fill_between(Rbin, lbound, ubound, alpha=0.5, color='r')
yscale('log')
# xlim([0, gpr.rprior])
# ylim([np.min(lbound),np.max(ubound)])
xlabel(r'$R [R_c]$')
ylabel(r'$\nu_{2D}(R) [\mathrm{Msun/pc/pc}]$')
savefig(gpr.get_dens_png(i))
ioff(); show(); clf()
# plot siglos
ion(); subplot(111)
print('Rbin = ',Rbin,' Rscale')
print('p_dvlos = ',p_dvlos,' km/s')
print('p_edvlos = ',p_edvlos, 'km/s')
plot(Rbin, p_dvlos, 'b', lw=1)
fill_between(Rbin, p_dvlos-p_edvlos, p_dvlos+p_edvlos, alpha=0.5, color='r')
# [rscale],2*[km/s]
xlabel(r'$R [\mathrm{Rscale}]$')
ylabel(r'$\langle\sigma_{\mathrm{LOS}}\rangle [\mathrm{km/s}]$')
ylim([-1, 30])
# xlim([0, 3])
savefig(gpr.get_siglos_png(comp))
ioff(); show(); clf()
# plot kappa
ion(); subplot(111)
print('Rbin = ', Rbin, ' Rscale')
print('p_kappa = ', p_kappa)
print('p_ekappa = ', p_ekappa)
plot(Rbin, p_kappa, 'b', lw=1)
fill_between(Rbin, p_kappa-p_ekappa, p_kappa+p_ekappa, alpha=0.5, color='r')
# [rscale], 2*[1]
xlabel(r'$R [\mathrm{Rscale}]$')
ylabel(r'$\langle\kappa_{\mathrm{LOS}}\rangle [1]$')
ylim([0, 5.])
# xlim([0, gpr.rprior])
savefig(gpr.get_kurtosis_png(comp))
ioff(); show(); clf()
def run():
# get radius, used for all binning
print('input:')
print(gpr.get_com_file(0))
if gfile.bufcount(gpr.get_com_file(0))<2: return
x,y,vlos = np.loadtxt(gpr.get_com_file(0), skiprows=1, unpack=True) #2*[rscale], [km/s]
totmass = 1.*len(x) # [munit], [Msun], where each star is weighted with the same mass
r = np.sqrt(x*x+y*y) # [rscale]
#set binning
#gpr.nbins = (max - min)*N^(1/3)/(2*(Q3-Q1)) #(method of wand)
rmin = 0. # [rscale]
rmax = max(r) if gpr.rprior<0 else 1.0*gpr.rprior # [rscale]
if gp.lograd:
# space logarithmically in radius
binmin, binmax, rbin = bin_r_log(rmax/gpr.nbins, rmax, gpr.nbins)
elif gp.consttr:
binmin, binmax, rbin = bin_r_const_tracers(r, len(r)/gpr.nbins)
else:
binmin, binmax, rbin = bin_r_linear(rmin, rmax, gpr.nbins)
# offset from the start!
rs = gpr.rerror*np.random.randn(len(r))+r #[rscale]
vlos = gpr.vrerror*np.random.randn(len(vlos))+vlos #[km/s]
print('output: ',gpr.get_siglos_file(0))
vfil = open(gpr.get_siglos_file(0),'w')
print('r','sigma_r(r)','error', file=vfil)
# 30 iterations for drawing a given radius in bin
dispvelocity = np.zeros((gpr.nbins,gpr.n))
a = np.zeros((gpr.nbins,gpr.n))
p_dvlos = np.zeros(gpr.nbins)
p_edvlos = np.zeros(gpr.nbins)
for k in range(gpr.n):
rsi = gpr.rerror*np.random.randn(len(rs))+rs #[rscale]
vlosi = gpr.vrerror*np.random.randn(len(vlos))+vlos #[km/s]
for i in range(gpr.nbins):
ind1 = np.argwhere(np.logical_and(rsi>binmin[i],rsi<binmax[i])).flatten()
a[i][k] = len(ind1) #[1]
vlos1 = vlosi[ind1] #[km/s]
if(len(ind1)<=1):
dispvelocity[i][k] = dispvelocity[i-1][k]
# attention! should be 0, uses last value
else:
dispvelocity[i][k] = meanbiweight(vlos1,ci_perc=68.4,\
ci_mean=True,ci_std=True)[1]
# [km/s], see BiWeight.py
for i in range(gpr.nbins):
dispvel = np.sum(dispvelocity[i])/gpr.n #[km/s]
ab = np.sum(a[i])/(1.*gpr.n) #[1]
if ab == 0:
dispvelerror = p_edvlos[i-1] #[km/s]
# attention! uses last error
else:
dispvelerror = dispvel/np.sqrt(ab) #[km/s]
p_dvlos[i] = dispvel #[km/s]
p_edvlos[i]= dispvelerror #[km/s]
maxvlos = max(p_dvlos) #[km/s]
print('maxvlos = ',maxvlos,'[km/s]')
fpars = open(gpr.get_params_file(0),'a')
print(maxvlos, file=fpars) #[km/s]
fpars.close()
import shutil
shutil.copy2(gpr.get_params_file(0), gpr.get_params_file(1))
for i in range(gpr.nbins):
# [rscale] [maxvlos] [maxvlos]
print(rbin[i], np.abs(p_dvlos[i]/maxvlos),np.abs(p_edvlos[i]/maxvlos), file=vfil) #/np.sqrt(n))
vfil.close()
if not gpr.showplots: return
ion(); subplot(111)
print('rbin = ',rbin,' rscale')
print('p_dvlos = ',p_dvlos,' km/s')
print('p_edvlos = ',p_edvlos, 'km/s')
plot(rbin,p_dvlos,'b',linewidth=3)
fill_between(rbin,p_dvlos-p_edvlos,p_dvlos+p_edvlos,alpha=0.5,color='r') #[rscale],[km/s],[km/s]
xlabel(r'$r [rscale]$')
ylabel(r'$\langle\sigma_{LOS}\rangle [km/s]$')
ylim([-5,30])
# xscale('log')
xlim([np.min(rbin),np.max(rbin)])
#plt.legend(['\rho','\rho'],'lower left'); #title(dwarf)
savefig(gpr.get_siglos_png(0))
if gpr.showplots:
ioff();show();clf()