-
Notifications
You must be signed in to change notification settings - Fork 10
/
2_eta_0.py
179 lines (149 loc) · 8.63 KB
/
2_eta_0.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
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
'''
Created on October 9, 2019
1) shuffle event magnitudes and times
2) computer NND values for all event pairs
3) use 1st percentile as estimate for eta_0
-> note that eta_0 is required to separate clustered
and background events in the following analyses steps
4) eta_0 is saved in /data directori
file_out = [file_in]_Mc_[mc]_eta_0.txt
@author: tgoebel
'''
#------------------------------------------------------------------------------
import matplotlib as mpl
#mpl.use( 'Agg') # turn off interactive plot
import matplotlib.pyplot as plt
import numpy as np
import scipy.io
import os
#------------------------------my modules--------------------------------------
import src.clustering as clustering
import src.data_utils as data_utils
from src.EqCat import *
eqCat = EqCat( ) # original cat
ranCat = EqCat() # randomized, Poissonian catalog
eqCatMc = EqCat() # catalog above completeness
np.random.seed( 123456)
#=================================1==============================================
# dir, file, params
#================================================================================
dir_in = 'data'
file_in= 'hs_1981_2011_all.mat'
#file_b = '%s_b_Mc_D.txt'%(fileIn.split('.')[0])
dPar = { 'aMc' : np.array([3.0, 4.0]), #np.array( [2.0, 2.5, 3.0, 3.5]),
# fractal dimension and b for eq. (1)
'D' : 1.6, # TODO: - these values should be constrained based on the data
'b' : 1.0, # use: https://github.com/tgoebel/magnitude-distribution for b-value
# number of bootstraps for randomized catalogs
'nBoot' : 100,
#=================plotting==============
'eta_binsize' : .3,
'cmin' : 1,
'xmin' : -13, 'xmax' : 0,
## R-T plot
'binx' : .1, 'biny' : .1,# used for density and gaussian smoothing
'sigma' : None, #if None: default = n**(-1./(d+4)), or set Gaussian bandwidth
'Tmin' : -8, 'Tmax' : 0,
'Rmin' : -5, 'Rmax' : 3,
'cmap' : plt.cm.RdYlGn_r,
'showPlot' : False,
}
#================================================================================
# load data, event selection
#================================================================================
eqCat.loadMatBin( os.path.join( dir_in, file_in))
print( 'total no. of events', eqCat.size())
eqCat.selectEvents( dPar['aMc'][0], None, 'Mag')
#eqCat.selectEvents( tmin, tmax, 'Time')
print( 'no. of events after initial selection', eqCat.size())
# project to equi-distant coordiante system for cartesian distances
eqCat.toCart_coordinates( projection = 'eqdc')#'eqdc')
for f_Mc in dPar['aMc']:
print( '-------------- current Mc:', f_Mc, '---------------------')
# select magnitude range
eqCatMc.copy( eqCat)
eqCatMc.selectEvents( f_Mc, None, 'Mag')
print( 'catalog size after MAG selection', eqCat.size())
# this dictionary is used in module: clustering
dConst = {'Mc' : f_Mc,
'b' : dPar['b'],
'D' : dPar['D']}
#=============================2===================================================
# randomize catalog
#=================================================================================
a_Eta_0 = np.zeros( dPar['nBoot'])
for i_Bs in range( dPar['nBoot']):
ranCat.copy( eqCatMc)
ranCat.data['X'] = np.random.uniform( eqCatMc.data['X'].min(), eqCatMc.data['X'].max(), size = eqCatMc.size())
ranCat.data['Y'] = np.random.uniform( eqCatMc.data['Y'].min(), eqCatMc.data['Y'].max(), size = eqCatMc.size())
ranCat.data['Time'] = clustering.rand_rate_uni( eqCatMc.size(), eqCatMc.data['Time'].min(), eqCatMc.data['Time'].max())
ranCat.sortCatalog( 'Time')
#==================================3=============================================
# compute space-time-magnitude distance, histogram
#================================================================================
dNND = clustering.NND_eta( ranCat, dConst, M0 = 0, correct_co_located = True,
verbose = False)
a_Eta_0[i_Bs] = round( np.percentile( np.log10(dNND['aNND']), 1), 5)
print( 'nBoot', i_Bs+1,'out of', dPar['nBoot'], 'eta 0 - 1st', np.percentile( np.log10(dNND['aNND']), 1))
if dPar['showPlot'] == True: # plots to check if everything is working
#=================================4==============================================
# plot NND histogram
#================================================================================
plt.figure( 1, figsize = (10,5))
ax = plt.axes( [.12, .12, .83, .83])
ax.hist( np.log10( dNND['aNND']), np.arange( dPar['xmin'], dPar['xmax'], dPar['eta_binsize']),
color = '.5', label = 'Mc = %.1f'%( f_Mc), align = 'mid', rwidth=.9)
ax.plot( [-5, -5], ax.get_ylim(), 'w-', lw = 2, )
ax.plot( [-5, -5], ax.get_ylim(), 'k--', lw = 2, )
ax.plot( [a_Eta_0[i_Bs], a_Eta_0[i_Bs]], ax.get_ylim(), 'w-', lw = 2, label = '$N_\mathrm{tot}$=%i'%( ranCat.size()))
ax.plot( [a_Eta_0[i_Bs], a_Eta_0[i_Bs]], ax.get_ylim(), 'r--', lw = 2, label = '$N_\mathrm{cl}$=%i'%( dNND['aNND'][dNND['aNND']<1e-5].shape[0]))
ax.legend( loc = 'upper left')
ax.set_xlabel( 'NND, log$_{10} \eta$')
ax.set_ylabel( 'Number of Events')
ax.grid( 'on')
ax.set_xlim( dPar['xmin'], dPar['xmax'])
#==================================4==============================================================
# T-R density plot
#=================================================================================================
catChild = EqCat()
catParent= EqCat()
catChild.copy( ranCat)
catParent.copy( ranCat)
catChild.selEventsFromID( dNND['aEqID_c'], repeats = True)
catParent.selEventsFromID( dNND['aEqID_p'], repeats = True)
print( catChild.size(), catParent.size(), eqCatMc.size())
a_R, a_T = clustering.rescaled_t_r( catChild, catParent, dConst, correct_co_located = True)
a_Tbin = np.arange( dPar['Tmin'], dPar['Tmax']+2*dPar['binx'], dPar['binx'])
a_Rbin = np.arange( dPar['Rmin'], dPar['Rmax']+2*dPar['biny'], dPar['biny'])
a_log_T = np.log10( a_T)
a_log_R = np.log10( a_R)
XX, YY, ZZ = data_utils.density_2D( a_log_T, a_log_R, a_Tbin, a_Rbin, sigma = dPar['sigma'])
plt.figure(2, figsize= (8,10))
ax = plt.subplot(111)
ax.set_title( 'Nearest Neighbor Pairs in R-T')
#------------------------------------------------------------------------------
normZZ = ZZ*( dPar['binx']*dPar['biny']*eqCatMc.size())
plot1 = ax.pcolormesh( XX, YY, normZZ, cmap=dPar['cmap'])
cbar = plt.colorbar(plot1, orientation = 'horizontal', shrink = .5, aspect = 20,)
#ax.plot( np.log10( a_T), np.log10( a_R), 'wo', ms = 1.5, alpha = .2)
# plot eta_0 to divide clustered and background mode
ax.plot( [dPar['Tmin'], dPar['Tmax']], -np.array([dPar['Tmin'], dPar['Tmax']])+a_Eta_0[i_Bs], '-', lw = 1.5, color = 'w' )
ax.plot( [dPar['Tmin'], dPar['Tmax']], -np.array([dPar['Tmin'], dPar['Tmax']])+a_Eta_0[i_Bs],'--', lw = 1.5, color = '.5' )
#-----------------------labels and legends-------------------------------------------------------
#cbar.set_label( 'Event Pair Density [#ev./dRdT]')
cbar.set_label( 'Number of Event Pairs',labelpad=-40)
ax.set_xlabel( 'Rescaled Time')
ax.set_ylabel( 'Rescaled Distance')
ax.set_xlim( dPar['Tmin'], dPar['Tmax'])
ax.set_ylim( dPar['Rmin'], dPar['Rmax'])
plt.show()
#=================================3==============================================
# save results
#================================================================================
f_eta_0 = a_Eta_0.mean()
print( 'medium eta_0', a_Eta_0.mean()
file_out = '%s/%s_Mc_%.1f_eta_0.txt'%(dir_in, file_in, f_Mc)
np.savetxt( file_out, np.array([f_eta_0]), fmt = '%10.3f', header='eta_0')
print( 'save results', file_out
scipy.io.savemat(file_out.replace('txt','mat'),
{'eta_0': f_eta_0, 'eta_BS' : a_Eta_0,}, do_compression=True)