def residual(A, zeta, realFlag=False, plotFlag=False):
"""Returns Euclidean norm of subhalo and core mass functions, where only subhalos and cores in 1:1 match are considered."""
# residual plots
print len(sh['subhalo_mass'][sh_mask][fmask])
print len(cc_filtered[m_evolved_col(A, zeta)][idx[:, 0][fmask]])
# realmask = cc_filtered['fof_halo_tag'][idx[:,0][fmask]]>=0
if realFlag:
realmask = np.invert(cc_filtered['wasInFragment'][idx[:, 0][fmask]])
else:
realmask = np.full(np.sum(fmask), True, dtype=bool)
shmf = sh['subhalo_mass'][sh_mask][fmask][realmask] / sh['M'][sh_mask][
fmask][realmask]
r = (-3, 0)
r_res = (-3, -2)
sh_xarr, sh_cnts = hist(np.log10(shmf),
bins=100,
normed=True,
normBinsize=True,
normCnts=False,
normLogCnts=True,
normScalar=nHalo,
plotFlag=plotFlag,
label='subhalos',
alpha=1,
range=r)
res_mask = (r_res[0] <= sh_xarr) & (sh_xarr <= r_res[1])
cmf = cc_filtered[m_evolved_col(A, zeta)][
idx[:, 0][fmask]][realmask] / cc_filtered['M'][idx[:,
0][fmask]][realmask]
cores_xarr, cores_cnts = hist(np.log10(cmf),
bins=100,
normed=True,
plotFlag=plotFlag,
label='cores',
alpha=1,
range=r,
normScalar=nHalo,
normCnts=False,
normBinsize=True,
normLogCnts=True)
return np.linalg.norm((sh_cnts - cores_cnts)[res_mask])
def shmass(A, zeta, realFlag=False, plotFlag=False, mergedCoreTagFlag=False):
"""Returns Euclidean norm of subhalo and core mass functions, where only subhalos and cores in 1:1 match are considered."""
# residual plots
# print len(sh['subhalo_mass'][sh_mask][fmask])
# print len(cc_filtered[m_evolved_col(A, zeta)][idx[:,0][fmask]])
# realmask = cc_filtered['fof_halo_tag'][idx[:,0][fmask]]>=0
realmask = np.full(np.sum(fmask), True, dtype=bool)
if realFlag:
realmask = np.invert(cc_filtered['wasInFragment'][idx[:, 0][fmask]])
if mergedCoreTagFlag:
realmask = realmask & (cc_filtered['mergedCoreTag'][idx[:, 0][fmask]]
== 0)
shmf = sh['subhalo_mass'][sh_mask][fmask][realmask]
r = (9, 15)
sh_xarr, sh_cnts = hist(np.log10(shmf),
bins=100,
normed=True,
normBinsize=False,
normCnts=True,
normLogCnts=True,
normScalar=1,
plotFlag=plotFlag,
label='subhalos',
alpha=1,
range=r)
cmf = cc_filtered[m_evolved_col(A, zeta)][idx[:, 0][fmask]][realmask]
cores_xarr, cores_cnts = hist(np.log10(cmf),
bins=100,
normed=True,
plotFlag=plotFlag,
label='cores',
alpha=1,
range=r,
normScalar=1,
normCnts=True,
normBinsize=False,
normLogCnts=True)
def mostMassiveCore(A, zeta, plotFlag):
iarr, carr, missing = [], [], []
# cnt = 0
for i in tqdm(range(len(distance_upper_bound))):
qres = cores_tree.query_ball_point(sh_arr[i],
r=distance_upper_bound[i])
# qres2 = []
# for qidx in qres:
# if SHMLM.dist(sh['X'][sh_mask][i], sh['Y'][sh_mask][i], sh['Z'][sh_mask][i], cc_filtered['x'][qidx], cc_filtered['y'][qidx], cc_filtered['z'][qidx]) != 0:
# qres2.append(qidx)
# else:
# cnt += 1
# qres = qres2
if len(qres) > 0:
idxmax = qres[np.argmax(cc_filtered[m_evolved_col(A, zeta)][qres])]
iarr.append(i)
carr.append(idxmax)
else:
missing.append(i)
percentexists = len(iarr) / np.sum(sh_mask) * 100
print '{}% of masked subhalos have at least 1 core within their search radius.'.format(
percentexists)
shmf = sh['subhalo_mass'][sh_mask][iarr]
r = (9, 15)
sh_xarr, sh_cnts = hist(np.log10(shmf),
bins=100,
normed=True,
normBinsize=True,
normCnts=False,
normLogCnts=True,
normScalar=nHalo,
plotFlag=plotFlag,
label='subhalos',
alpha=1,
range=r)
cmf = cc_filtered[m_evolved_col(A, zeta)][carr]
cores_xarr, cores_cnts = hist(np.log10(cmf),
bins=100,
normed=True,
plotFlag=plotFlag,
label='cores',
alpha=1,
range=r,
normScalar=nHalo,
normCnts=False,
normBinsize=True,
normLogCnts=True)
# print "Cnt: "+str(cnt)
plt.figure()
plt.hist((sh['subhalo_mass'][sh_mask][iarr] -
cc_filtered[m_evolved_col(A, zeta)][carr]) /
sh['subhalo_mass'][sh_mask][iarr],
range=(-5, 5),
bins=100) #, alpha=.5)
plt.axvline(x=0, c='k')
plt.figure()
shmassmask = sh['subhalo_mass'][sh_mask][iarr] >= (SHMLM.PARTICLES100MASS *
10)
plt.hist((sh['subhalo_mass'][sh_mask][iarr][shmassmask] -
cc_filtered[m_evolved_col(A, zeta)][carr][shmassmask]) /
sh['subhalo_mass'][sh_mask][iarr][shmassmask],
range=(-5, 5),
bins=100) #, alpha=.5)
plt.axvline(x=0, c='k')
def subhalo_core_mass_plots(A,
zeta,
M1,
M2,
plotFlag,
residualRange,
normLogCntsFlag=True,
logmOverMFlag=False,
phaseMergedFlag=False,
normCnts=False,
MdepPlots=False,
vlinepart=True,
normScalar=False,
normBinsize=False,
shPlot=True,
part100cutFlag=False,
bins=50,
ratio=False,
plotHMFlag=True):
"""Plots subhalo and core mass histogram for all satellite subhalos and cores with at least 100 particles.
Returns their residual (L2 norm) in `residualRange`."""
if MdepPlots:
fig, ((ax1, ax2), (ax3,
ax4)) = plt.subplots(2,
2,
sharex=True,
sharey=True,
gridspec_kw={
'hspace': 0,
'wspace': 0
},
figsize=[4.8 * 2, 4.8 * 1.5],
dpi=150)
# fig.suptitle('z={}'.format(z))
for ax, lM in zip([ax1, ax2, ax3, ax4], Mlist):
M1, M2 = 10**lM, 10**(lM + massbinsize)
# ax.set_title( '{} $\leq$ log(M/$h^{{-1}}M_\odot$)$\leq$ {}'.format(np.log10(M1),np.log10(M2)) )
nH_core_SV, cc_filtered_SV = cc_filtered_dict_SV[lM]
nH_core_HM, cc_filtered_HM = cc_filtered_dict_HM[lM]
if part100cutFlag:
cores100mask_SV = cc_filtered_SV[m_evolved_col(
A, zeta)] >= SHMLM_SV.PARTICLECUTMASS
cores100mask_HM = cc_filtered_HM[m_evolved_col(
A, zeta)] >= SHMLM_HM.PARTICLECUTMASS
else:
cores100mask_SV = np.full_like(
cc_filtered_SV[m_evolved_col(A, zeta)] >= 0, True)
cores100mask_HM = np.full_like(
cc_filtered_HM[m_evolved_col(A, zeta)] >= 0, True)
nH_sh_param, nH_cHM_param, nH_cSV_param = 1, 1, 1
if logmOverMFlag:
r = (-3, 0)
if part100cutFlag:
shmf, nH_sh = shmfdict_cut100[lM]
else:
shmf, nH_sh = shmfdict[lM]
if normScalar:
nH_sh_param = nH_core_SV #nH_sh
nH_cHM_param = nH_core_HM
nH_cSV_param = nH_core_SV
print 'nH_sh_param: ', nH_sh_param
print 'nH_cSV_param: ', nH_cSV_param
print 'nH_cHM_param: ', nH_cHM_param
print ''
if ratio:
plotFlag = False
if shPlot:
sh_xarr, sh_cnts = hist(np.log10(shmf),
bins=bins,
normed=True,
normBinsize=normBinsize,
normCnts=normCnts,
normLogCnts=normLogCntsFlag,
normScalar=nH_sh_param,
plotFlag=plotFlag,
label='subhalos SV',
alpha=.7,
range=r,
ax=ax)
cmf_SV = cc_filtered_SV[m_evolved_col(
A, zeta
)][cores100mask_SV] / cc_filtered_SV['M'][cores100mask_SV]
cores_xarr_SV, cores_cnts_SV = hist(
np.log10(cmf_SV),
bins=bins,
normed=True,
plotFlag=plotFlag,
label='cores SV',
alpha=.7,
range=r,
normScalar=nH_cSV_param,
normCnts=normCnts,
normBinsize=normBinsize,
normLogCnts=normLogCntsFlag,
ax=ax)
if plotHMFlag:
cmf_HM = cc_filtered_HM[m_evolved_col(
A, zeta
)][cores100mask_HM] / cc_filtered_HM['M'][cores100mask_HM]
cores_xarr_HM, cores_cnts_HM = hist(
np.log10(cmf_HM),
bins=bins,
normed=True,
plotFlag=plotFlag,
label='cores HM',
alpha=1,
range=r,
normScalar=nH_cHM_param,
normCnts=normCnts,
normBinsize=normBinsize,
normLogCnts=normLogCntsFlag,
ax=ax)
if ratio:
# assert np.array_equal(sh_xarr, cores_xarr_SV) and np.array_equal(cores_xarr_SV, cores_xarr_HM)
isfin_SV = np.isfinite(10**(cores_cnts_SV - sh_cnts))
ax.axhline(1, ls='--', alpha=1, zorder=0, c='k')
ax.plot(sh_xarr[isfin_SV],
(10**(cores_cnts_SV - sh_cnts))[isfin_SV],
label='cores/subhalos (SV)')
if plotHMFlag:
isfin_HM = np.isfinite(10**(cores_cnts_HM - sh_cnts))
ax.plot(sh_xarr[isfin_HM],
(10**(cores_cnts_HM - sh_cnts))[isfin_HM],
label='cores HM')
ax.set_xlim(-3.2, 0)
ax.set_ylim(0, 3.4)
else:
ax.set_xlim(-3.2, 0)
# ax.set_ylim(-2, 1.4)
# ax.set_xlabel(r'$\log(m/M)$')
# ax.set_ylabel(r'$\log \left[ \mathrm{d}N/\mathrm{d} \log(m/M) \right]$')
else:
#r = (9, 15)
r = (np.log10(SHMLM_HM.PARTICLECUTMASS / 100.), 15)
if part100cutFlag:
shmf, nH = shfdict_cut100[lM]
else:
shmf, nH = shfdict[lM]
if normScalar:
nH_sh_param = nH_core_SV #nH_sh
nH_cHM_param = nH_core_HM
nH_cSV_param = nH_core_SV
print 'nH_sh_param: ', nH_sh_param
print 'nH_cSV_param: ', nH_cSV_param
print 'nH_cHM_param: ', nH_cHM_param
print ''
if shPlot:
sh_xarr, sh_cnts = hist(np.log10(shmf),
bins=bins,
normed=True,
normBinsize=normBinsize,
normCnts=normCnts,
normLogCnts=normLogCntsFlag,
normScalar=nH_sh_param,
plotFlag=plotFlag,
label='subhalos SV',
alpha=1,
range=r,
ax=ax)
cmf_SV = cc_filtered_SV[m_evolved_col(A,
zeta)][cores100mask_SV]
cores_xarr_SV, cores_cnts_SV = hist(
np.log10(cmf_SV),
bins=bins,
normed=True,
plotFlag=plotFlag,
label='cores SV',
alpha=1,
range=r,
normScalar=nH_cHM_param,
normCnts=normCnts,
normBinsize=normBinsize,
normLogCnts=normLogCntsFlag,
ax=ax)
if plotHMFlag:
cmf_HM = cc_filtered_HM[m_evolved_col(
A, zeta)][cores100mask_HM]
cores_xarr_HM, cores_cnts_HM = hist(
np.log10(cmf_HM),
bins=bins,
normed=True,
plotFlag=plotFlag,
label='cores HM',
alpha=1,
range=r,
normScalar=nH_cHM_param,
normCnts=normCnts,
normBinsize=normBinsize,
normLogCnts=normLogCntsFlag,
ax=ax)
ax.set_ylim(-5, 3.6)
ax.set_xlim(7.5, 14.2)
if vlinepart:
ax.axvline(x=np.log10(SHMLM_SV.PARTICLES100MASS),
c='r',
lw=0.5,
label='100 part. SV',
ymax=0.90)
ax.axvline(x=np.log10(SHMLM_HM.PARTICLES100MASS),
c='k',
lw=0.5,
label='100 part. HM')
# ax.axvline(x=np.log10(SHMLM.PARTICLES100MASS/100.*20), c='r', lw=0.5, label='20 part.')
assert np.array_equal(cores_xarr_SV, sh_xarr)
##ax.set_title( '[{},{}]'.format(np.log10(M1),np.log10(M2)) )
ax.set_title('{} $\leq$ log(M/$h^{{-1}}M_\odot$)$\leq$ {}'.format(
np.log10(M1), np.log10(M2)),
pad=-16,
x=0.60)
##ax.text(0.5, 0.7, 'sh:{}'.format(np.format_float_scientific(len(shmf))), horizontalalignment='center',verticalalignment='center', transform=ax.transAxes)
##ax.text(0.5, 0.6, 'c:{}'.format(np.format_float_scientific(len(cmf))), horizontalalignment='center',verticalalignment='center', transform=ax.transAxes)
##ax.text(0.5, 0.5, 'H:{}'.format(np.format_float_scientific(nH)), horizontalalignment='center',verticalalignment='center', transform=ax.transAxes)
if logmOverMFlag:
fig.text(0.5, 0.08, r'$\log(m/M)$', ha="center", va="center")
if ratio:
fig.text(
0.08,
0.5,
r'ratio $\left[ \mathrm{d}N/\mathrm{d} \log(m/M) \right]$',
ha="center",
va="center",
rotation=90)
else:
fig.text(
0.08,
0.5,
r'$\log \left[ \mathrm{d}N/\mathrm{d} \log(m/M) \right]$',
ha="center",
va="center",
rotation=90)
else:
fig.text(0.5, 0.08, r'$\log(m)$', ha="center", va="center")
fig.text(0.08,
0.5,
r'$\log \left[ \mathrm{d}N/\mathrm{d} \log(m) \right]$',
ha="center",
va="center",
rotation=90)
ax4.legend(loc=4)
return fig, ((ax1, ax2), (ax3, ax4))
'''
fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2)
fig.suptitle('z={}'.format(z))
for ax,lM in zip([ax1, ax2, ax3, ax4],Mlist):
M1, M2 = 10**lM, 10**(lM+massbinsize)
# ax.set_title( '{} $\leq$ log(M/$h^{{-1}}M_\odot$)$\leq$ {}'.format(np.log10(M1),np.log10(M2)) )
cc_filtered = cc_filtered_dict[lM] #generate_cores_kdtree(M1=M1, M2=M2, s1=False, disrupt=None, onlyFiltered=True)
cores100mask = cc_filtered[m_evolved_col(A, zeta)]>=SHMLM.PARTICLECUTMASS
if logmOverMFlag:
normScalarVal = 1
r = (-3, 0)
# shmf = sh['subhalo_mass'][sh_mask][sh100mask] / sh['M'][sh_mask][sh100mask]
shmf, nH = shmfdict[lM]
sh_xarr, sh_cnts = hist(np.log10(shmf), bins=100, normed=True, normBinsize=normBinsize, normCnts=normCnts, normLogCnts=normLogCntsFlag, normScalar=normScalarVal, plotFlag=plotFlag, label='subhalos', alpha=1, range=r, ax=ax)
cmf = cc_filtered[m_evolved_col(A, zeta)][cores100mask] / cc_filtered['M'][cores100mask]
cores_xarr, cores_cnts = hist(np.log10(cmf), bins=100, normed=True, plotFlag=plotFlag, label='cores', alpha=1, range=r, normScalar=normScalarVal, normCnts=normCnts, normBinsize=normBinsize, normLogCnts=normLogCntsFlag, ax=ax)
ax.set_xlim(-3.2, 0)
#ax.set_ylim(-0.8, 3)
else:
r = (9, 15)
# shmf = sh['subhalo_mass'][sh_mask][sh100mask] / sh['M'][sh_mask][sh100mask]
shmf, nH = shfdict[lM]
normScalarVal = 1
if normScalar:
normScalarVal = nH
sh_xarr, sh_cnts = hist(np.log10(shmf), bins=100, normed=True, normBinsize=normBinsize, normCnts=normCnts, normLogCnts=normLogCntsFlag, normScalar=normScalarVal, plotFlag=plotFlag, label='subhalos', alpha=1, range=r, ax=ax)
cmf = cc_filtered[m_evolved_col(A, zeta)][cores100mask]
cores_xarr, cores_cnts = hist(np.log10(cmf), bins=100, normed=True, plotFlag=plotFlag, label='cores', alpha=1, range=r, normScalar=normScalarVal, normCnts=normCnts, normBinsize=normBinsize, normLogCnts=normLogCntsFlag, ax=ax)
ax.set_xlim(8.8, 15)
ax.set_ylim(-0.8, 6)
if vlinepart:
ax.axvline(x=np.log10(SHMLM.PARTICLES100MASS), c='k', lw=0.5, label='100 part.')
ax.axvline(x=np.log10(SHMLM.PARTICLES100MASS/100.*20), c='r', lw=0.5, label='20 part.')
assert np.array_equal(cores_xarr, sh_xarr)
##ax.set_title( '[{},{}]'.format(np.log10(M1),np.log10(M2)) )
ax.set_title( '{} $\leq$ log(M/$h^{{-1}}M_\odot$)$\leq$ {}'.format(np.log10(M1),np.log10(M2)) )
##ax.text(0.5, 0.7, 'sh:{}'.format(np.format_float_scientific(len(shmf))), horizontalalignment='center',verticalalignment='center', transform=ax.transAxes)
##ax.text(0.5, 0.6, 'c:{}'.format(np.format_float_scientific(len(cmf))), horizontalalignment='center',verticalalignment='center', transform=ax.transAxes)
##ax.text(0.5, 0.5, 'H:{}'.format(np.format_float_scientific(nH)), horizontalalignment='center',verticalalignment='center', transform=ax.transAxes)
if logmOverMFlag:
fig.text(0.5,0.04, r'$\log(m/M)$', ha="center", va="center")
fig.text(0.05,0.5, r'$\log \left[ \mathrm{d}N/\mathrm{d} \log(m/M) \right]$', ha="center", va="center", rotation=90)
else:
fig.text(0.5,0.04, r'$\log(m)$', ha="center", va="center")
fig.text(0.05,0.5, r'$\log \left[ \mathrm{d}N/\mathrm{d} \log(m) \right]$', ha="center", va="center", rotation=90)
ax1.legend()
plt.subplots_adjust(hspace = 0.4)
return fig, ((ax1, ax2), (ax3, ax4))
'''
else:
# sh_mask = (sh['subhalo_tag']!=0)&(M1<=sh['M'])&(sh['M']<=M2)#&(distance_mask) #&(sh['subhalo_mass']>=SHMLM.SUBHALOMASSCUT)
# sh100mask = sh['subhalo_mass'][sh_mask]>=SHMLM.PARTICLECUTMASS
# cores_tree, cc_filtered, nHalo = generate_cores_kdtree(M1, M2, s1=False, disrupt=None, onlyFiltered=False, giveFragmentsFofMass=False)
global sh_mask, sh100mask, cores_tree, cc_filtered, nHalo
cores100mask = cc_filtered[m_evolved_col(
A, zeta)] >= SHMLM.PARTICLECUTMASS
if phaseMergedFlag:
cores100mask = cores100mask & (cc_filtered['phaseSpaceMerged'] !=
1)
if logmOverMFlag:
if normScalar:
normScalarVal = nHalo
else:
normScalarVal = 1
r = (-3, 0)
shmf = sh['subhalo_mass'][sh_mask][sh100mask] / sh['M'][sh_mask][
sh100mask]
sh_xarr, sh_cnts = hist(np.log10(shmf),
bins=100,
normed=True,
normBinsize=normBinsize,
normCnts=normCnts,
normLogCnts=normLogCntsFlag,
normScalar=normScalarVal,
plotFlag=plotFlag,
label='subhalos',
alpha=1,
range=r)
cmf = cc_filtered[m_evolved_col(
A, zeta)][cores100mask] / cc_filtered['M'][cores100mask]
cores_xarr, cores_cnts = hist(np.log10(cmf),
bins=100,
normed=True,
plotFlag=plotFlag,
label='cores',
alpha=1,
range=r,
normScalar=normScalarVal,
normCnts=normCnts,
normBinsize=normBinsize,
normLogCnts=normLogCntsFlag)
else:
r = (9, 15)
shmf = sh['subhalo_mass'][sh_mask][sh100mask]
sh_xarr, sh_cnts = hist(np.log10(shmf),
bins=100,
normed=True,
normBinsize=False,
normCnts=normCnts,
normLogCnts=normLogCntsFlag,
normScalar=1,
plotFlag=plotFlag,
label='subhalos',
alpha=1,
range=r)
cmf = cc_filtered[m_evolved_col(A, zeta)][cores100mask]
cores_xarr, cores_cnts = hist(np.log10(cmf),
bins=100,
normed=True,
plotFlag=plotFlag,
label='cores',
alpha=1,
range=r,
normScalar=1,
normCnts=normCnts,
normBinsize=False,
normLogCnts=normLogCntsFlag)
assert np.array_equal(cores_xarr, sh_xarr)
# print sh_cnts, cores_cnts
res_mask = (residualRange[0] <= sh_xarr) & (sh_xarr <=
residualRange[1])
return np.linalg.norm(
np.true_divide(((sh_cnts - cores_cnts)[res_mask]),
sh_cnts[res_mask]))