forked from alexander-jaeger-2101/DC-filtrations-homology
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathInhomogeneousFiltration.py
320 lines (289 loc) · 15.5 KB
/
InhomogeneousFiltration.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
# Import Modules
import numpy as np
import pandas as pd
import scipy as sp
import dionysus as d
import matplotlib.pyplot as plt
#import pecan as pc
import sys
import os
#from __future__ import print_function
from ipywidgets import interact, interactive, fixed, interact_manual
import ipywidgets as widgets
from sklearn.metrics.pairwise import euclidean_distances
import networkx as nx
import matplotlib.pyplot as plt
import argparse
import sys
import matplotlib.collections
import matplotlib.lines
import matplotlib.animation as animation
import matplotlib.pyplot as plt
import gudhi as gd
import time
from basic_functions import *
from visualizations import *
class OperatorPowerFiltrationInhomogeneous:
"""
Class to generate a filtration of Rips complex over the power tau of a diffusion operator
"""
def __init__(self,
experiment_path,
experiment_name,
mode,
kernel,
e,
b,
t):
"""
Args:
experiment_path [string]: path to the pecan .npz file.
experiment_name [string]: name of the pecan .npz file .
mode [string]: restng / non_resting.
kernel [string]: 'gaussian_aniso' / 'gaussian' (gaussian)
e [float]: kernel bandwith
b [float]: gaussian_aniso exp
t [int]: diffusion condensation time
"""
# load diffused data sets
self.data=np.load(experiment_path+experiment_name+'.npz')
parsed_keys = parse_keys(self.data)
assert 'data' in parsed_keys
X = make_tensor(self.data, parsed_keys['data'])
T = X.shape[-1]
N=X.shape[0]
name_t=str('data_t_'+str(t))
self.X_t=self.data[name_t]
# create diffusion operator
if mode=='resting':
K=kernel_gaussian_resting(self.X_t,e)
if mode=='non_resting':
K=kernel_gaussian_nonresting(self.X_t,e)
if kernel== 'gaussian_aniso':
K_=kernel_gaussian_aniso_norm(K,b)
K=K_
if kernel== 'gaussian':
K=K
self.P_t=diffusion_operator(K)
self.N=len(self.P_t)
self.experiment_path=experiment_path
self.experiment_name=experiment_name
self.mode=mode
self.kernel=kernel
self.e=e
self.b=b
self.t=t
def compute_edge_weights(self, tau_max):
"""
Function to compute all N_E=(N c 2) edge weights for all steps tau of the filtration. Result stored in
tau_max x N_E matrix W. Order of elements in a row of W according to upper triangle matrix of P_t. Edge weights
are mean of both corresponding elements of P_t^tau: 0.5*(P_t[i][j]+P_t[j][i])
Args:
tau_max [int]: filtration from tau=1 to tau=tau_max
"""
P_t_tau=self.P_t
for tau in range(0,tau_max):
if tau ==0:
P_t_tau=self.P_t
P_t_tau_flat=upper_triu(self.P_t).flatten()
P_t_tau_flat_nonzero=P_t_tau_flat[P_t_tau_flat>0]
N_E=len(P_t_tau_flat_nonzero)
W_tensor=np.zeros((tau_max,N_E))
W_tensor[0]=P_t_tau_flat_nonzero
else:
name=str('data_t_'+str(tau))
X_t=self.data[name]
if self.mode=='resting':
K=kernel_gaussian_resting(X_t,self.e)
if self.mode=='non_resting':
K=kernel_gaussian_nonresting(X_t,self.e)
if self.kernel== 'gaussian_aniso':
K_=kernel_gaussian_aniso_norm(K,self.b)
K=K_
P_t_=diffusion_operator(K)
P_t_tau=P_t_tau@P_t_
P_t_tau_mean=0.5*(P_t_tau+P_t_tau.T)
P_t_tau=P_t_tau_mean
P_t_tau_flat=upper_triu(P_t_tau).flatten()
P_t_tau_flat_nonzero=P_t_tau_flat[P_t_tau_flat>0]
W_tensor[tau]=P_t_tau_flat_nonzero
self.W_tensor=W_tensor
return self.W_tensor
def compute_filtration(self,weight_threshold,dist_threshold,tau_max, max_dimension,plot_embedding='fix',merge=False,show_fig=True,save_fig=True,save_path=''):
"""
Args:
tau_max [int]: filtration from tau=1 to tau=tau_max
max_dimension (int): graph expansion until this given dimension.
threshold (float): minimum edge weight
dist_threshold (float): distance merge threshold
merge : False - do not merge data points closter than weight_threshold / True - do not merge data points closter than weight_threshold
plot_embedding (string): plot_X_t (string): visualization of data: fix - at the fix diffsion time t choosen above / var - time variable
"""
print(plot_embedding)
P_t_tau=self.P_t
betti_tensor=np.zeros(3)
simplices_tensor=np.zeros(3)
simplex_list=[]
time_list=[]
for f in range(0,tau_max):
if f==0:
P_t_tau=self.P_t
name=str('data_t_'+str(f))
X_0=self.data[name]
X_t=X_0
else:
name=str('data_t_'+str(f))
X_t=self.data[name]
if self.mode=='resting':
K=kernel_gaussian_resting(X_t,self.e)
if self.mode=='non_resting':
K=kernel_gaussian_nonresting(X_t,self.e)
if self.kernel== 'gaussian_aniso':
K_=kernel_gaussian_aniso_norm(K,self.b)
K=K_
P_t_=diffusion_operator(K)
P_t_tau= P_t_tau@P_t_
P_t_tau_mean=0.5*(P_t_tau+P_t_tau.T)
P_t_tau=P_t_tau_mean
if merge==True:
dist=sp.spatial.distance.cdist(X_t,X_t)
tf_dist=dist<=dist_threshold
tf_weight=P_t_tau>=weight_threshold
tf=np.logical_or(tf_dist,tf_weight)
if merge == False:
tf=P_t_tau>=weight_threshold
A_list=(np.where(tf))
A_bol_1=np.zeros((self.N,self.N))
for v in range(0,len(A_list[0])):
A_bol_1[A_list[0][v]][A_list[1][v]]=1
np.fill_diagonal(A_bol_1,0)
G_f = nx.from_numpy_matrix(A_bol_1)
T_f=find_triangles(A_bol_1)
G_f_list=list(G_f.edges())
st = gd.SimplexTree()
for r in range(0,self.N):
st.insert([r],0) #-1
for u in range(0,len(G_f_list)):
G_u=G_f_list[u]
st.insert([G_u[0],G_u[1]])
st.expansion(max_dimension)
if f==1:
st_gen = st.get_skeleton(2)
for splx in st_gen :
simplex_list.append(list(sorted(splx[0])))
time_list.append([1])
else:
simplex_t=[]
st_gen = st.get_skeleton(2)
for splx in st_gen :
simplex_t.append(list(sorted(splx[0])))
simplex_list,time_list=BarCodesUpdate(simplex_list,time_list,simplex_t,f)
BarCodes=st.persistence(min_persistence=-1,persistence_dim_max=True)
betti_f=(st.betti_numbers())
if len (betti_f)==3:
betti_tensor=np.vstack((betti_tensor,betti_f))
if len (betti_f)==2:
betti_tensor=np.vstack((betti_tensor,np.array([betti_f[0],betti_f[1],0])))
if len (betti_f)==1:
betti_tensor=np.vstack((betti_tensor,np.array([betti_f[0],0,0])))
simplices_tensor=np.vstack((simplices_tensor,np.array([len(G_f.nodes()),len(G_f.edges()),len(T_f)])))
print('τ=',f)
if show_fig == True:
if plot_embedding=='fix':
X_plot=self.X_t
if plot_embedding=='var':
name=str('data_t_'+str(f))
X_plot=self.data[name]
plt.figure(figsize=(8,8))
plt.xlim(-1.5,1.5)
plt.ylim(-1.5,1.5)
for z in range(0,len(T_f),1):
plt.fill(*X_plot[get_tf(T_f[z],self.N)].T,alpha=0.1,c='r')
nx.draw(G_f, X_plot,node_size=15,width=0.5,edge_color='k',node_color='k')
if save_fig == True:
if not os.path.exists(save_path+"/"+self.experiment_name):
os.mkdir(save_path+"/"+self.experiment_name)
if not os.path.exists(save_path+"/"+self.experiment_name+"/inhomogeneous/"):
os.mkdir(save_path+"/"+self.experiment_name+"/inhomogeneous/")
if not os.path.exists(save_path+"/"+self.experiment_name+"/inhomogeneous/simplicial_complex/"):
os.mkdir(save_path+"/"+self.experiment_name+"/inhomogeneous/simplicial_complex/")
if not os.path.exists(save_path+"/"+self.experiment_name+"/inhomogeneous/simplicial_complex/"+str(weight_threshold).replace(".", "_")):
os.mkdir(save_path+"/"+self.experiment_name+"/inhomogeneous/simplicial_complex/"+str(weight_threshold).replace(".", "_"))
if merge == True:
plt.savefig(save_path+"/"+self.experiment_name+"/inhomogeneous/simplicial_complex/"+str(weight_threshold).replace(".", "_")+'/simplicialcomplex_inhomo_'+self.experiment_name+'_e_'+str(self.e).replace(".", "_")+'_b_'+str(self.b).replace(".", "_")+'_kernel_'+str(self.kernel)+'_mode_'+str(self.mode)+'_t_'+str(self.t)+'_thres_'+str(weight_threshold).replace(".", "_")+'_tau_'+str(f+1)+'_merge_'+str(merge)+'_'+str(dist_threshold).replace(".", "_")+'_embedding_'+plot_embedding+'.png')
if merge == False:
plt.savefig(save_path+"/"+self.experiment_name+"/inhomogeneous/simplicial_complex/"+str(weight_threshold).replace(".", "_")+'/simplicialcomplex_inhomo_'+self.experiment_name+'_e_'+str(self.e).replace(".", "_")+'_b_'+str(self.b).replace(".", "_")+'_kernel_'+str(self.kernel)+'_mode_'+str(self.mode)+'_t_'+str(self.t)+'_thres_'+str(weight_threshold).replace(".", "_")+'_tau_'+str(f+1)+'_merge_'+str(merge)+'_embedding_'+plot_embedding+'.png')
plt.show()
if len(betti_f) >1:
print( 'β0='+str(betti_f[0]),'' ,'β1='+str(betti_f[1]))
if len(betti_f) ==1:
print( 'β0='+str(betti_f[0]))
print()
f = d.Filtration(simplex_list)
zz, dgms, cells = d.zigzag_homology_persistence(f, time_list)
BarCodes_=[]
for i,dgm in enumerate(dgms):
for p in dgm:
if i <2:
BarCodes_.append((i,(p.birth,p.death)))
return betti_tensor[1:], simplices_tensor[1:],BarCodes_
def plot_betti(self,betti_tensor,merge,weight_threshold,dist_threshold,scale='linear',show_fig=True,save_fig=True,save_path=''):
plt.figure(figsize=(10,5))
for z in range(0,len(betti_tensor[0])):
if z==0:
plt.plot(betti_tensor[:,0],c='r',label=r'$\beta_{0}$')
plt.plot(betti_tensor[:,1],c='g',label=r'$\beta_{1}$')
else:
plt.plot(betti_tensor[:,0],c='r')
plt.plot(betti_tensor[:,1],c='g')
plt.grid()
plt.yscale(scale)
plt.legend()
plt.xlabel(r'$\tau$', fontsize=20)
plt.ylabel('#', fontsize=20)
if save_fig==True:
if not os.path.exists(save_path+"/"+self.experiment_name):
os.mkdir(save_path+"/"+self.experiment_name)
if not os.path.exists(save_path+"/"+self.experiment_name+"/inhomogeneous/"):
os.mkdir(save_path+"/"+self.experiment_name+"/inhomogeneous/")
if not os.path.exists(save_path+"/"+self.experiment_name+"/inhomogeneous/features/"):
os.mkdir(save_path+"/"+self.experiment_name+"/inhomogeneous/features/")
if not os.path.exists(save_path+"/"+self.experiment_name+"/inhomogeneous/features/betti_numbers/"):
os.mkdir(save_path+"/"+self.experiment_name+"/inhomogeneous/features/betti_numbers/")
if merge == True:
plt.savefig(save_path+"/"+self.experiment_name+"/inhomogeneous/features/betti_numbers/"+'/bettinumbers_inhomo_'+self.experiment_name+'_e_'+str(self.e).replace(".", "_")+'_b_'+str(self.b).replace(".", "_")+'_kernel_'+str(self.kernel)+'_mode_'+str(self.mode)+'_t_'+str(self.t)+'_thres_'+str(weight_threshold).replace(".", "_")+'_merge_'+str(merge)+'_'+str(dist_threshold).replace(".", "_")+'_'+scale+'.png')
if merge == False:
plt.savefig(save_path+"/"+self.experiment_name+"/inhomogeneous/features/betti_numbers/"+'/bettinumbers_inhomo_'+self.experiment_name+'_e_'+str(self.e).replace(".", "_")+'_b_'+str(self.b).replace(".", "_")+'_kernel_'+str(self.kernel)+'_mode_'+str(self.mode)+'_t_'+str(self.t)+'_thres_'+str(weight_threshold).replace(".", "_")+'_merge_'+str(merge)+'_'+scale+'.png')
if show_fig==True:
plt.show()
def plot_num_simplices(self,simplices_tensor,merge,weight_threshold,dist_threshold,scale='log',show_fig=True,save_fig=True,save_path=''):
plt.figure(figsize=(10,5))
for z in range(0,len(simplices_tensor[0])):
if z==0:
plt.plot(simplices_tensor[:,0],c='r',label='0-simplices')
plt.plot(simplices_tensor[:,1],c='g',label='1-simplices')
plt.plot(simplices_tensor[:,2],c='b',label='2-simplices')
else:
plt.plot(simplices_tensor[:,0],c='r')
plt.plot(simplices_tensor[:,1],c='g')
plt.plot(simplices_tensor[:,2],c='b')
plt.grid()
plt.yscale(scale)
plt.legend()
plt.xlabel(r'$\tau$', fontsize=20)
plt.ylabel('#', fontsize=20)
if save_fig==True:
if not os.path.exists(save_path+"/"+self.experiment_name):
os.mkdir(save_path+"/"+self.experiment_name)
if not os.path.exists(save_path+"/"+self.experiment_name+"/inhomogeneous/"):
os.mkdir(save_path+"/"+self.experiment_name+"/inhomogeneous/")
if not os.path.exists(save_path+"/"+self.experiment_name+"/inhomogeneous/features/"):
os.mkdir(save_path+"/"+self.experiment_name+"/inhomogeneous/features/")
if not os.path.exists(save_path+"/"+self.experiment_name+"/inhomogeneous/features/num_simplices/"):
os.mkdir(save_path+"/"+self.experiment_name+"/inhomogeneous/features/num_simplices/")
if merge == True:
plt.savefig(save_path+"/"+self.experiment_name+"/inhomogeneous/features/num_simplices/"+'/numsimplices_inhomo_'+self.experiment_name+'_e_'+str(self.e).replace(".", "_")+'_b_'+str(self.b).replace(".", "_")+'_kernel_'+str(self.kernel)+'_mode_'+str(self.mode)+'_t_'+str(self.t)+'_thres_'+str(weight_threshold).replace(".", "_")+'_merge_'+str(merge)+'_'+str(dist_threshold).replace(".", "_")+'_'+scale+'.png')
if merge == False:
plt.savefig(save_path+"/"+self.experiment_name+"/inhomogeneous/features/num_simplices/"+'/numsimplices_inhomo_'+self.experiment_name+'_e_'+str(self.e).replace(".", "_")+'_b_'+str(self.b).replace(".", "_")+'_kernel_'+str(self.kernel)+'_mode_'+str(self.mode)+'_t_'+str(self.t)+'_thres_'+str(weight_threshold).replace(".", "_")+'_merge_'+str(merge)+'_'+scale+'.png')
if show_fig==True:
plt.show()