-
Notifications
You must be signed in to change notification settings - Fork 2
/
drum_ncoh.py
364 lines (267 loc) · 14.6 KB
/
drum_ncoh.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
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
#Source codes of the publication "Boundary Graph Neural Networks for 3D Simulations"
#Copyright (C) 2023 Sebastian Lehner, Andreas Mayr
#This program is free software: you can redistribute it and/or modify
#it under the terms of the GNU General Public License as published by
#the Free Software Foundation, version 3 of the License.
#This program is distributed in the hope that it will be useful,
#but WITHOUT ANY WARRANTY; without even the implied warranty of
#MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
#GNU General Public License for more details.
#You should have received a copy of the GNU General Public License
#along with this program. If not, see <https://www.gnu.org/licenses/>.
#
#The license file for GNU General Public License v3.0 is available here:
#https://github.com/ml-jku/bgnn/blob/master/licenses/own/LICENSE_GPL3
import os
import numpy as np
import matplotlib
import matplotlib.pyplot as plt
from matplotlib.offsetbox import AnchoredOffsetbox, TextArea, HPacker, VPacker
from PIL import Image
import warnings
warnings.filterwarnings("ignore")
homeDir=os.getenv('HOME')
plotDir=os.path.join(os.environ['HOME'], "bgnnPlots")
runid = '103'
gt=np.load("/system/user/mayr-data/BGNNRuns/trajectories/drum/runs2/model/gtp_"+runid+"_6.npy")
pred=np.load("/system/user/mayr-data/BGNNRuns/trajectories/drum/runs2/model/predp_"+runid+"_6.npy")
runs_gt = []
run_ids = list(range(105,120))
time1=50
time2=150
ind=time1 #use shift of 50
runs_gt.append(gt[ind:(gt.shape[0])])
ind=time2 #use shift of 150
runs_gt.append(gt[ind:(gt.shape[0])])
#particle number variation with same distributions
#10%: 105+(int(runid)-100)*5+0
#20%: 105+(int(runid)-100)*5+1
#30%: 105+(int(runid)-100)*5+2
#40%: 105+(int(runid)-100)*5+3
#50%: 105+(int(runid)-100)*5+4
accessId=105+(int(runid)-100)*5+4
ind=time1
runs_gt.append(np.load("/system/user/mayr-data/BGNNRuns/trajectories/drum/runs2/model/gtp_"+str(accessId)+"_6.npy")[ind:(gt.shape[0])])
ind=time2
runs_gt.append(np.load("/system/user/mayr-data/BGNNRuns/trajectories/drum/runs2/model/gtp_"+str(accessId)+"_6.npy")[ind:(gt.shape[0])])
#cut to have the same length for all
startInd=time2
endInd=gt.shape[0]-startInd
gt=gt[:endInd]
pred=pred[:endInd]
for i in range(0,len(runs_gt)):
runs_gt[i]=runs_gt[i][0:endInd]
#main line added to uncertainty computation
runs_gt.append(gt)
#(time, particle, coord) -> (particle, time, coord)
pred = pred.transpose((1,0,2))
gt = gt.transpose((1,0,2))
for n in range(len(runs_gt)):
runs_gt[n] = runs_gt[n].transpose((1,0,2))
#get total min & max positions for binnings
minTot = np.min(gt,axis=(0,1))
maxTot = np.max(gt,axis=(0,1))
xbins = np.linspace(minTot[0],maxTot[0],20)
zbins = np.linspace(minTot[2],maxTot[2],20)
#input is array: part,time,dim
def get_zarr(pos):
vlist=[]
sum=0
tsStart = 30
v = np.diff(pos,axis=1)
for i in range(len(zbins)-1):
mask = (pos[:,tsStart:,2]>=zbins[i]) * (pos[:,tsStart:,2]<zbins[i+1])
mask = mask[:,:-1]
vSel = v[:,tsStart:,:][mask]
sum+=vSel.shape[0]
VSel = vSel.sum(axis=0)
vlist.append(VSel)
return np.array(vlist)
zbinarr = np.array(zbins[:-1])
vpred = get_zarr(pred)
vgt = get_zarr(gt)
vgt_run_list = []
for gtrun in runs_gt:
vgt_run_list.append(get_zarr(gtrun))
rms_v_runs = np.std(np.array(vgt_run_list),axis=0)
matplotlib.rc('font',**{'family':'serif','serif':['Times'], 'size' : 20})
matplotlib.rc('axes', labelsize=30)
matplotlib.rc('legend', fontsize=30)
matplotlib.rc('text', usetex=True)
params= {'text.latex.preamble' : r'\usepackage{amsmath}'}
params["legend.labelspacing"] = 0.15
params["legend.columnspacing"] = 0.7
params["legend.borderpad"] = 0.2
params["legend.handletextpad"] = 0.3
plt.rcParams.update(params)
matplotlib.rc('font',**{'family':'serif','serif':['Times'], 'size' : 70})
axlsize=130
matplotlib.rc('axes', labelsize=axlsize)
matplotlib.rc('legend', fontsize=100)
plt.figure(figsize=(20,20))
plt.subplots_adjust(bottom=0.20, top=0.95, left=0.25, right=0.95)
plt.grid(b=True, which='major', linewidth='1', linestyle='--', color='grey', dashes=(2,40,2,40))
plt.plot(zbinarr,vpred[:,0],'b--', dashes=(4, 4), linewidth=7.0)
plt.plot(zbinarr,vgt[:,0],'b-', linewidth=5.0)
plt.fill_between(zbinarr,vgt[:,0] - rms_v_runs[:,0], vgt[:,0] + rms_v_runs[:,0], color='b', alpha=0.2, linewidth=0.0)
plt.plot(zbinarr,vpred[:,2],'r--', dashes=(4, 4), linewidth=7.0)
plt.plot(zbinarr,vgt[:,2],'r-', linewidth=5.0)
plt.fill_between(zbinarr,vgt[:,2] - rms_v_runs[:,2], vgt[:,2] + rms_v_runs[:,2], color='r', alpha=0.2, linewidth=0.0)
inarry=np.concatenate([vpred[:,0], vpred[:,2], vgt[:,0], vgt[:,2]])
inarr=zbinarr
myarr=[inarr.min()+0.1*inarr.ptp(), inarr.max()-0.1*inarr.ptp()]
nr1=10**(np.ceil(np.log10(np.abs(myarr[0]))))
nr2=10**(np.ceil(np.log10(np.abs(myarr[1]))))
lsp=np.linspace(np.round(myarr[0]/min(nr1,nr2), 1)*min(nr1,nr2), np.round(myarr[1]/min(nr1,nr2), 1)*min(nr1,nr2), 3)
plt.axes().xaxis.set_ticks(lsp)
myarr=[inarr.min()-0.0*inarr.ptp(), inarr.max()+0.0*inarr.ptp()]
plt.xlim([myarr[0], myarr[1]])
inarr=inarry
myarr=[inarr.min()+0.1*inarr.ptp(), inarr.max()-0.1*inarr.ptp()]
nr1=10**(np.ceil(np.log10(np.abs(myarr[0]))))*10
nr2=10**(np.ceil(np.log10(np.abs(myarr[1]))))*10
lsp=np.linspace(np.round(myarr[0]/min(nr1,nr2), 1)*min(nr1,nr2), np.round(myarr[1]/min(nr1,nr2), 1)*min(nr1,nr2), 3)
plt.axes().yaxis.set_ticks(lsp)
myarr=[inarr.min()-0.1*inarr.ptp(), inarr.max()+0.1*inarr.ptp()]
plt.ylim([myarr[0]-0.35*(myarr[1]-myarr[0]), myarr[1]])
plt.tick_params(axis='both', which='major', pad=15)
ybox6=TextArea(r".\hspace*{50cm}\phantom{Flow $x$/}$z$\hspace*{50cm}.", textprops=dict(color="r",size=axlsize,rotation=90,ha='left',va='center'))
ybox5=TextArea(r".\hspace*{50cm}\phantom{Flow $x$}/\phantom{$z$}\hspace*{50cm}.", textprops=dict(color="black", size=axlsize,rotation=90,ha='left',va='center'))
ybox2=TextArea(r".\hspace*{50cm}\phantom{Flow }$x$\phantom{/$z$}\hspace*{50cm}.", textprops=dict(color="b",size=axlsize,rotation=90,ha='left',va='center'))
ybox1=TextArea(r".\hspace*{50cm}Flow \phantom{$x$/$z$}\hspace*{50cm}.", textprops=dict(color="black", size=axlsize,rotation=90,ha='left',va='center'))
xbox1=TextArea(r"Position z", textprops=dict(color="black", size=axlsize, ha='center', va='top'))
anchored_ybox6=AnchoredOffsetbox(loc='lower left', child=ybox6, pad=0.0, frameon=False, bbox_to_anchor=(-0.30, 0.5), bbox_transform=plt.axes().transAxes, borderpad=0.)
anchored_ybox5=AnchoredOffsetbox(loc='lower left', child=ybox5, pad=0.0, frameon=False, bbox_to_anchor=(-0.30, 0.5), bbox_transform=plt.axes().transAxes, borderpad=0.)
anchored_ybox2=AnchoredOffsetbox(loc='lower left', child=ybox2, pad=0.0, frameon=False, bbox_to_anchor=(-0.30, 0.5), bbox_transform=plt.axes().transAxes, borderpad=0.)
anchored_ybox1=AnchoredOffsetbox(loc='lower left', child=ybox1, pad=0.0, frameon=False, bbox_to_anchor=(-0.30, 0.5), bbox_transform=plt.axes().transAxes, borderpad=0.)
anchored_xbox1=AnchoredOffsetbox(loc='lower left', child=xbox1, pad=0.0, frameon=False, bbox_to_anchor=(0.5, -0.15), bbox_transform=plt.axes().transAxes, borderpad=0.)
plt.axes().add_artist(anchored_ybox6)
plt.axes().add_artist(anchored_ybox5)
plt.axes().add_artist(anchored_ybox2)
plt.axes().add_artist(anchored_ybox1)
plt.axes().add_artist(anchored_xbox1)
legend_elements=[plt.Line2D([0], [0], color='black', linestyle='--', dashes=(4, 6), linewidth=7.0, label='Prediction'),
plt.Line2D([0], [0], color='black', linewidth=5.0, label='Ground Truth')]
plt.legend(handles=legend_elements, loc="lower center", bbox_to_anchor=(0.5, -0.05), ncol=1, fancybox=False, shadow=False, frameon=False, handletextpad=0.5, labelspacing=0.0, handlelength=2.25, columnspacing=10, bbox_transform=plt.axes().transAxes)
plt.savefig(os.path.join(plotDir, "noncoh_zFlow_drum_"+runid+".png"))
plt.savefig(os.path.join(plotDir, "noncoh_zFlow_drum_"+runid+".pdf"))
plt.close()
im = Image.open(os.path.join(plotDir, "noncoh_zFlow_drum_"+runid+".png"))
im
def get_entropy_histos(arr,dim=2):
ts = 0
x_max=0.25
x_min=-0.25
y_max=0.25
y_min=-0.25
z_max=0.25
z_min=-0.25
xbins = np.linspace(x_min,x_max, num=10)
ybins = np.linspace(y_min,y_max, num=10)
zbins = np.linspace(z_min,z_max, num=10)
arr_bins=[xbins,ybins,zbins]
bins = arr_bins[dim]
med = np.median(arr[:,ts,dim])
idx = (np.abs(bins - med)).argmin()
med = bins[idx]
ind_up = arr[:,ts,dim]>=med
ind_down = arr[:,ts,dim]<med
gt_up = arr[ind_up]
gt_down = arr[ind_down]
harr_up_gt = []
for ts in range(gt_up.shape[1]):
histo_up, _ = np.histogramdd(gt_up[:,ts,:], bins=[xbins,ybins,zbins])
harr_up_gt.append(histo_up)
harr_down_gt = []
for ts in range(gt_down.shape[1]):
histo_down, _ = np.histogramdd(gt_down[:,ts,:], bins=[xbins,ybins,zbins])
harr_down_gt.append(histo_down)
return harr_down_gt, harr_up_gt
def get_entropy_series(histos1, histos2):
entr_ser = []
for ts in range(len(histos1)):
n_tot = np.sum(histos1[ts])+np.sum(histos2[ts])
if n_tot == 0:
entr_ser.append(0)
continue
entr = 0
for hbin1,hbin2 in zip(histos1[ts].flatten(),histos2[ts].flatten()):
if hbin1==0 or hbin2==0:
continue
fr1 = hbin1/(hbin1+hbin2)
fr2 = hbin2/(hbin1+hbin2)
entr += (hbin1+hbin2)/n_tot * -(fr1*np.log(fr1) + fr2*np.log(fr2))
entr_ser.append(entr)
return entr_ser
arr_up,arr_down = get_entropy_histos(pred,dim=0)
entr_ser_pred_x = get_entropy_series(arr_up,arr_down)
arr_up,arr_down = get_entropy_histos(pred)
entr_ser_pred_z = get_entropy_series(arr_up,arr_down)
arr_up,arr_down = get_entropy_histos(gt,dim=0)
entr_ser_gt_x = get_entropy_series(arr_up,arr_down)
arr_up,arr_down = get_entropy_histos(gt)
entr_ser_gt_z = get_entropy_series(arr_up,arr_down)
entr_list_x=[]
entr_list_z=[]
for run in runs_gt:
arr_up, arr_down = get_entropy_histos(run,dim=0)
entr_list_x.append(get_entropy_series(arr_up,arr_down))
arr_up, arr_down = get_entropy_histos(run)
entr_list_z.append(get_entropy_series(arr_up,arr_down))
entr_rms_x= np.std(np.array(entr_list_x),axis=0)
entr_rms_z= np.std(np.array(entr_list_z),axis=0)
matplotlib.rc('font',**{'family':'serif','serif':['Times'], 'size' : 70})
axlsize=130
matplotlib.rc('axes', labelsize=axlsize)
matplotlib.rc('legend', fontsize=100)
plt.figure(figsize=(20,20))
plt.subplots_adjust(bottom=0.20, top=0.95, left=0.25, right=0.95)
plt.grid(b=True, which='major', linewidth='1', linestyle='--', color='grey', dashes=(2,40,2,40))
plt.plot(np.linspace(1,235600,len(entr_ser_gt_z[::30])),entr_ser_pred_z[::30],'r--', dashes=(4, 4), linewidth=7.0)
plt.plot(np.linspace(1,235600,len(entr_ser_gt_z[::30])),entr_ser_gt_z[::30],'r-', linewidth=5.0)
plt.fill_between(np.linspace(1,235600,len(entr_ser_gt_z[::30])),entr_ser_gt_z[::30] - entr_rms_z[::30], entr_ser_gt_z[::30] + entr_rms_z[::30], color='r', alpha=0.2, linewidth=0.0)
plt.plot(np.linspace(1,235600,len(entr_ser_gt_z[::30])),entr_ser_pred_x[::30],'b--', dashes=(4, 4), linewidth=7.0)
plt.plot(np.linspace(1,235600,len(entr_ser_gt_z[::30])),entr_ser_gt_x[::30],'b-', linewidth=5.0)
plt.fill_between(np.linspace(1,235600,len(entr_ser_gt_z[::30])),entr_ser_gt_x[::30] - entr_rms_x[::30], entr_ser_gt_x[::30] + entr_rms_x[::30], color='b', alpha=0.2, linewidth=0.0)
plt.axes().yaxis.set_ticks([0,0.10,0.20,0.30])
inarr=np.linspace(1,235600,len(entr_ser_gt_z[::30]))
myarr=[inarr.min()+0.1*inarr.ptp(), inarr.max()-0.1*inarr.ptp()]
nr1=10**(np.ceil(np.log10(np.abs(myarr[0]))))*1
nr2=10**(np.ceil(np.log10(np.abs(myarr[1]))))*1
lsp=np.linspace(np.round(myarr[0]/min(nr1,nr2), 1)*min(nr1,nr2), np.round(myarr[1]/min(nr1,nr2), 1)*min(nr1,nr2), 3)
plt.axes().xaxis.set_ticks(lsp)
myarr=[inarr.min()-0.0*inarr.ptp(), inarr.max()+0.0*inarr.ptp()]
plt.xlim([myarr[0], myarr[1]])
inarr=np.concatenate([entr_ser_pred_z[::30], entr_ser_gt_z[::30], entr_ser_pred_x[::30], entr_ser_gt_x[::30]])
myarr=[inarr.min()+0.1*inarr.ptp(), inarr.max()-0.1*inarr.ptp()]
nr1=10**(np.ceil(np.log10(np.abs(myarr[0]))))*10
nr2=10**(np.ceil(np.log10(np.abs(myarr[1]))))*10
lsp=np.linspace(np.round(myarr[0]/min(nr1,nr2), 1)*min(nr1,nr2), np.round(myarr[1]/min(nr1,nr2), 1)*min(nr1,nr2), 3)
plt.axes().yaxis.set_ticks(lsp)
myarr=[inarr.min()-0.1*inarr.ptp(), inarr.max()+0.1*inarr.ptp()]
plt.ylim([myarr[0]-0.35*(myarr[1]-myarr[0]), myarr[1]])
plt.tick_params(axis='both', which='major', pad=15)
plt.axes().ticklabel_format(style='sci',axis='x',scilimits=(0,0))
ybox6=TextArea(r".\hspace*{50cm}\phantom{Mixing S $x$/}$z$\hspace*{50cm}.", textprops=dict(color="r",size=axlsize,rotation=90,ha='left',va='center'))
ybox5=TextArea(r".\hspace*{50cm}\phantom{Mixing S $x$}/\phantom{$z$}\hspace*{50cm}.", textprops=dict(color="black", size=axlsize,rotation=90,ha='left',va='center'))
ybox2=TextArea(r".\hspace*{50cm}\phantom{Mixing S }$x$\phantom{/$z$}\hspace*{50cm}.", textprops=dict(color="b",size=axlsize,rotation=90,ha='left',va='center'))
ybox1=TextArea(r".\hspace*{50cm}Mixing S \phantom{$x$/$z$}\hspace*{50cm}.", textprops=dict(color="black", size=axlsize,rotation=90,ha='left',va='center'))
xbox1=TextArea(r"Time", textprops=dict(color="black", size=axlsize, ha='center', va='top'))
anchored_ybox6=AnchoredOffsetbox(loc='lower left', child=ybox6, pad=0.0, frameon=False, bbox_to_anchor=(-0.30, 0.5), bbox_transform=plt.axes().transAxes, borderpad=0.)
anchored_ybox5=AnchoredOffsetbox(loc='lower left', child=ybox5, pad=0.0, frameon=False, bbox_to_anchor=(-0.30, 0.5), bbox_transform=plt.axes().transAxes, borderpad=0.)
anchored_ybox2=AnchoredOffsetbox(loc='lower left', child=ybox2, pad=0.0, frameon=False, bbox_to_anchor=(-0.30, 0.5), bbox_transform=plt.axes().transAxes, borderpad=0.)
anchored_ybox1=AnchoredOffsetbox(loc='lower left', child=ybox1, pad=0.0, frameon=False, bbox_to_anchor=(-0.30, 0.5), bbox_transform=plt.axes().transAxes, borderpad=0.)
anchored_xbox1=AnchoredOffsetbox(loc='lower left', child=xbox1, pad=0.0, frameon=False, bbox_to_anchor=(0.5, -0.15), bbox_transform=plt.axes().transAxes, borderpad=0.)
plt.axes().add_artist(anchored_ybox6)
plt.axes().add_artist(anchored_ybox5)
plt.axes().add_artist(anchored_ybox2)
plt.axes().add_artist(anchored_ybox1)
plt.axes().add_artist(anchored_xbox1)
legend_elements=[plt.Line2D([0], [0], color='black', linestyle='--', dashes=(4, 6), linewidth=7.0, label='Prediction'),
plt.Line2D([0], [0], color='black', linewidth=5.0, label='Ground Truth')]
plt.legend(handles=legend_elements, loc='lower center', bbox_to_anchor=(0.5, -0.05), ncol=1, fancybox=False, shadow=False, frameon=False, handletextpad=0.5, labelspacing=0.0, handlelength=2.25, columnspacing=10, bbox_transform=plt.axes().transAxes)
plt.savefig(os.path.join(plotDir, "noncoh_entropy"+runid+".png"))
plt.savefig(os.path.join(plotDir, "noncoh_entropy"+runid+".pdf"))
plt.close()
im = Image.open(os.path.join(plotDir, "noncoh_entropy"+runid+".png"))
im