-
Notifications
You must be signed in to change notification settings - Fork 3
/
tmfg_core.py
120 lines (91 loc) · 4.49 KB
/
tmfg_core.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
import copy
from itertools import combinations, chain
import pandas as pd
from numpy.linalg import inv
from utils import *
class TMFG:
def __init__(self, correlation_matrix):
self.W = correlation_matrix
self.original_W = copy.copy(correlation_matrix)
self.N = self.W.shape[1]
self.P = np.zeros((self.N, self.N))
self.max_clique_gains = np.zeros(((3 * self.N) - 6))
self.best_vertex = np.array([-1] * ((3 * self.N) - 6))
self.cliques = []
self.separators = []
self.triangles = []
self.vertex_list = None
self.peo = None
self.JS = None
def compute_TMFG(self):
self.cliques.append(list(max_clique(self.W)))
self.vertex_list = np.setdiff1d(range(self.N), self.cliques[0])
self.triangles.append(list(pd.Series(self.cliques[0])[[0, 1, 2]]))
self.triangles.append(list(pd.Series(self.cliques[0])[[0, 1, 3]]))
self.triangles.append(list(pd.Series(self.cliques[0])[[0, 2, 3]]))
self.triangles.append(list(pd.Series(self.cliques[0])[[1, 2, 3]]))
self.peo = copy.copy(self.cliques[0])
self.W = np.array(self.W)
self.W[np.diag_indices_from(self.W)] = 0
peo_combinations_list = []
for n in range(len(self.cliques[0]) + 1):
two_d_lists = len(list(combinations(self.cliques[0], n))[0])
if two_d_lists == 2:
peo_combinations_list += list(combinations(self.cliques[0], n))
for i in peo_combinations_list:
self.P[int(i[0]), int(i[1])] = self.W[int(i[0]), int(i[1])]
for t in range(0, 4):
index_max, max_element = get_best_gain(self.N, self.vertex_list, self.triangles[t], self.W, None)
self.max_clique_gains[t] = max_element
self.best_vertex[t] = index_max
for u in range(0, (self.N - 4)):
nt = np.argmax(self.max_clique_gains)
nv = self.best_vertex[nt]
self.peo.append(nv)
thetraedron = [nv] + self.triangles[nt]
self.cliques.append(thetraedron)
newsep = self.triangles[nt]
peo_combinations_list = []
thetraedron_tbc = [nv] + newsep
for n in range(len(thetraedron_tbc) + 1):
two_d_lists = len(list(combinations(thetraedron_tbc, n))[0])
if two_d_lists == 2:
peo_combinations_list += list(combinations(thetraedron_tbc, n))
for i in peo_combinations_list:
self.P[int(i[0]), int(i[1])] = self.W[int(i[0]), int(i[1])]
self.separators.append(newsep)
self.triangles[nt] = [newsep[0], newsep[1], nv]
self.triangles.append([newsep[0], newsep[2], nv])
self.triangles.append([newsep[1], newsep[2], nv])
self.vertex_list = np.setdiff1d(self.vertex_list, nv)
no_vertex_list = np.setdiff1d(range(self.N), self.vertex_list)
if len(self.vertex_list) > 0:
indices_of_interest = np.argwhere(self.best_vertex == nv)
indices_of_interest = list(chain(*indices_of_interest))
for t in indices_of_interest:
index_max, max_element = get_best_gain(self.N, self.vertex_list, self.triangles[t], self.W, no_vertex_list)
self.max_clique_gains[t] = max_element
self.best_vertex[t] = index_max
self.max_clique_gains[nt] = 0
ct = len(self.triangles) - 1
if len(self.vertex_list) > 0:
for t in [nt, (ct - 1), ct]:
index_max, max_element = get_best_gain(self.N, self.vertex_list, self.triangles[t], self.W, no_vertex_list)
self.max_clique_gains[t] = max_element
self.best_vertex[t] = index_max
#self.__logo()
self.__unweighted_tmfg()
return self.cliques, self.separators, self.JS
def __unweighted_tmfg(self):
self.JS = np.zeros((self.original_W.shape[0], self.original_W.shape[0]))
for c in self.cliques:
self.JS[np.ix_(c, c)] = 1
np.fill_diagonal(self.JS, 0)
def __logo(self):
self.JS = np.zeros((self.original_W.shape[0], self.original_W.shape[0]))
W = self.original_W.to_numpy()
for c in self.cliques:
self.JS[np.ix_(c, c)] = self.JS[np.ix_(c, c)] + inv(W[np.ix_(c, c)])
for s in self.separators:
self.JS[np.ix_(s, s)] = self.JS[np.ix_(s, s)] - inv(W[np.ix_(s, s)])
np.fill_diagonal(self.JS, 0)