stats-arrays 0.7.dev8__tar.gz → 1.0__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/CHANGES.md +22 -9
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/PKG-INFO +8 -5
- stats_arrays-1.0/docs/distributions/beta_pert.rst +9 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/index.rst +5 -2
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/pyproject.toml +9 -6
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/__init__.py +11 -5
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/dataset_statistics.py +6 -5
- stats_arrays-1.0/src/stats_arrays/distributions/__init__.py +35 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/distributions/base.py +62 -28
- stats_arrays-1.0/src/stats_arrays/distributions/bernoulli.py +50 -0
- stats_arrays-1.0/src/stats_arrays/distributions/beta.py +144 -0
- stats_arrays-1.0/src/stats_arrays/distributions/beta_pert.py +124 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/distributions/discrete_uniform.py +22 -13
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/distributions/extreme.py +14 -6
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/distributions/gamma.py +14 -6
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/distributions/geometric.py +49 -29
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/distributions/lognormal.py +19 -11
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/distributions/normal.py +19 -11
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/distributions/student.py +14 -6
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/distributions/undefined.py +14 -6
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/distributions/weibull.py +14 -6
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/random.py +53 -30
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/uncertainty_choices.py +16 -10
- stats_arrays-1.0/src/stats_arrays/utils.py +113 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays.egg-info/PKG-INFO +8 -5
- stats_arrays-1.0/src/stats_arrays.egg-info/SOURCES.txt +55 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays.egg-info/requires.txt +2 -2
- stats_arrays-1.0/tests/test_base.py +274 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/tests/test_lhc.py +102 -8
- stats_arrays-1.0/tests/test_mc.py +193 -0
- stats_arrays-1.0/tests/test_rng.py +239 -0
- stats_arrays-0.7.dev8/stats_arrays/distributions/__init__.py +0 -30
- stats_arrays-0.7.dev8/stats_arrays/distributions/bernoulli.py +0 -45
- stats_arrays-0.7.dev8/stats_arrays/distributions/beta.py +0 -137
- stats_arrays-0.7.dev8/stats_arrays/utils.py +0 -47
- stats_arrays-0.7.dev8/stats_arrays.egg-info/SOURCES.txt +0 -53
- stats_arrays-0.7.dev8/tests/test_base.py +0 -127
- stats_arrays-0.7.dev8/tests/test_mc.py +0 -43
- stats_arrays-0.7.dev8/tests/test_rng.py +0 -89
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/CODE_OF_CONDUCT.md +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/CONTRIBUTING.md +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/LICENSE +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/MANIFEST.in +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/README.md +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/distributions/base.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/distributions/bernoulli.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/distributions/beta.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/distributions/discrete-uniform.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/distributions/extreme.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/distributions/gamma.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/distributions/lognormal.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/distributions/normal.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/distributions/student.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/distributions/triangular.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/distributions/uniform.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/distributions/weibull.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/lhc.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/mcrng.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/docs/rng.rst +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/setup.cfg +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays/errors.py +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays.egg-info/dependency_links.txt +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0/src}/stats_arrays.egg-info/top_level.txt +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/tests/test_choices.py +0 -0
- {stats_arrays-0.7.dev8 → stats_arrays-1.0}/tests/test_conversions.py +0 -0
|
@@ -1,43 +1,56 @@
|
|
|
1
1
|
# stats_arrays Changelog
|
|
2
2
|
|
|
3
|
-
# 0
|
|
3
|
+
# 1.0 (2025-10-27)
|
|
4
|
+
|
|
5
|
+
* Added types
|
|
6
|
+
* **BREAKING CHANGE**: Bernoulli distribution is now correctly strictly limited to outputting 0 or 1; `minimum` and `maximum` are ignored.
|
|
7
|
+
* Fixed triangular distribution `.pdf` method which didn't account for the minimum to maximum range but gave values for a `(0, 1)` range.
|
|
8
|
+
* Added many more tests
|
|
9
|
+
|
|
10
|
+
## 0.8 (2025-10-23)
|
|
11
|
+
|
|
12
|
+
* [#19: Add Beta PERT distribution](https://github.com/brightway-lca/stats_arrays/pull/19).
|
|
13
|
+
* Moved to a `src` layout.
|
|
14
|
+
* Switch to absolute imports
|
|
15
|
+
|
|
16
|
+
## 0.7 (2024-08-19)
|
|
4
17
|
|
|
5
18
|
* Moved to `pyproject.toml` packaging and `pytest` tests
|
|
6
19
|
* Numpy 2.0 compatiblity
|
|
7
20
|
|
|
8
|
-
|
|
21
|
+
### 0.6.6 (2023-10-18)
|
|
9
22
|
|
|
10
23
|
* Merged [PR #13](https://github.com/brightway-lca/stats_arrays/pull/13). Move argsort to save on execution time. Thanks @Loisel!
|
|
11
24
|
* Merged [PR #12](https://github.com/brightway-lca/stats_arrays/pull/12). Update links in readme. Thanks @mfastudillo!
|
|
12
25
|
* Merged [PR #10](https://github.com/brightway-lca/stats_arrays/pull/10). Fix flaky test on distributions::extreme.py, distributions::gama.py, and distributions::student.py. Thanks @lonly7star!
|
|
13
26
|
|
|
14
|
-
|
|
27
|
+
### 0.6.5 (2021-05-06)
|
|
15
28
|
|
|
16
29
|
* Updates for changes in Numpy API
|
|
17
30
|
|
|
18
|
-
|
|
31
|
+
### 0.6.4 (2020-01-31)
|
|
19
32
|
|
|
20
33
|
* Merged [PR #5](https://bitbucket.org/cmutel/stats_arrays/pull-requests/5/use-meanround-0-instead-of-round-mean/diff), fix rounding function.
|
|
21
34
|
|
|
22
|
-
|
|
35
|
+
### 0.6.3 (2019-11-29)
|
|
23
36
|
|
|
24
37
|
* Improve speed of lognormal CDF and PPF under common conditions
|
|
25
38
|
|
|
26
|
-
|
|
39
|
+
### 0.6.2 (2019-11-13)
|
|
27
40
|
|
|
28
41
|
* Restore Python 2.7 compatibility
|
|
29
42
|
|
|
30
|
-
|
|
43
|
+
### 0.6.1 (2019-07-29)
|
|
31
44
|
|
|
32
45
|
* Several bug fixes from Daniel de Koning
|
|
33
46
|
* Start improvement of beta distribution
|
|
34
47
|
|
|
35
|
-
|
|
48
|
+
## 0.6 (2019-05-19)
|
|
36
49
|
|
|
37
50
|
* Allow `MCRandomNumberGenerator` to generate multiple samples at once
|
|
38
51
|
* Move tests to pytest
|
|
39
52
|
* Add tests for MCRNG
|
|
40
53
|
|
|
41
|
-
|
|
54
|
+
### 0.5.1 (2019-05-11)
|
|
42
55
|
|
|
43
56
|
* Import from `collections.abc` when possible, for Python 3.8 support.
|
|
@@ -1,6 +1,6 @@
|
|
|
1
|
-
Metadata-Version: 2.
|
|
1
|
+
Metadata-Version: 2.4
|
|
2
2
|
Name: stats_arrays
|
|
3
|
-
Version: 0
|
|
3
|
+
Version: 1.0
|
|
4
4
|
Summary: Standard NumPy array interface for defining uncertain parameters
|
|
5
5
|
Author-email: Chris Mutel <cmutel@gmail.com>
|
|
6
6
|
Maintainer-email: Chris Mutel <cmutel@gmail.com>
|
|
@@ -14,14 +14,16 @@ Classifier: Programming Language :: Python :: 3.9
|
|
|
14
14
|
Classifier: Programming Language :: Python :: 3.10
|
|
15
15
|
Classifier: Programming Language :: Python :: 3.11
|
|
16
16
|
Classifier: Programming Language :: Python :: 3.12
|
|
17
|
+
Classifier: Programming Language :: Python :: 3.13
|
|
18
|
+
Classifier: Programming Language :: Python :: 3.14
|
|
17
19
|
Classifier: Natural Language :: English
|
|
18
20
|
Classifier: Operating System :: OS Independent
|
|
19
21
|
Classifier: Topic :: Scientific/Engineering
|
|
20
|
-
Requires-Python: >=3.
|
|
22
|
+
Requires-Python: >=3.9
|
|
21
23
|
Description-Content-Type: text/markdown
|
|
22
24
|
License-File: LICENSE
|
|
23
|
-
Requires-Dist: numpy
|
|
24
|
-
Requires-Dist: scipy
|
|
25
|
+
Requires-Dist: numpy
|
|
26
|
+
Requires-Dist: scipy
|
|
25
27
|
Provides-Extra: testing
|
|
26
28
|
Requires-Dist: stats_arrays; extra == "testing"
|
|
27
29
|
Requires-Dist: pytest; extra == "testing"
|
|
@@ -35,6 +37,7 @@ Requires-Dist: pytest; extra == "dev"
|
|
|
35
37
|
Requires-Dist: pytest-cov; extra == "dev"
|
|
36
38
|
Requires-Dist: pytest-randomly; extra == "dev"
|
|
37
39
|
Requires-Dist: setuptools; extra == "dev"
|
|
40
|
+
Dynamic: license-file
|
|
38
41
|
|
|
39
42
|
The `stats_arrays` package provides a standard NumPy array interface for defining uncertain parameters used in models, and classes for Monte Carlo sampling. It also plays well with others.
|
|
40
43
|
|
|
@@ -118,13 +118,14 @@ No uncertainty 1 **static value**
|
|
|
118
118
|
:ref:`normal` [#]_ 3 :math:`\boldsymbol{\mu}` :math:`\boldsymbol{\sigma}` *lower bound* *upper bound*
|
|
119
119
|
:ref:`uniform` [#]_ 4 *minimum* [#]_ **maximum**
|
|
120
120
|
:ref:`triangular` [#]_ 5 *mode* [#]_ *minimum* [#]_ **maximum**
|
|
121
|
-
:ref:`bernoulli` [#]_ 6 **p**
|
|
121
|
+
:ref:`bernoulli` [#]_ 6 **p**
|
|
122
122
|
:ref:`discreteu` [#]_ 7 *minimum* [#]_ **upper bound** [#]_
|
|
123
123
|
:ref:`weibull` [#]_ 8 *offset* [#]_ :math:`\boldsymbol{\lambda}` :math:`\boldsymbol{k}`
|
|
124
124
|
:ref:`gamma` [#]_ 9 *offset* [#]_ :math:`\boldsymbol{\theta}` :math:`\boldsymbol{k}`
|
|
125
125
|
:ref:`beta` [#]_ 10 :math:`\boldsymbol{\alpha}` :math:`\boldsymbol{\beta}` *lower bound* *upper bound*
|
|
126
126
|
:ref:`extreme` [#]_ 11 :math:`\boldsymbol{\mu}` :math:`\boldsymbol{\sigma}` :math:`\boldsymbol{\xi}`
|
|
127
127
|
:ref:`students` [#]_ 12 *median* *scale* :math:`\boldsymbol{\nu}`
|
|
128
|
+
:ref:`beta_pert` [#]_ 13 **mean** *lambda* **minimum** **maximum**
|
|
128
129
|
======================= === =========================== ============================= =========================== =============== ===============
|
|
129
130
|
|
|
130
131
|
Items in **bold** are required, items in *italics* are optional.
|
|
@@ -147,12 +148,13 @@ Items in **bold** are required, items in *italics* are optional.
|
|
|
147
148
|
.. [#] `Beta distribution <https://en.wikipedia.org/wiki/Beta_distribution>`_
|
|
148
149
|
.. [#] `Extreme value distribution <https://en.wikipedia.org/wiki/Generalized_extreme_value_distribution>`_
|
|
149
150
|
.. [#] `Student's T distribution <https://en.wikipedia.org/wiki/Student%27s_t-distribution>`_
|
|
151
|
+
.. [#] `Beta PERT distribution <https://en.wikipedia.org/wiki/PERT_distribution>`_
|
|
150
152
|
|
|
151
153
|
Unused columns can be given any value, but it is recommended that they are set to ``np.nan``.
|
|
152
154
|
|
|
153
155
|
.. warning::
|
|
154
156
|
|
|
155
|
-
|
|
157
|
+
Optional columns without data **must** be set to ``np.nan`` to avoid unexpected behaviour!
|
|
156
158
|
|
|
157
159
|
Extending parameter arrays
|
|
158
160
|
--------------------------
|
|
@@ -180,6 +182,7 @@ Probability distributions
|
|
|
180
182
|
distributions/student
|
|
181
183
|
distributions/gamma
|
|
182
184
|
distributions/weibull
|
|
185
|
+
distributions/beta_pert
|
|
183
186
|
|
|
184
187
|
Random number generators
|
|
185
188
|
------------------------
|
|
@@ -23,16 +23,16 @@ classifiers = [
|
|
|
23
23
|
"Programming Language :: Python :: 3.10",
|
|
24
24
|
"Programming Language :: Python :: 3.11",
|
|
25
25
|
"Programming Language :: Python :: 3.12",
|
|
26
|
+
"Programming Language :: Python :: 3.13",
|
|
27
|
+
"Programming Language :: Python :: 3.14",
|
|
26
28
|
"Natural Language :: English",
|
|
27
29
|
"Operating System :: OS Independent",
|
|
28
30
|
"Topic :: Scientific/Engineering"
|
|
29
31
|
]
|
|
30
|
-
requires-python = ">=3.
|
|
32
|
+
requires-python = ">=3.9"
|
|
31
33
|
dependencies = [
|
|
32
|
-
|
|
33
|
-
|
|
34
|
-
"numpy>=1.6.0",
|
|
35
|
-
"scipy>=1.1.0",
|
|
34
|
+
"numpy",
|
|
35
|
+
"scipy",
|
|
36
36
|
]
|
|
37
37
|
|
|
38
38
|
[project.urls]
|
|
@@ -62,6 +62,7 @@ dev = [
|
|
|
62
62
|
[tool.setuptools]
|
|
63
63
|
license-files = ["LICENSE"]
|
|
64
64
|
include-package-data = true
|
|
65
|
+
package-dir = {"" = "src"}
|
|
65
66
|
packages = [
|
|
66
67
|
"stats_arrays",
|
|
67
68
|
"stats_arrays.distributions",
|
|
@@ -77,7 +78,8 @@ norecursedirs = [
|
|
|
77
78
|
"build",
|
|
78
79
|
".tox"
|
|
79
80
|
]
|
|
80
|
-
testpaths = ["tests
|
|
81
|
+
testpaths = ["tests"]
|
|
82
|
+
pythonpath = ["src"]
|
|
81
83
|
|
|
82
84
|
[tool.flake8]
|
|
83
85
|
# Some sane defaults for the code style checker flake8
|
|
@@ -91,6 +93,7 @@ exclude = [
|
|
|
91
93
|
"dist",
|
|
92
94
|
".eggs",
|
|
93
95
|
"docs/conf.py",
|
|
96
|
+
"src/stats_arrays/__pycache__",
|
|
94
97
|
]
|
|
95
98
|
|
|
96
99
|
[tool.black]
|
|
@@ -1,8 +1,9 @@
|
|
|
1
|
-
__version__ = "0
|
|
1
|
+
__version__ = "1.0"
|
|
2
2
|
|
|
3
3
|
__all__ = (
|
|
4
4
|
"BernoulliUncertainty",
|
|
5
5
|
"BetaUncertainty",
|
|
6
|
+
"BetaPERTUncertainty",
|
|
6
7
|
"BoundedUncertaintyBase",
|
|
7
8
|
"DiscreteUniform",
|
|
8
9
|
"GammaUncertainty",
|
|
@@ -30,8 +31,9 @@ __all__ = (
|
|
|
30
31
|
"WeibullUncertainty",
|
|
31
32
|
)
|
|
32
33
|
|
|
33
|
-
from .distributions import (
|
|
34
|
+
from stats_arrays.distributions import (
|
|
34
35
|
BernoulliUncertainty,
|
|
36
|
+
BetaPERTUncertainty,
|
|
35
37
|
BetaUncertainty,
|
|
36
38
|
BoundedUncertaintyBase,
|
|
37
39
|
DiscreteUniform,
|
|
@@ -47,7 +49,7 @@ from .distributions import (
|
|
|
47
49
|
UniformUncertainty,
|
|
48
50
|
WeibullUncertainty,
|
|
49
51
|
)
|
|
50
|
-
from .errors import (
|
|
52
|
+
from stats_arrays.errors import (
|
|
51
53
|
ImproperBoundsError,
|
|
52
54
|
InvalidParamsError,
|
|
53
55
|
MaximumIterationsError,
|
|
@@ -57,5 +59,9 @@ from .errors import (
|
|
|
57
59
|
UnknownUncertaintyType,
|
|
58
60
|
UnreasonableBoundsError,
|
|
59
61
|
)
|
|
60
|
-
from .random import
|
|
61
|
-
|
|
62
|
+
from stats_arrays.random import (
|
|
63
|
+
LatinHypercubeRNG,
|
|
64
|
+
MCRandomNumberGenerator,
|
|
65
|
+
RandomNumberGenerator,
|
|
66
|
+
)
|
|
67
|
+
from stats_arrays.uncertainty_choices import uncertainty_choices
|
|
@@ -1,3 +1,6 @@
|
|
|
1
|
+
import numpy.typing as npt
|
|
2
|
+
from numpy import array
|
|
3
|
+
|
|
1
4
|
# Weighted formulas from
|
|
2
5
|
# http://en.wikipedia.org/wiki/Weighted_mean#Weighted_sample_variance
|
|
3
6
|
|
|
@@ -5,25 +8,23 @@
|
|
|
5
8
|
# http://pygsl.sourceforge.net/reference/pygsl/node36.html
|
|
6
9
|
|
|
7
10
|
|
|
8
|
-
def weighted_mean(values, weights):
|
|
11
|
+
def weighted_mean(values: npt.NDArray, weights: npt.NDArray) -> npt.NDArray:
|
|
9
12
|
assert values.shape == weights.shape
|
|
10
13
|
return (values * weights).sum() / weights.sum()
|
|
11
14
|
|
|
12
15
|
|
|
13
|
-
def weighted_sample_variance(values, weights):
|
|
16
|
+
def weighted_sample_variance(values: npt.NDArray, weights: npt.NDArray) -> npt.NDArray:
|
|
14
17
|
assert values.shape == weights.shape
|
|
15
18
|
return (
|
|
16
19
|
(weights * values**2).sum() * weights.sum() - ((weights * values).sum()) ** 2
|
|
17
20
|
) / ((weights.sum()) ** 2 - (weights**2).sum())
|
|
18
21
|
|
|
19
22
|
|
|
20
|
-
def weighted_sample_stddev(values, weights):
|
|
23
|
+
def weighted_sample_stddev(values: npt.NDArray, weights: npt.NDArray) -> npt.NDArray:
|
|
21
24
|
return weighted_sample_variance(values, weights) ** 0.5
|
|
22
25
|
|
|
23
26
|
|
|
24
27
|
def test():
|
|
25
|
-
from numpy import array
|
|
26
|
-
|
|
27
28
|
a = array((1, 2, 3, 4))
|
|
28
29
|
print(a**2)
|
|
29
30
|
b = array((0.5, 0.5, 1, 1))
|
|
@@ -0,0 +1,35 @@
|
|
|
1
|
+
__all__ = (
|
|
2
|
+
"BernoulliUncertainty",
|
|
3
|
+
"BetaPERTUncertainty",
|
|
4
|
+
"BetaUncertainty",
|
|
5
|
+
"BoundedUncertaintyBase",
|
|
6
|
+
"DiscreteUniform",
|
|
7
|
+
"GammaUncertainty",
|
|
8
|
+
"GeneralizedExtremeValueUncertainty",
|
|
9
|
+
"LognormalUncertainty",
|
|
10
|
+
"NormalUncertainty",
|
|
11
|
+
"NoUncertainty",
|
|
12
|
+
"StudentsTUncertainty",
|
|
13
|
+
"TriangularUncertainty",
|
|
14
|
+
"UncertaintyBase",
|
|
15
|
+
"UndefinedUncertainty",
|
|
16
|
+
"UniformUncertainty",
|
|
17
|
+
"WeibullUncertainty",
|
|
18
|
+
)
|
|
19
|
+
|
|
20
|
+
from stats_arrays.distributions.base import BoundedUncertaintyBase, UncertaintyBase
|
|
21
|
+
from stats_arrays.distributions.bernoulli import BernoulliUncertainty
|
|
22
|
+
from stats_arrays.distributions.beta import BetaUncertainty
|
|
23
|
+
from stats_arrays.distributions.beta_pert import BetaPERTUncertainty
|
|
24
|
+
from stats_arrays.distributions.discrete_uniform import DiscreteUniform
|
|
25
|
+
from stats_arrays.distributions.extreme import GeneralizedExtremeValueUncertainty
|
|
26
|
+
from stats_arrays.distributions.gamma import GammaUncertainty
|
|
27
|
+
from stats_arrays.distributions.geometric import (
|
|
28
|
+
TriangularUncertainty,
|
|
29
|
+
UniformUncertainty,
|
|
30
|
+
)
|
|
31
|
+
from stats_arrays.distributions.lognormal import LognormalUncertainty
|
|
32
|
+
from stats_arrays.distributions.normal import NormalUncertainty
|
|
33
|
+
from stats_arrays.distributions.student import StudentsTUncertainty
|
|
34
|
+
from stats_arrays.distributions.undefined import NoUncertainty, UndefinedUncertainty
|
|
35
|
+
from stats_arrays.distributions.weibull import WeibullUncertainty
|
|
@@ -1,14 +1,21 @@
|
|
|
1
|
-
from
|
|
1
|
+
from typing import Optional, Tuple
|
|
2
2
|
|
|
3
3
|
import numpy as np
|
|
4
|
+
import numpy.typing as npt
|
|
4
5
|
|
|
5
|
-
from
|
|
6
|
+
from stats_arrays.errors import (
|
|
6
7
|
ImproperBoundsError,
|
|
7
8
|
InvalidParamsError,
|
|
8
9
|
MaximumIterationsError,
|
|
9
10
|
UnreasonableBoundsError,
|
|
10
11
|
)
|
|
11
|
-
from
|
|
12
|
+
from stats_arrays.utils import (
|
|
13
|
+
ParamsArray,
|
|
14
|
+
construct_params_array,
|
|
15
|
+
one_row_params_array,
|
|
16
|
+
rescale_to_unitary_interval,
|
|
17
|
+
rescale_vector_to_params,
|
|
18
|
+
)
|
|
12
19
|
|
|
13
20
|
np.seterr(invalid="ignore")
|
|
14
21
|
|
|
@@ -26,12 +33,13 @@ class UncertaintyBase:
|
|
|
26
33
|
|
|
27
34
|
"""
|
|
28
35
|
|
|
36
|
+
id = None
|
|
29
37
|
default_number_points_in_pdf = 200
|
|
30
38
|
standard_deviations_in_default_range = 2.2
|
|
31
39
|
|
|
32
40
|
# Conversion utilities ###
|
|
33
41
|
@classmethod
|
|
34
|
-
def from_tuples(cls, *data):
|
|
42
|
+
def from_tuples(cls, *data: tuple) -> ParamsArray:
|
|
35
43
|
"""
|
|
36
44
|
Construct a :ref:`hpa` from parameter tuples.
|
|
37
45
|
|
|
@@ -72,11 +80,16 @@ class UncertaintyBase:
|
|
|
72
80
|
"""
|
|
73
81
|
params = construct_params_array(len(data), True)
|
|
74
82
|
for index, row in enumerate(data):
|
|
83
|
+
# Check if uncertainty_type is valid (must be < 256)
|
|
84
|
+
uncertainty_type = row[6] # uncertainty_type is the 7th element (index 6)
|
|
85
|
+
if uncertainty_type >= 256:
|
|
86
|
+
raise InvalidParamsError("`uncertainty_type` must be less than 256")
|
|
75
87
|
params[index] = row
|
|
88
|
+
|
|
76
89
|
return params
|
|
77
90
|
|
|
78
91
|
@classmethod
|
|
79
|
-
def from_dicts(cls, *dicts):
|
|
92
|
+
def from_dicts(cls, *dicts: dict) -> ParamsArray:
|
|
80
93
|
"""
|
|
81
94
|
Construct a :ref:`hpa` from parameter dictionaries.
|
|
82
95
|
|
|
@@ -119,12 +132,16 @@ class UncertaintyBase:
|
|
|
119
132
|
data = [obj.get(key, default) for key, default in LABELS]
|
|
120
133
|
if "uncertainty type" in obj and "uncertainty_type" not in obj:
|
|
121
134
|
data[-1] = obj["uncertainty type"]
|
|
135
|
+
# Check if uncertainty_type is valid (must be < 256)
|
|
136
|
+
uncertainty_type = data[6] # uncertainty_type is the last element
|
|
137
|
+
if uncertainty_type >= 256:
|
|
138
|
+
raise InvalidParamsError("`uncertainty_type` must be less than 256")
|
|
122
139
|
params[index] = tuple(data)
|
|
123
140
|
return params
|
|
124
141
|
|
|
125
142
|
# Utility methods ###
|
|
126
143
|
@classmethod
|
|
127
|
-
def validate(cls, params):
|
|
144
|
+
def validate(cls, params: ParamsArray) -> None:
|
|
128
145
|
"""
|
|
129
146
|
Validate the parameter array for uncertainty distribution.
|
|
130
147
|
|
|
@@ -141,7 +158,7 @@ class UncertaintyBase:
|
|
|
141
158
|
raise ImproperBoundsError
|
|
142
159
|
|
|
143
160
|
@classmethod
|
|
144
|
-
def check_2d_inputs(cls, params, vector):
|
|
161
|
+
def check_2d_inputs(cls, params: ParamsArray, vector: npt.NDArray) -> npt.NDArray:
|
|
145
162
|
"""Convert ``vector`` to 2 dimensions if not already, and raise ``stats_arrays.InvalidParamsError`` if ``vector`` and ``params`` dimensions don't match."""
|
|
146
163
|
if len(vector.shape) == 1:
|
|
147
164
|
# Slices from structured arrays can't always be resized
|
|
@@ -157,7 +174,9 @@ class UncertaintyBase:
|
|
|
157
174
|
|
|
158
175
|
@classmethod
|
|
159
176
|
@one_row_params_array
|
|
160
|
-
def check_bounds_reasonableness(
|
|
177
|
+
def check_bounds_reasonableness(
|
|
178
|
+
cls, params: ParamsArray, threshold: float = 0.1
|
|
179
|
+
) -> None:
|
|
161
180
|
"""
|
|
162
181
|
Test if there is at least a ``threshold`` percent chance of generating random numbers within the provided bounds.
|
|
163
182
|
|
|
@@ -182,9 +201,13 @@ class UncertaintyBase:
|
|
|
182
201
|
# Used for Monte Carlo ###
|
|
183
202
|
@classmethod
|
|
184
203
|
def bounded_random_variables(
|
|
185
|
-
cls,
|
|
186
|
-
|
|
187
|
-
|
|
204
|
+
cls,
|
|
205
|
+
params: ParamsArray,
|
|
206
|
+
size: int,
|
|
207
|
+
seeded_random: Optional[np.random.RandomState] = None,
|
|
208
|
+
maximum_iterations: int = 50,
|
|
209
|
+
) -> npt.NDArray:
|
|
210
|
+
"""Generate random variables repeatedly until all variables are within the bounds of each distribution. Raise MaximumIterationsError if this takes more that `maximum_iterations`. Uses `random_variables` for random number generation.
|
|
188
211
|
|
|
189
212
|
.. rubric:: Inputs
|
|
190
213
|
|
|
@@ -228,7 +251,12 @@ class UncertaintyBase:
|
|
|
228
251
|
return data
|
|
229
252
|
|
|
230
253
|
@classmethod
|
|
231
|
-
def random_variables(
|
|
254
|
+
def random_variables(
|
|
255
|
+
cls,
|
|
256
|
+
params: ParamsArray,
|
|
257
|
+
size: int,
|
|
258
|
+
seeded_random: Optional[np.random.RandomState] = None,
|
|
259
|
+
) -> npt.NDArray:
|
|
232
260
|
"""Generate random variables for the given uncertainty. Should **not check** to ensure that random samples are with the (minimum, maximum bounds). Bounds checking is provided by the `bounded_random_variables` class method.
|
|
233
261
|
|
|
234
262
|
.. rubric:: Inputs
|
|
@@ -242,9 +270,8 @@ class UncertaintyBase:
|
|
|
242
270
|
An array of random values, with dimensions `params` rows by `size`."""
|
|
243
271
|
raise NotImplementedError
|
|
244
272
|
|
|
245
|
-
# Used for Latin Hypercube Monte Carlo ###
|
|
246
273
|
@classmethod
|
|
247
|
-
def ppf(cls, params, percentages):
|
|
274
|
+
def ppf(cls, params: ParamsArray, percentages: npt.NDArray) -> npt.NDArray:
|
|
248
275
|
"""Return percent point function (inverse of CDF, e.g. value in distribution where x percent of the distribution is less than value) for various distributions.
|
|
249
276
|
|
|
250
277
|
.. rubric:: Inputs
|
|
@@ -256,11 +283,10 @@ class UncertaintyBase:
|
|
|
256
283
|
|
|
257
284
|
An array of values within the ranges of each distribtion, with `params` rows and `percentages` columns.
|
|
258
285
|
"""
|
|
259
|
-
percentages = cls.check_ppf_inputs(params, percentages)
|
|
260
286
|
raise NotImplementedError
|
|
261
287
|
|
|
262
288
|
@classmethod
|
|
263
|
-
def cdf(cls, params, vector):
|
|
289
|
+
def cdf(cls, params: ParamsArray, vector: npt.NDArray) -> npt.NDArray:
|
|
264
290
|
"""Used when a distribution is bounded, to determine where to begin or end the percentages used in calculating hypercube sampling space.
|
|
265
291
|
|
|
266
292
|
.. rubric:: Inputs
|
|
@@ -272,13 +298,12 @@ class UncertaintyBase:
|
|
|
272
298
|
|
|
273
299
|
An array of cumulative densities, bounded on (0,1), with `params` rows and `vector` columns.
|
|
274
300
|
"""
|
|
275
|
-
vector = cls.check_cdf_inputs(params, vector)
|
|
276
301
|
raise NotImplementedError
|
|
277
302
|
|
|
278
303
|
# Used for graphing ###
|
|
279
304
|
@classmethod
|
|
280
305
|
@one_row_params_array
|
|
281
|
-
def statistics(cls, params):
|
|
306
|
+
def statistics(cls, params: ParamsArray) -> dict:
|
|
282
307
|
"""Build a dictionary of mean, mode, median, and 95% confidence interval upper and lower values.
|
|
283
308
|
|
|
284
309
|
.. rubric:: Inputs
|
|
@@ -299,7 +324,7 @@ class UncertaintyBase:
|
|
|
299
324
|
|
|
300
325
|
@classmethod
|
|
301
326
|
@one_row_params_array
|
|
302
|
-
def pdf(cls, params, xs=None):
|
|
327
|
+
def pdf(cls, params: ParamsArray, xs: Optional[npt.NDArray] = None) -> npt.NDArray:
|
|
303
328
|
"""Provide a standard interface to calculate the probability distribution function of a uncertainty distribution. Default is `cls.default_number_points_in_pdf` points between min to max range if bounds are present, or `cls.standard_deviations_in_default_range` standard distributions.
|
|
304
329
|
|
|
305
330
|
.. rubric:: Inputs
|
|
@@ -320,7 +345,7 @@ class BoundedUncertaintyBase(UncertaintyBase):
|
|
|
320
345
|
"""An uncertainty distribution where minimum and maximum bounds are required. No bounds checking is required for these distributions, as bounds are integral inputs into the sample space generator."""
|
|
321
346
|
|
|
322
347
|
@classmethod
|
|
323
|
-
def validate(cls, params):
|
|
348
|
+
def validate(cls, params: ParamsArray) -> None:
|
|
324
349
|
super(BoundedUncertaintyBase, cls).validate(params)
|
|
325
350
|
if np.isnan(params["minimum"]).sum() or np.isnan(params["maximum"]).sum():
|
|
326
351
|
raise ImproperBoundsError(
|
|
@@ -328,21 +353,30 @@ class BoundedUncertaintyBase(UncertaintyBase):
|
|
|
328
353
|
)
|
|
329
354
|
|
|
330
355
|
@classmethod
|
|
331
|
-
def
|
|
332
|
-
|
|
333
|
-
|
|
334
|
-
|
|
335
|
-
|
|
356
|
+
def rescale_to_unitary_interval(
|
|
357
|
+
cls, params: ParamsArray
|
|
358
|
+
) -> Tuple[npt.NDArray, npt.NDArray]:
|
|
359
|
+
return rescale_to_unitary_interval(params=params)
|
|
360
|
+
|
|
361
|
+
@classmethod
|
|
362
|
+
def rescale_vector_to_params(
|
|
363
|
+
cls, params: npt.NDArray, vector: npt.NDArray
|
|
364
|
+
) -> npt.NDArray:
|
|
365
|
+
return rescale_vector_to_params(params=params, vector=vector)
|
|
336
366
|
|
|
337
367
|
@classmethod
|
|
338
368
|
def bounded_random_variables(
|
|
339
|
-
cls,
|
|
340
|
-
|
|
369
|
+
cls,
|
|
370
|
+
params: ParamsArray,
|
|
371
|
+
size: int,
|
|
372
|
+
seeded_random: Optional[np.random.RandomState] = None,
|
|
373
|
+
maximum_iterations: Optional[int] = None,
|
|
374
|
+
) -> npt.NDArray:
|
|
341
375
|
"""No bounds checking because the bounds do not exclude any of the distribution."""
|
|
342
376
|
return cls.random_variables(params, size, seeded_random)
|
|
343
377
|
|
|
344
378
|
@classmethod
|
|
345
379
|
@one_row_params_array
|
|
346
|
-
def check_bounds_reasonableness(cls, params):
|
|
380
|
+
def check_bounds_reasonableness(cls, params: ParamsArray) -> None:
|
|
347
381
|
"""Always true because the bounds do not exclude any of the distribution."""
|
|
348
382
|
return
|
|
@@ -0,0 +1,50 @@
|
|
|
1
|
+
from typing import Optional
|
|
2
|
+
|
|
3
|
+
import numpy as np
|
|
4
|
+
import numpy.typing as npt
|
|
5
|
+
|
|
6
|
+
from stats_arrays.distributions.base import UncertaintyBase
|
|
7
|
+
from stats_arrays.errors import InvalidParamsError
|
|
8
|
+
from stats_arrays.utils import ParamsArray
|
|
9
|
+
|
|
10
|
+
|
|
11
|
+
class BernoulliUncertainty(UncertaintyBase):
|
|
12
|
+
id = 6
|
|
13
|
+
description = "Bernoulli uncertainty"
|
|
14
|
+
|
|
15
|
+
@classmethod
|
|
16
|
+
def validate(cls, params: ParamsArray) -> None:
|
|
17
|
+
"""Validate that loc is between 0 and 1 (inclusive)."""
|
|
18
|
+
if (params["loc"] < 0).sum() or (params["loc"] > 1).sum():
|
|
19
|
+
raise InvalidParamsError(
|
|
20
|
+
"Bernoulli uncertainty requires loc values between 0 and 1 (inclusive)."
|
|
21
|
+
)
|
|
22
|
+
|
|
23
|
+
@classmethod
|
|
24
|
+
def random_variables(
|
|
25
|
+
cls,
|
|
26
|
+
params: ParamsArray,
|
|
27
|
+
size: int,
|
|
28
|
+
seeded_random: Optional[np.random.RandomState] = None,
|
|
29
|
+
) -> npt.NDArray:
|
|
30
|
+
if not seeded_random:
|
|
31
|
+
seeded_random = np.random.RandomState()
|
|
32
|
+
data = np.zeros((params.shape[0], size))
|
|
33
|
+
mask = (
|
|
34
|
+
seeded_random.random_sample(size * params.shape[0]).reshape(
|
|
35
|
+
(params.shape[0], size)
|
|
36
|
+
)
|
|
37
|
+
<= params["loc"]
|
|
38
|
+
)
|
|
39
|
+
data[mask] = 1
|
|
40
|
+
return data
|
|
41
|
+
|
|
42
|
+
@classmethod
|
|
43
|
+
def cdf(cls, params: ParamsArray, vector: npt.NDArray) -> npt.NDArray:
|
|
44
|
+
vector = cls.check_2d_inputs(params, vector)
|
|
45
|
+
return (vector <= params["loc"].reshape(-1, 1)) * 1.0
|
|
46
|
+
|
|
47
|
+
@classmethod
|
|
48
|
+
def ppf(cls, params: ParamsArray, percentages: npt.NDArray) -> npt.NDArray:
|
|
49
|
+
percentages = cls.check_2d_inputs(params, percentages)
|
|
50
|
+
return (percentages <= params["loc"].reshape(-1, 1)) * 1.0
|