Skip to content
Snippets Groups Projects

Analysis of truth/reconstruction associations

Merged Tooba Ali requested to merge truth_reconstruction into master
Compare and
5 files
+ 386
0
Compare changes
  • Side-by-side
  • Inline
Files
5
+ 364
0
 
import os
 
import numpy as np
 
import uproot as ur
 
import awkward as ak
 
import matplotlib.pyplot as plt
 
import matplotlib as mpl
 
import mplhep
 
import argparse
 
 
parser = argparse.ArgumentParser()
 
parser.add_argument('--rec_file', type=str, help='Reconstructed track file.')
 
parser.add_argument('--config', type=str, help='Momentum configuration.')
 
parser.add_argument('--nevents', type=float, help='Number of events to process.')
 
parser.add_argument('--results_path', type=str, help='Output directory.')
 
args = parser.parse_args()
 
kwargs = vars(args)
 
 
rec_file = args.rec_file
 
config = args.config
 
Nevents = int(args.nevents)
 
r_path = args.results_path + '/truth_reconstruction/'
 
 
for array in ur.iterate(rec_file + ':events',['MCParticles/MCParticles.generatorStatus',
 
'MCParticles/MCParticles.mass',
 
'MCParticles/MCParticles.PDG',
 
'MCParticles/MCParticles.momentum.x',
 
'MCParticles/MCParticles.momentum.y',
 
'MCParticles/MCParticles.momentum.z',
 
'ReconstructedParticles/ReconstructedParticles.energy',
 
'ReconstructedParticles/ReconstructedParticles.PDG',
 
'ReconstructedParticles/ReconstructedParticles.momentum.x',
 
'ReconstructedParticles/ReconstructedParticles.momentum.y',
 
'ReconstructedParticles/ReconstructedParticles.momentum.z',
 
'ReconstructedParticlesAssoc/ReconstructedParticlesAssoc.simID',
 
'ReconstructedParticlesAssoc/ReconstructedParticlesAssoc.recID',
 
],step_size=Nevents):
 
# generatorStatus = array['MCParticles/MCParticles.generatorStatus']
 
mass = array['MCParticles/MCParticles.mass']
 
PDG_mc = array['MCParticles/MCParticles.PDG']
 
px_mc = array['MCParticles/MCParticles.momentum.x']
 
py_mc = array['MCParticles/MCParticles.momentum.y']
 
pz_mc = array['MCParticles/MCParticles.momentum.z']
 
PDG_rc = array['ReconstructedParticles/ReconstructedParticles.PDG']
 
px_rc = array['ReconstructedParticles/ReconstructedParticles.momentum.x']
 
py_rc = array['ReconstructedParticles/ReconstructedParticles.momentum.y']
 
pz_rc = array['ReconstructedParticles/ReconstructedParticles.momentum.z']
 
simID = array['ReconstructedParticlesAssoc/ReconstructedParticlesAssoc.simID']
 
recID = array['ReconstructedParticlesAssoc/ReconstructedParticlesAssoc.recID']
 
 
momentum_mc = np.sqrt(((px_mc**2)+(py_mc**2)+(pz_mc**2)))
 
momentum_rc = np.sqrt(((px_rc**2)+(py_rc**2)+(pz_rc**2)))
 
 
theta_mc = np.arctan2(np.sqrt(px_mc**2+py_mc**2), pz_mc)
 
phi_mc = np.arctan2(py_mc, px_mc)
 
theta_rc = np.arctan2(np.sqrt(px_rc**2+py_rc**2), pz_rc)
 
phi_rc = np.arctan2(py_rc, px_rc)
 
 
booll = (PDG_mc[simID])==(PDG_rc[recID])
 
MC_list = [ak.Array(momentum_mc[simID][booll]),
 
ak.Array(theta_mc[simID][booll]),
 
ak.Array(phi_mc[simID][booll]),
 
-np.log(np.tan((ak.Array(theta_mc[simID][booll]))/2))]
 
RC_list = [ak.Array(momentum_rc[recID][booll]),
 
ak.Array(theta_rc[recID][booll]),
 
ak.Array(phi_rc[recID][booll]),
 
-np.log(np.tan((ak.Array(theta_rc[recID][booll]))/2))]
 
title_list = ['Momentum','Theta','Phi','Eta']
 
boolean_pion = np.logical_or(ak.Array(PDG_mc[simID][booll])==-211, ak.Array(PDG_mc[simID][booll])==+211)
 
boolean_proton = np.logical_or(ak.Array(PDG_mc[simID][booll])==-2212, ak.Array(PDG_mc[simID][booll])==+2212)
 
boolean_electron = ak.Array(PDG_mc[simID][booll])==11
 
boolean_neutron = ak.Array(PDG_mc[simID][booll])==2112
 
boolean_photon = ak.Array(PDG_mc[simID][booll])==22
 
 
if Nevents == 100:
 
ssize = 1
 
else:
 
ssize = 0.01
 
 
particle = config.split('-')[0].strip()
 
particle_dict = {'e':[boolean_electron,'Electrons'],'pi':[boolean_pion,'Pions']}
 
###################
 
 
 
for i in range(len(MC_list)):
 
X = MC_list[i]
 
Y = RC_list[i]
 
 
X_len = ak.count(X,axis=None)
 
Y_len = ak.count(Y,axis=None)
 
 
if X_len > Y_len:
 
F_boolean = np.ones_like(Y) == 1
 
else:
 
F_boolean = np.ones_like(X) == 1
 
X_F = X[F_boolean]
 
Y_F = Y[F_boolean]
 
 
X_s = np.array(ak.flatten(X_F))
 
Y_s = np.array(ak.flatten(Y_F))
 
 
if i == 3:
 
h, xedges, yedges, image = plt.hist2d(x=X_s,y= Y_s, bins = 11,range = [[-5,5],[-5,5]])
 
plt.close()
 
else:
 
h, xedges, yedges, image = plt.hist2d(x=X_s,y= Y_s, bins = 10)
 
plt.close()
 
 
col_sum = ak.sum(h,axis=-1) #number of events in each (verticle) column
 
norm_h = [] #norm_h is the normalized matrix
 
norm_h_text = [] #display labels matrix
 
for j in range(len(col_sum)):
 
if col_sum[j] != 0:
 
norm_c = h[j]/col_sum[j] #normalized column = column values divide by sum of the column
 
else:
 
norm_c = h[j]
 
norm_h.append(norm_c)
 
norm_c_text = [ '%.3f' % elem for elem in norm_c ] #display value to 3 dp
 
norm_h_text.append(norm_c_text)
 
 
fig, axs = plt.subplots(1, 2, figsize=(20, 10))
 
if i == 3:
 
axs[0].hist2d(x=X_s,y=Y_s, bins = 11,range = [[-5,5],[-5,5]])
 
else:
 
axs[0].hist2d(x=X_s,y=Y_s, bins = 10)
 
mplhep.hist2dplot(H=norm_h,norm=mpl.colors.LogNorm(vmin= 1e-4, vmax= 1),labels=norm_h_text, xbins = xedges, ybins = yedges, ax=axs[1])
 
 
axs[0].set_title('%s Histogram'%(title_list[i]))
 
axs[0].set_xlabel('%s_mc'%(title_list[i]))
 
axs[0].set_ylabel('%s_rc'%(title_list[i]))
 
axs[1].set_xlabel('%s_mc'%(title_list[i]))
 
axs[1].set_ylabel('%s_rc'%(title_list[i]))
 
axs[1].set_title('%s Correlation'%(title_list[i]))
 
 
fig.suptitle('%s %s events'%(config,Nevents))
 
 
plt.savefig(os.path.join(r_path, '%s_correlation_%s.png' % (title_list[i],config)))
 
 
 
#################
 
 
 
 
for i in range(len(MC_list)):
 
X1 = MC_list[i]
 
Y1 = RC_list[i]
 
X_list = [ak.Array(X1),
 
ak.Array(X1[boolean_pion]),
 
ak.Array(X1[boolean_proton]),
 
ak.Array(X1[boolean_electron]),
 
ak.Array(X1[boolean_neutron]),
 
ak.Array(X1[boolean_photon])]
 
Y_list = [ak.Array(Y1),
 
ak.Array(Y1[boolean_pion]),
 
ak.Array(Y1[boolean_proton]),
 
ak.Array(Y1[boolean_electron]),
 
ak.Array(Y1[boolean_neutron]),
 
ak.Array(Y1[boolean_photon])]
 
 
X_plot = list(np.zeros(6))
 
Y_plot = list(np.zeros(6))
 
 
for j in range(len(X_list)):
 
X = X_list[j]
 
Y = Y_list[j]
 
X_len = ak.count(X,axis=None)
 
Y_len = ak.count(Y,axis=None)
 
 
if X_len > Y_len:
 
F_boolean = np.ones_like(Y) == 1
 
else:
 
F_boolean = np.ones_like(X) == 1
 
X_F = X[F_boolean]
 
Y_F = Y[F_boolean]
 
 
X_s = np.array(ak.flatten(X_F))
 
Y_s = np.array(ak.flatten(Y_F))
 
 
if i == 0:
 
ratio = np.array((ak.Array(Y_s)/ak.Array(X_s)))
 
else:
 
ratio = np.array((ak.Array(Y_s)-(ak.Array(X_s))))
 
X_plot[j] = X_s
 
Y_plot[j] = ratio
 
 
fig = plt.figure()
 
gs = fig.add_gridspec(3, 2, wspace=0)
 
(ax1, ax2), (ax3, ax4),(ax5, ax6) = gs.subplots(sharex=True, sharey=True)
 
# fig.suptitle('')
 
if i == 1:
 
X_plot[0],X_plot[1],X_plot[2],X_plot[3],X_plot[4],X_plot[5] = -X_plot[0],-X_plot[1],-X_plot[2],-X_plot[3],-X_plot[4],-X_plot[5]
 
 
ax1.scatter(X_plot[0], Y_plot[0], s = ssize)
 
ax2.scatter(X_plot[1], Y_plot[1], s = ssize)
 
ax3.scatter(X_plot[2], Y_plot[2], s = ssize)
 
ax4.scatter(X_plot[3], Y_plot[3], s = ssize)
 
ax5.scatter(X_plot[4], Y_plot[4], s = ssize)
 
ax6.scatter(X_plot[5], Y_plot[5], s = ssize)
 
 
ax_list = [ax1,ax2,ax3,ax4,ax5]
 
if i == 0:
 
ax1.set_ylabel('rc/mc')
 
ax3.set_ylabel('rc/mc')
 
ax5.set_ylabel('rc/mc')
 
tratio ='ratio'
 
for ax in ax_list:
 
ax.set_yscale('log')
 
ax.set_xscale('log')
 
else:
 
ax1.set_ylabel('rc-mc')
 
ax3.set_ylabel('rc-mc')
 
ax5.set_ylabel('rc-mc')
 
tratio ='difference'
 
 
ax2.set_title('Pions')
 
ax3.set_title('Protons')
 
ax4.set_title('Electrons')
 
ax5.set_title('Neutrons')
 
ax6.set_title('Photons')
 
if i == 1:
 
ax5.set_xlabel('- %s mc'%(title_list[i]))
 
ax6.set_xlabel('- %s mc'%(title_list[i]))
 
else:
 
ax5.set_xlabel('%s'%(title_list[i]))
 
ax6.set_xlabel('%s'%(title_list[i]))
 
fig.set_figwidth(20)
 
fig.set_figheight(10)
 
ax1.set_title('%s %s %s %s events'%(title_list[i],tratio,config,Nevents))
 
plt.savefig(os.path.join(r_path, '%s_%s_%s.png' % (title_list[i],tratio,config)))
 
 
 
###############
 
 
 
M_list = [ak.Array(momentum_mc[simID][booll]),
 
ak.Array(momentum_mc[simID][booll][boolean_pion]),
 
ak.Array(momentum_mc[simID][booll][boolean_proton]),
 
ak.Array(momentum_mc[simID][booll][boolean_electron]),
 
ak.Array(momentum_mc[simID][booll][boolean_neutron]),
 
ak.Array(momentum_mc[simID][booll][boolean_photon])]
 
 
for i in range(1,len(MC_list),1):
 
X1 = MC_list[i]
 
Y1 = RC_list[i]
 
X_list = [ak.Array(X1),
 
ak.Array(X1[boolean_pion]),
 
ak.Array(X1[boolean_proton]),
 
ak.Array(X1[boolean_electron]),
 
ak.Array(X1[boolean_neutron]),
 
ak.Array(X1[boolean_photon])]
 
Y_list = [ak.Array(Y1),
 
ak.Array(Y1[boolean_pion]),
 
ak.Array(Y1[boolean_proton]),
 
ak.Array(Y1[boolean_electron]),
 
ak.Array(Y1[boolean_neutron]),
 
ak.Array(Y1[boolean_photon])]
 
 
X_plot = list(np.zeros(6))
 
Y_plot = list(np.zeros(6))
 
 
for j in range(len(X_list)):
 
X = X_list[j]
 
Y = Y_list[j]
 
M_mc = M_list[j]
 
boolean_M = np.ones_like(M_mc) == 1
 
X_F = X[boolean_M]
 
Y_F = Y[boolean_M]
 
X_s = np.array(ak.flatten(X_F))
 
Y_s = np.array(ak.flatten(Y_F))
 
M_s = np.array(ak.flatten(M_mc))
 
ratio = np.array((ak.Array(Y_s)-(ak.Array(X_s))))
 
X_plot[j] = M_s
 
Y_plot[j] = ratio
 
 
 
fig = plt.figure()
 
gs = fig.add_gridspec(3, 2, wspace=0)
 
(ax1, ax2), (ax3, ax4),(ax5, ax6) = gs.subplots(sharex=True, sharey=True)
 
# fig.suptitle('')
 
 
ax1.scatter(X_plot[0], Y_plot[0], s = ssize)
 
ax2.scatter(X_plot[1], Y_plot[1], s = ssize)
 
ax3.scatter(X_plot[2], Y_plot[2], s = ssize)
 
ax4.scatter(X_plot[3], Y_plot[3], s = ssize)
 
ax5.scatter(X_plot[4], Y_plot[4], s = ssize)
 
ax6.scatter(X_plot[5], Y_plot[5], s = ssize)
 
 
ax_list = [ax1,ax2,ax3,ax4,ax5]
 
for ax in ax_list:
 
ax.set_xscale('log')
 
else:
 
ax1.set_ylabel('rc-mc')
 
ax3.set_ylabel('rc-mc')
 
ax5.set_ylabel('rc-mc')
 
 
 
ax2.set_title('Pions')
 
ax3.set_title('Protons')
 
ax4.set_title('Electrons')
 
ax5.set_title('Neutrons')
 
ax6.set_title('Photons')
 
ax5.set_xlabel('Momentum mc')
 
ax6.set_xlabel('Momentum mc')
 
fig.set_figwidth(20)
 
fig.set_figheight(10)
 
 
ax1.set_title('%s Difference Vs Momentum %s %s events'%(title_list[i],config,Nevents))
 
 
plt.savefig(os.path.join(r_path, '%s_difference_vs_momentum_%s.png' % (title_list[i],config)))
 
 
 
################
 
if particle in particle_dict.keys():
 
boolean_particle = particle_dict[particle][0]
 
else:
 
boolean_particle = boolean_photon
 
theta_mc_fil = ak.Array(theta_mc[simID][booll])[boolean_particle]
 
theta_rc_fil = ak.Array(theta_rc[recID][booll])[boolean_particle]
 
phi_mc_fil = ak.Array(phi_mc[simID][booll])[boolean_particle]
 
phi_rc_fil = ak.Array(phi_rc[recID][booll])[boolean_particle]
 
 
theta_mc_fil_len = ak.count(theta_mc_fil,axis=None)
 
theta_rc_fil_len = ak.count(theta_rc_fil,axis=None)
 
 
if theta_mc_fil_len > theta_rc_fil_len:
 
F_boolean = np.ones_like(theta_rc_fil) == 1
 
else:
 
F_boolean = np.ones_like(theta_mc_fil) == 1
 
theta_mc_F = np.array(ak.flatten(theta_mc_fil[F_boolean]))
 
theta_rc_F = np.array(ak.flatten(theta_rc_fil[F_boolean]))
 
phi_mc_F = np.array(ak.flatten(phi_mc_fil[F_boolean]))
 
phi_rc_F = np.array(ak.flatten(phi_rc_fil[F_boolean]))
 
ratio = np.array((ak.Array(theta_rc_F)-(ak.Array(theta_mc_F))))
 
 
fig = plt.figure()
 
gs = fig.add_gridspec(2, 2, wspace=0.01)
 
(ax1, ax2), (ax3, ax4) = gs.subplots(sharex=True, sharey=True)
 
 
ax1.scatter(-theta_mc_F, ratio, s = ssize)
 
ax2.scatter(-theta_rc_F, ratio, s = ssize)
 
ax3.scatter(-theta_mc_F, phi_mc_F, s = ssize)
 
ax4.scatter(-theta_rc_F, phi_rc_F, s = ssize)
 
 
ax1.set_ylabel('rc-mc')
 
ax2.set_ylabel('rc-mc')
 
ax3.set_ylabel('Phi mc')
 
ax4.set_ylabel('Phi rc')
 
ax3.set_xlabel('- Theta mc')
 
ax4.set_xlabel('- Theta rc')
 
fig.set_figwidth(20)
 
fig.set_figheight(10)
 
 
 
if particle in particle_dict.keys():
 
particle_name = particle_dict[particle][1]
 
else:
 
particle_name = 'Photons'
 
 
fig.suptitle('%s in %s %s events'%(particle_name,config,Nevents))
 
plt.savefig(os.path.join(r_path, '%s_%s.png' % (particle_name,config)))
 
 
 
 
 
Loading