-
Notifications
You must be signed in to change notification settings - Fork 2
/
propagator_catalog_calc.py
136 lines (110 loc) · 4.66 KB
/
propagator_catalog_calc.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
import numpy as np
import time
from nbodykit.lab import *
from argparse import ArgumentParser
import pandas as pd
ap = ArgumentParser(description='PropagatorCalc')
ap.add_argument('--data',type=str)
ap.add_argument('--ran',type=str)
ap.add_argument('--out',type=str)
#ap.add_argument('--pk',type=str)
ap.add_argument('--Gf', type=float,default=0.61207)
ap.add_argument('--b', type=float, default=1.4)
ap.add_argument('--Nmu',type=int,default=10)
ap.add_argument('--Nmesh',type=int,default=512)
ns = ap.parse_args()
Nmu = ns.Nmu
Nmesh=ns.Nmesh
N=Nmesh
path = '/global/project/projectdirs/desi/users/UNIT-BAO-RSD-challenge/Reconstruction/Stage1/'
initname = '/global/cscratch1/sd/yuyu22//unitic/den%04d.bin'%Nmesh
box_size=1000
t0=time.time()
print(f"Loading initial density field {initname}", flush=True)
init_dat = np.fromfile(initname,dtype=np.float32,sep='')
print("Done")
arr_init = init_dat.reshape((N,N,N),order='F')
print("Generating mesh")
mesh_init = ArrayMesh(arr_init,BoxSize=box_size)
print("Done")
#datfile = path+ns.data
#ranfile = path+ns.ran
datfile = ns.data
ranfile = ns.ran
#May have to add a 'z_rsd' column to names depending on how teams format their files.
names = ['x','y','z']
print(f"Loading {datfile}", flush=True)
try:
catp = CSVCatalog(datfile,names=names)
except:
catdf = np.genfromtxt(datfile, names=names, usecols=(0, 1, 2))
print("Loaded with numpy successfully")
catp = ArrayCatalog(catdf)
print(f"Loading {ranfile}", flush=True)
try:
catran = CSVCatalog(ranfile,names=names)
except:
randf = np.genfromtxt(ranfile, names=names, usecols=(0, 1, 2))
print("Loaded with numpy successfully")
catran = ArrayCatalog(randf)
print("Done")
xp = catp['x'].compute()
yp = catp['y'].compute()
zp = catp['z'].compute()
#replace zp with z_rsd when relevant
positionp = np.stack([xp,yp,zp],axis=1)
n_elem = xp.shape[0]
xran = catran['x'].compute()
yran = catran['y'].compute()
zran = catran['z'].compute()
positionran = np.stack([xran,yran,zran],axis=1)
n_rand = xran.shape[0]
catp['Position'] = positionp
catran['Position'] = positionran
compensated=True
interlaced=True
meshp = catp.to_mesh(Nmesh=Nmesh,BoxSize=box_size,compensated=compensated,interlaced=interlaced)
meshran = catran.to_mesh(Nmesh=Nmesh,BoxSize=box_size,compensated=compensated,interlaced=interlaced)
array_d = meshp.compute()
array_s = meshran.compute()
arrayrecon = array_d-array_s
mesh_recon = ArrayMesh(arrayrecon,BoxSize=box_size)
LOS = [0,0,1]
kmin=0.005 #leftmost edge of first bin
dk=0.005
kmax=1 #rightmost edge of last bin
kbins = np.arange(kmin, kmax, dk)
kbincenters = 0.5 * (kbins[1:] + kbins[:-1])
r_cross = FFTPower(mesh_init, mode='2d', Nmesh=Nmesh, Nmu=2*Nmu, dk=dk, second=mesh_recon,los=LOS, kmin=kmin, kmax=kmax)
r_cross_1d = FFTPower(mesh_init, mode='1d', Nmesh=Nmesh, dk=dk, second=mesh_recon, kmin=kmin, kmax=kmax)
r_auto_recon = FFTPower(mesh_recon,mode='2d', Nmesh=Nmesh, Nmu=2*Nmu, dk=dk, los=LOS, poles=[0,2,4], kmin=kmin, kmax=kmax)
r_auto_recon_1d = FFTPower(mesh_recon, mode='1d', Nmesh=Nmesh, dk=dk, kmin=kmin, kmax=kmax)
r_auto_init = FFTPower(mesh_init,mode='2d', Nmesh=Nmesh, Nmu=2*Nmu, dk=dk, los=LOS, kmin=kmin, kmax=kmax)
r_auto_init_1d = FFTPower(mesh_recon, mode='1d', Nmesh=Nmesh, dk=dk, kmin=kmin, kmax=kmax)
Gf = ns.Gf
b = ns.b
out = ns.out
pg2d=[]
pg1d=[]
pg1d.append(r_cross_1d.power['k'])
pg1d.append(r_cross_1d.power['power'].real/r_auto_init_1d.power['power'].real/(Gf*b))
np.savetxt(out+'1Dpropagator.txt',np.column_stack([pg1d[0],pg1d[1]]),header='Gf= %lf \nb + %lf \n dk=dk#\nkmean \t \t C(k)' % (Gf,b))
pg2d.append(r_cross.power[:,Nmu]['k'])
pg2d.append(r_cross.power[:,Nmu]['power'].real/r_auto_init.power[:,Nmu]['power'].real/(Gf*b))
pg2d.append(r_cross.power[:,2*Nmu-1]['power'].real/r_auto_init.power[:,2*Nmu-1]['power'].real/(Gf*b))
np.savetxt(out+'2Dpropagator.txt',np.column_stack([pg2d[0],pg2d[1],pg2d[2]]),header='Gf = %lf \nb = %lf \ndk=dk \nkmean \t \t C(k,mu=0.05) \t \t C(k,mu=0.95' %(Gf,b))
reconell=[]
poles = r_auto_recon.poles
#reconell.append(poles['k'])
reconell.append(kbincenters)
for ell in [0,2,4]:
P = poles['power_%d' %ell].real
if ell==0:
if poles.attrs['shotnoise']==0:
print(f"Nbodykit computed shot noise 0. Correcting with 1/nbar")
poles.attrs['shotnoise'] = (box_size**3 / n_elem) + (box_size**3 / n_rand)
print(f"New shot noise = {poles.attrs['shotnoise']}")
P = P-poles.attrs['shotnoise']
reconell.append(P)
np.savetxt(out+'reconstructed_multipoles.txt',np.column_stack([reconell[0],reconell[1],reconell[2],reconell[3]]),header='dk=%lf \nshotnoise= %lf \nkcenter \t \t \t P_0-shot \t \t \t P_2 \t \t \t P_4' %(dk, poles.attrs['shotnoise']))
print('%lf seconds' %(time.time()-t0))