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
|
# -*- coding: utf-8 -*-
import itertools as it
import numpy as np
from . import diff
from . import utils
def _calc_dmetric(X, Z, metric, diff_op):
dmetric = np.zeros((3,) + metric.shape)
res = diff_op(metric, 3, X[0, 1] - X[0, 0])
dmetric[0] = diff_op(metric, 3, X[0, 1] - X[0, 0])
dmetric[2] = diff_op(metric, 2, Z[1, 0] - Z[0, 0])
dmetric[1, 0, 0] = 0.0
dmetric[1, 1, 1] = 0.0
dmetric[1, 2, 2] = 0.0
dmetric[1, 0, 1] = np.where(np.abs(X) > 1e-8, (metric[0, 0] - metric[1, 1]) / X, dmetric[0, 0, 0] - dmetric[0, 1, 1])
dmetric[1, 1, 0] = dmetric[1, 0, 1]
dmetric[1, 0, 2] = 0.0
dmetric[1, 2, 0] = 0.0
dmetric[1, 1, 2] = np.where(np.abs(X) > 1e-8, metric[0, 2] / X, dmetric[0, 0, 2])
dmetric[1, 2, 1] = dmetric[1, 1, 2]
return dmetric
def _calc_d2metric(X, Z, metric, dmetric, diff_op, diff2_op):
d2metric = np.empty((3,) + dmetric.shape)
dx = X[0, 1] - X[0, 0]
dz = Z[1, 0] - Z[0, 0]
X2 = X * X
d2metric[0, 0] = diff2_op(metric, 3, dx)
d2metric[2, 2] = diff2_op(metric, 2, dz)
d2metric[0, 2] = diff_op(dmetric[0], 2, dz)
d2metric[2, 0] = d2metric[0, 2]
d2metric[0, 1, 0, 0] = 0
d2metric[1, 0, 0, 0] = 0
d2metric[1, 2, 0, 0] = 0
d2metric[2, 1, 0, 0] = 0
d2metric[1, 1, 0, 0] = np.where(np.abs(X) > 1e-8, dmetric[0, 0, 0] / X - 2.0 * (metric[0, 0] - metric[1, 1]) / X2, d2metric[0, 0, 1, 1])
d2metric[1, 1, 1, 1] = np.where(np.abs(X) > 1e-8, dmetric[0, 1, 1] / X + 2.0 * (metric[0, 0] - metric[1, 1]) / X2, d2metric[0, 0, 0, 0])
d2metric[0, 1, 1, 1] = 0
d2metric[1, 0, 1, 1] = 0
d2metric[2, 1, 1, 1] = 0
d2metric[1, 2, 1, 1] = 0
d2metric[1, 1, 2, 2] = np.where(np.abs(X) > 1e-8, dmetric[0, 2, 2] / X, d2metric[0, 0, 2, 2])
d2metric[0, 1, 2, 2] = 0
d2metric[1, 0, 2, 2] = 0
d2metric[2, 1, 2, 2] = 0
d2metric[1, 2, 2, 2] = 0
d2metric[1, 1, 0, 1] = 0
d2metric[0, 1, 0, 1] = np.where(np.abs(X) > 1e-8, (dmetric[0, 0, 0] - dmetric[0, 1, 1]) / X - (metric[0, 0] - metric[1, 1]) / X2, 0.5 * (d2metric[0, 0, 0, 0] - d2metric[0, 0, 1, 1]))
d2metric[1, 0, 0, 1] = d2metric[0, 1, 0, 1]
d2metric[1, 2, 0, 1] = np.where(np.abs(X) > 1e-8, (dmetric[2, 0, 0] - dmetric[2, 1, 1]) / X, d2metric[0, 2, 0, 0] - d2metric[0, 2, 1, 1])
d2metric[2, 1, 0, 1] = d2metric[1, 2, 0, 1]
d2metric[1, 1, 0, 2] = np.where(np.abs(X) > 1e-8, dmetric[0, 0, 2] / X - metric[0, 2] / X2, 0.5 * d2metric[0, 0, 0, 2])
d2metric[0, 1, 0, 2] = 0
d2metric[1, 0, 0, 2] = 0
d2metric[1, 2, 0, 2] = 0
d2metric[2, 1, 0, 2] = 0
d2metric[1, 1, 1, 2] = 0
d2metric[0, 1, 1, 2] = np.where(np.abs(X) > 1e-8, dmetric[0, 0, 2] / X - metric[0, 2] / X2, 0.5 * d2metric[0, 0, 0, 2])
d2metric[1, 0, 1, 2] = d2metric[0, 1, 1, 2]
d2metric[1, 2, 1, 2] = np.where(np.abs(X) > 1e-8, dmetric[2, 0, 2] / X, d2metric[0, 2, 0, 2])
d2metric[2, 1, 1, 2] = d2metric[1, 2, 1, 2]
d2metric[:, :, 1, 0] = d2metric[:, :, 0, 1]
d2metric[:, :, 2, 0] = d2metric[:, :, 0, 2]
d2metric[:, :, 2, 1] = d2metric[:, :, 1, 2]
return d2metric
def _calc_christoffel(metric, metric_u, dmetric):
Gamma = np.empty_like(dmetric)
for i, j, k in it.product(range(3), repeat = 3):
Gamma[i, j, k] = 0.5 * np.einsum('k...,k...', metric_u[i], dmetric[j, k] + dmetric[k, j] - dmetric[:, k, j])
return Gamma
def calc_christoffel(x, z, metric, diff_op = diff.fd4):
"""
Calculate Christoffel symbols
i 1 il / \
Γ = -γ | ∂ γ + ∂ γ - ∂ γ |
jk 2 \ j kl k jl l jk /
using finite differences.
:param array_like x: 1D array of x coordinates.
:param array_like z: 1D array of z-coordinates.
:param array_like metric: 4D array of spatial metric values at the grid
formed by x and z. metric[i, j, k, l] is the ijth
component of the metric at the point (X=x[l],
Z=z[k]).
:rtype: array_like, shape (3, 3, 3, z.shape[0], x.shape[0])
:return: Christoffel symbols, first axis is the upper index, following two
axes are the two lower indices, final two axes correspond to the z
and x grid points respectively.
"""
X, Z = np.meshgrid(x, z)
metric_u = utils.matrix_invert(metric)
dmetric = _calc_dmetric(X, Z, metric, diff_op)
return _calc_christoffel(metric, metric_u, dmetric)
def _calc_dchristoffel(metric, metric_u, dmetric, dmetric_u, d2metric):
dGamma = np.empty_like(d2metric)
for i, j, k, l in it.product(range(3), repeat = 4):
dGamma[i, j, k, l] = 0.5 * (np.einsum('k...,k...', dmetric_u[i, j], dmetric[l, k] + dmetric[k, l] - dmetric[:, k, l]) +
np.einsum('k...,k...', metric_u[j], d2metric[i, l, k] + d2metric[i, k, l] - d2metric[i, :, l, k]))
return dGamma
def calc_riemann(metric, metric_u, dmetric, d2metric):
dmetric_u = -np.einsum('ij...,km...,ljk...->lim...', metric_u, metric_u, dmetric)
Gamma = _calc_christoffel(metric, metric_u, dmetric)
dGamma = _calc_dchristoffel(metric, metric_u, dmetric, dmetric_u, d2metric)
Riemann_uddd = (np.einsum('cabd...->abcd...', dGamma) - np.einsum('dabc...->abcd...', dGamma) +
np.einsum('akc...,kbd...->abcd...', Gamma, Gamma) - np.einsum('akd...,kbc...->abcd...', Gamma, Gamma))
Riemann = np.einsum('ik...,klmn...->ilmn...', metric, Riemann_uddd)
return Riemann
def calc_ricci(metric, metric_u, dmetric, d2metric):
dmetric_u = -np.einsum('ij...,km...,ljk...->lim...', metric_u, metric_u, dmetric)
Gamma = _calc_christoffel(metric, metric_u, dmetric)
dGamma = _calc_dchristoffel(metric, metric_u, dmetric, dmetric_u, d2metric)
Ricci = (np.einsum('mmjk...->jk...', dGamma) - np.einsum('kmjm...->jk...', dGamma) +
np.einsum('llm...,mjk...->jk...', Gamma, Gamma) - np.einsum('lkm...,mjl...->jk...', Gamma, Gamma))
return Ricci
def calc_constraint_ham(metric, metric_u, dmetric, d2metric, curv):
Ricci = calc_ricci(metric, metric_u, dmetric, d2metric)
R_scalar = np.einsum('ij...,ij...', metric_u, Ricci)
curv_ud = np.einsum('ik...,kj...->ij...', metric_u, curv)
curv_trace = curv_ud[0, 0] + curv_ud[1, 1] + curv_ud[2, 2]
k2 = np.einsum('ij...,ji...', curv_ud, curv_ud)
return R_scalar + curv_trace ** 2 - k2
def calc_constraint_mom(metric, metric_u, dmetric, curv_m, dcurv_m):
Gamma = _calc_christoffel(metric, metric_u, dmetric)
curv_trace = curv_m[0, 0] + curv_m[1, 1] + curv_m[2, 2]
M = np.empty((3,) + curv_trace.shape)
for j in range(3):
M[j] = np.einsum('ii...', dcurv_m[:, :, j]) + np.einsum('iik...,k...', Gamma, curv_m[:, j]) - np.einsum('ki...,ik...', Gamma[:, :, j], curv_m)
return M
|