-
Notifications
You must be signed in to change notification settings - Fork 0
/
match486_sym90.py
110 lines (70 loc) · 4.04 KB
/
match486_sym90.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
import importlib
import pickle
import pandas as pd
import time, os
import numpy as np
import matplotlib.pyplot as plt
from xedrift.tpcClass import TPC_X1T
import xedrift
from xedrift.chargeMatching import *
import sys
# Parameters
print("Hello!")
model_dir = './models' #path to electric field data
data_dir = './data' #path to experimental data
n_iterations=1000 #number of matching iterations to run
tpc_x1t = TPC_X1T() #for drift velocity, tpc walls for out of bounds condition
model_name = 'XE1T_3D_sp_eff_pot_48x6' #folder with electric field data inside model_dir
charges_n_phi = 48 #number of angular subdivisions for charge basis vectors
charges_n_z = 6 #number of vertical subdivisions for charge basis vectors
nudge_single = True #only change one coefficient per iteration (gibbs sampling)
nudge_scale = 0.2 #gaussian proposal distribution sigma
data_name = 'kr83m/kr83m_sr1_{}'.format(sys.argv[1]) #folder with experimental data in data_dir, get dataset id from command line argument
num_coefficients = charges_n_z * (charges_n_phi/2)/4 #number of coefficients for matching, /2 for uncharged fixed reflectors, /4 for 90 degree symmetry
state_0 = np.array([0.0] * num_coefficients) #initial state: zero charge
import itertools
#create rectangular grid in observation space for tetrahedralization
xys = np.linspace(-46,46,12)
dts = np.linspace(0,780,10)
grid = np.array(list(itertools.product(xys,xys,dts)))
# End parameters
model_path = model_dir + model_name
data_path = data_dir + data_name
results_dir = '{}/matchresults/{}/{}/'.format(data_dir,data_name,model_name) #save results here
make_dir(results_dir)
print("Results in: " + results_dir)
# Prepare experimental data
data = pd.read_hdf(data_path + '.hdf')
data['drift_time_us'] = data.drift_time/1e3
grid = preprocess_grid(grid, data) #Remove grid points that fall outside the data
triangulation = triangulate_grid(grid)
simp_counts = count_events(triangulation,data) #count events in observation space simplices
tri_mask = mask_triangulation(triangulation,grid,simp_counts) #sanity checks on simplex aspect ratio, volume and event count
simps_obs = triangulation.simplices.copy()[tri_mask]
print("Events in triangulation: {:.2f}%".format(simp_counts.events.sum()/len(data)))
print("Events in masked triangulation: {:.2f}%".format(simp_counts[tri_mask].events.sum()/len(data)))
#superpos = makeSuperposition(model_path,charges_n_phi,charges_n_z) #code for matching without symmetry
#load electric field data (basis vectors) with uncharged fixed reflectors, 90 degree symmetry
fields = [model_path + '/no_charge/electric_field.txt']
for i in range(num_coefficients):
fields += [model_path + '/symmetry_90_no_fixed/field_{}.npz'.format(i)]
assert(len(state_0)==len(fields)-1)
superpos = Superposition(fields) #load electric field data (basis vectors)
from functools import partial
log_l_f = partial(log_like_no_pool,tpc_x1t,superpos,grid,triangulation,tri_mask,simp_counts) #log-likelihood function for metropolis algorithm
match_data = init_match_data(results_dir, log_l_f, state_0) #set initial state or load saved state from previous run if found
state = match_data.states[-1]
log_l = match_data.log_ls[-1]
for i in range(n_iterations):
start_time = time.time()
state, log_l, accepted, *meta = MHStep(state,log_l_f,log_l,scale=nudge_scale,nudgeSingle=nudge_single) #single step of metropolis algorithm
if(accepted):
print("Accepted! {}".format("single" if nudge_single else "all"))
time_elapsed = (time.time() - start_time)
match_data.states.append(state)
match_data.log_ls.append(log_l)
match_data.accs.append(accepted)
match_data.metas.append(meta)
print("---------------- {} | {:.2f} | {:.2f}%".format(i,time_elapsed,100*sum(match_data.accs[-50:])/50))
if(accepted): #save results
pickle.dump({'match_data' : match_data, 'model_name' : model_name, 'data_name' : data_name, 'grid' : grid, 'tri_data' : (triangulation.simplices,tri_mask), 'simp_counts': simp_counts, 'symmetry' : 'sym486_90'} ,open(results_dir+'/matchresult.p','wb'))