test_rdists.py
8.59 KB
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
from __future__ import print_function
from __future__ import division
from past.utils import old_div
from collections import defaultdict
import unittest
import numpy as np
import numpy.testing as npt
from hyperopt.rdists import (
loguniform_gen,
lognorm_gen,
quniform_gen,
qloguniform_gen,
qnormal_gen,
qlognormal_gen,
)
from scipy import stats
from scipy.stats.tests.test_continuous_basic import (
check_cdf_logcdf,
check_pdf_logpdf,
check_pdf,
check_cdf_ppf,
)
class TestLogUniform(unittest.TestCase):
def test_cdf_logcdf(self):
check_cdf_logcdf(loguniform_gen(0, 1), (0, 1), "")
check_cdf_logcdf(loguniform_gen(0, 1), (-5, 5), "")
def test_cdf_ppf(self):
check_cdf_ppf(loguniform_gen(0, 1), (0, 1), "")
check_cdf_ppf(loguniform_gen(-2, 1), (-5, 5), "")
def test_pdf_logpdf(self):
check_pdf_logpdf(loguniform_gen(0, 1), (0, 1), "")
check_pdf_logpdf(loguniform_gen(low=-4, high=-0.5), (-2, 1), "")
def test_pdf(self):
check_pdf(loguniform_gen(0, 1), (0, 1), "")
check_pdf(loguniform_gen(low=-4, high=-2), (-3, 2), "")
def test_distribution_rvs(self):
alpha = 0.01
loc = 0
scale = 1
arg = (loc, scale)
distfn = loguniform_gen(0, 1)
D, pval = stats.kstest(distfn.rvs, distfn.cdf, args=arg, N=1000)
if pval < alpha:
npt.assert_(
pval > alpha,
"D = %f; pval = %f; alpha = %f; args=%s" % (D, pval, alpha, arg),
)
class TestLogNormal(unittest.TestCase):
def test_cdf_logcdf(self):
check_cdf_logcdf(lognorm_gen(0, 1), (), "")
check_cdf_logcdf(lognorm_gen(0, 1), (), "")
def test_cdf_ppf(self):
check_cdf_ppf(lognorm_gen(0, 1), (), "")
check_cdf_ppf(lognorm_gen(-2, 1), (), "")
def test_pdf_logpdf(self):
check_pdf_logpdf(lognorm_gen(0, 1), args=(), msg="base case")
check_pdf_logpdf(
lognorm_gen(mu=-4, sigma=0.5), args=(), msg="non-default mu, sigma"
)
def test_pdf(self):
check_pdf(lognorm_gen(0, 1), (), "")
check_pdf(lognorm_gen(mu=-4, sigma=2), (), "")
def test_distribution_rvs(self):
import warnings
warnings.warn("test_distribution_rvs is being skipped!")
return # XXX
alpha = 0.01
loc = 0
scale = 1
arg = (loc, scale)
distfn = lognorm_gen(0, 1)
D, pval = stats.kstest(distfn.rvs, distfn.cdf, args=arg, N=1000)
if pval < alpha:
npt.assert_(
pval > alpha,
"D = %f; pval = %f; alpha = %f; args=%s" % (D, pval, alpha, arg),
)
def check_d_samples(dfn, n, rtol=1e-2, atol=1e-2):
counts = defaultdict(lambda: 0)
# print 'sample', dfn.rvs(size=n)
inc = old_div(1.0, n)
for s in dfn.rvs(size=n):
counts[s] += inc
for ii, p in sorted(counts.items()):
t = np.allclose(dfn.pmf(ii), p, rtol=rtol, atol=atol)
if not t:
print(("Error in sampling frequencies", ii))
print("value\tpmf\tfreq")
for jj in sorted(counts):
print(("%.2f\t%.3f\t%.4f" % (jj, dfn.pmf(jj), counts[jj])))
npt.assert_(t, "n = %i; pmf = %f; p = %f" % (n, dfn.pmf(ii), p))
class TestQUniform(unittest.TestCase):
def test_smallq(self):
low, high, q = (0, 1, 0.1)
qu = quniform_gen(low, high, q)
check_d_samples(qu, n=10000)
def test_bigq(self):
low, high, q = (-20, -1, 3)
qu = quniform_gen(low, high, q)
check_d_samples(qu, n=10000)
def test_offgrid_int(self):
qn = quniform_gen(0, 2, 2)
assert qn.pmf(0) > 0.0
assert qn.pmf(1) == 0.0
assert qn.pmf(2) > 0.0
assert qn.pmf(3) == 0.0
assert qn.pmf(-1) == 0.0
def test_offgrid_float(self):
qn = quniform_gen(0, 1, 0.2)
assert qn.pmf(0) > 0.0
assert qn.pmf(0.1) == 0.0
assert qn.pmf(0.2) > 0.0
assert qn.pmf(0.4) > 0.0
assert qn.pmf(0.8) > 0.0
assert qn.pmf(-0.2) == 0.0
assert qn.pmf(0.99) == 0.0
assert qn.pmf(-0.99) == 0.0
def test_output_type_int(self):
result = quniform_gen(0, 10, 1).rvs()
assert int == type(result)
def test_output_type_float(self):
assert float == type(quniform_gen(0, 10, 1.0).rvs())
class TestQLogUniform(unittest.TestCase):
def logp(self, x, low, high, q):
return qloguniform_gen(low, high, q).logpmf(x)
def test_smallq(self):
low, high, q = (0, 1, 0.1)
qlu = qloguniform_gen(low, high, q)
check_d_samples(qlu, n=10000)
def test_bigq(self):
low, high, q = (-20, 4, 3)
qlu = qloguniform_gen(low, high, q)
check_d_samples(qlu, n=10000)
def test_point(self):
low, high, q = (np.log(0.05), np.log(0.15), 0.5)
qlu = qloguniform_gen(low, high, q)
check_d_samples(qlu, n=10000)
def test_2points(self):
low, high, q = (np.log(0.05), np.log(0.75), 0.5)
qlu = qloguniform_gen(low, high, q)
check_d_samples(qlu, n=10000)
def test_point_logpmf(self):
assert np.allclose(self.logp(0, np.log(0.25), np.log(0.5), 1), 0.0)
def test_rounding_logpmf(self):
assert self.logp(0, np.log(0.25), np.log(0.75), 1) > self.logp(
1, np.log(0.25), np.log(0.75), 1
)
assert (
self.logp(-1, np.log(0.25), np.log(0.75), 1)
== self.logp(2, np.log(0.25), np.log(0.75), 1)
== -np.inf
)
def test_smallq_logpmf(self):
assert (
self.logp(0.2, np.log(0.16), np.log(0.55), 0.1)
> self.logp(0.3, np.log(0.16), np.log(0.55), 0.1)
> self.logp(0.4, np.log(0.16), np.log(0.55), 0.1)
> self.logp(0.5, np.log(0.16), np.log(0.55), 0.1)
> -10
)
assert (
self.logp(0.1, np.log(0.16), np.log(0.55), 1)
== self.logp(0.6, np.log(0.16), np.log(0.55), 1)
== -np.inf
)
def test_output_type_int(self):
result = qloguniform_gen(0, 10, 1).rvs()
assert int == type(result)
def test_output_type_float(self):
assert float == type(qloguniform_gen(0, 10, 1.0).rvs())
class TestQNormal(unittest.TestCase):
def test_smallq(self):
mu, sigma, q = (0, 1, 0.1)
qn = qnormal_gen(mu, sigma, q)
check_d_samples(qn, n=10000)
def test_bigq(self):
mu, sigma, q = (-20, 4, 3)
qn = qnormal_gen(mu, sigma, q)
check_d_samples(qn, n=10000)
def test_offgrid_int(self):
qn = qnormal_gen(0, 1, 2)
assert qn.pmf(0) > 0.0
assert qn.pmf(1) == 0.0
assert qn.pmf(2) > 0.0
def test_offgrid_float(self):
qn = qnormal_gen(0, 1, 0.2)
assert qn.pmf(0) > 0.0
assert qn.pmf(0.1) == 0.0
assert qn.pmf(0.2) > 0.0
assert qn.pmf(0.4) > 0.0
assert qn.pmf(-0.2) > 0.0
assert qn.pmf(-0.4) > 0.0
assert qn.pmf(0.99) == 0.0
assert qn.pmf(-0.99) == 0.0
def test_numeric(self):
qn = qnormal_gen(0, 1, 1)
assert qn.pmf(500) > -np.inf
def test_output_type_int(self):
result = qnormal_gen(0, 10, 1).rvs()
assert int == type(result)
def test_output_type_float(self):
assert float == type(qnormal_gen(0, 10, 1.0).rvs())
class TestQLogNormal(unittest.TestCase):
def test_smallq(self):
mu, sigma, q = (0, 1, 0.1)
qn = qlognormal_gen(mu, sigma, q)
check_d_samples(qn, n=10000)
def test_bigq(self):
mu, sigma, q = (-20, 4, 3)
qn = qlognormal_gen(mu, sigma, q)
check_d_samples(qn, n=10000)
def test_offgrid_int(self):
mu, sigma, q = (1, 2, 2)
qn = qlognormal_gen(mu, sigma, q)
assert qn.pmf(0) > qn.pmf(2) > qn.pmf(20) > 0
assert qn.pmf(1) == qn.pmf(2 - 0.001) == qn.pmf(-1) == 0
def test_offgrid_float(self):
mu, sigma, q = (-0.5, 2, 0.2)
qn = qlognormal_gen(mu, sigma, q)
assert qn.pmf(0) > qn.pmf(0.2) > qn.pmf(2) > 0
assert qn.pmf(0.1) == qn.pmf(0.2 - 0.001) == qn.pmf(-0.2) == 0
def test_numeric(self):
# XXX we don't have a numerically accurate computation for this guy
# qn = qlognormal_gen(0, 1, 1)
# assert -np.inf < qn.logpmf(1e-20) < -50
# assert -np.inf < qn.logpmf(1e20) < -50
pass
def test_output_type_int(self):
result = qlognormal_gen(0, 10, 1).rvs()
assert int == type(result)
def test_output_type_float(self):
assert float == type(qlognormal_gen(0, 10, 1.0).rvs())
# -- non-empty last line for flake8