/
test_likelihoodfns.py
91 lines (75 loc) · 3.71 KB
/
test_likelihoodfns.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
import numpy as _np
from pygsti import circuits
from pygsti.modelpacks.legacy import std1Q_XYI as std
from pygsti.tools import likelihoodfns as lfn
from . import fixtures as pkg
from ..util import BaseCase
class LikelihoodFunctionsBase(BaseCase):
def setUp(self):
self.ds = pkg.dataset.copy()
self.model = pkg.mdl_lsgst_go.copy()
#We Removed the forbidden_prob function
#class LikelihoodFunctionsTester(LikelihoodFunctionsBase):
# def test_forbidden_probablity(self):
# prob = lfn.forbidden_prob(std.target_model(), self.ds)
# self.assertAlmostEqual(prob, 1.276825378318927e-13)
class LogLTester(LikelihoodFunctionsBase):
def setUp(self):
super(LogLTester, self).setUp()
self.circuits = circuits.to_circuits([('Gx',), ('Gy',), ('Gx', 'Gx')])
def test_logl(self):
L1 = lfn.logl(self.model, self.ds, self.circuits,
prob_clip_interval=(-1e6, 1e6),
poisson_picture=True)
# Non-poisson-picture
L2 = lfn.logl(self.model, self.ds, self.circuits,
prob_clip_interval=(-1e6, 1e6),
poisson_picture=False)
# TODO assert correctness
def test_logl_jacobian(self):
dL1 = lfn.logl_jacobian(self.model, self.ds, self.circuits,
prob_clip_interval=(-1e6, 1e6), radius=1e-4,
poisson_picture=True)
dL2 = lfn.logl_jacobian(self.model, self.ds, self.circuits,
prob_clip_interval=(-1e6, 1e6), radius=1e-4,
poisson_picture=False)
# test None as mdl list
dL2b = lfn.logl_jacobian(self.model, self.ds, None,
prob_clip_interval=(-1e6, 1e6), radius=1e-4,
poisson_picture=False)
# TODO assert correctness
def test_logl_hessian(self):
# TODO optimize
hL1 = lfn.logl_hessian(self.model, self.ds, self.circuits,
prob_clip_interval=(-1e6, 1e6), radius=1e-4,
poisson_picture=True)
hL2 = lfn.logl_hessian(self.model, self.ds, self.circuits,
prob_clip_interval=(-1e6, 1e6), radius=1e-4,
poisson_picture=False)
# test None as mdl list
hL2b = lfn.logl_hessian(self.model, self.ds, None,
prob_clip_interval=(-1e6, 1e6), radius=1e-4,
poisson_picture=False)
# TODO assert correctness
def test_logl_max(self):
maxL1 = lfn.logl_max(self.model, self.ds, self.circuits, poisson_picture=True)
maxL2 = lfn.logl_max(self.model, self.ds, self.circuits, poisson_picture=False)
# TODO assert correctness
#Removed this function
#def test_cptp_penalty(self):
# lfn.cptp_penalty(self.model, include_spam_penalty=True)
# # TODO assert correctness
def test_two_delta_logl(self):
twoDelta1 = lfn.two_delta_logl_term(n=100, p=0.5, f=0.6, min_prob_clip=1e-6, poisson_picture=True)
twoDelta2 = lfn.two_delta_logl_term(n=100, p=0.5, f=0.6, min_prob_clip=1e-6, poisson_picture=False)
# TODO assert correctness
def test_no_gatestrings(self):
# TODO what edge case does this cover?
model = std.target_model()
L1 = lfn.logl(model, self.ds,
prob_clip_interval=(-1e6, 1e6),
poisson_picture=True)
self.assertAlmostEqual(L1, -21393568.52986, 2)
L2 = lfn.logl_max(model, self.ds)
L2 = _np.array(L2) # convert from LocalNumpyArray => ndarray so __round__ works (?)
self.assertAlmostEqual(L2, -14028782.1039, 2)