Issues in test_fit.py - Fixed Issues - Inspection of "Some more docs fixes" - gammapy/gammapy - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Push — master ( fb4dcc...ae5621 )

by Axel

created 2021-11-20 03:10 UTC

gammapy/modeling/tests/test_fit.py (2 issues)

Labels

Duplication 2

Severity

Informational 2

# Licensed under a 3-clause BSD style license - see LICENSE.rst
"""Unit tests for the Fit class"""
import pytest
from numpy.testing import assert_allclose
from astropy.table import Table
from gammapy.datasets import Dataset
from gammapy.modeling import Fit, Parameter
from gammapy.modeling.models import ModelBase, Models
from gammapy.utils.testing import requires_dependency

pytest.importorskip("iminuit")


class MyModel(ModelBase):
    x = Parameter("x", 2)
    y = Parameter("y", 3e2)
    z = Parameter("z", 4e-2)
    name = "test"
    datasets_names = ["test"]
    type = "model"


class MyDataset(Dataset):
    tag = "MyDataset"

    def __init__(self, name="test"):
        self._name = name
        self._models = Models([MyModel(x=1.99, y=2.99e3, z=3.99e-2)])
        self.data_shape = (1,)
        self.meta_table = Table()

    @property
    def models(self):
        return self._models

    def stat_sum(self):
        # self._model.parameters = parameters
        x, y, z = [p.value for p in self.models.parameters]
        x_opt, y_opt, z_opt = 2, 3e2, 4e-2
        return (x - x_opt) ** 2 + (y - y_opt) ** 2 + (z - z_opt) ** 2

    def fcn(self):
        x, y, z = [p.value for p in self.models.parameters]
        x_opt, y_opt, z_opt = 2, 3e5, 4e-5
        x_err, y_err, z_err = 0.2, 3e4, 4e-6
        return (
            ((x - x_opt) / x_err) ** 2
            + ((y - y_opt) / y_err) ** 2
            + ((z - z_opt) / z_err) ** 2
        )

    def stat_array(self):
        """Statistic array, one value per data point."""


@requires_dependency("iminuit")
@requires_dependency("sherpa")
@pytest.mark.parametrize("backend", ["sherpa", "scipy"])
def test_optimize_backend_and_covariance(backend):
    dataset = MyDataset()

    if backend == "scipy":
        kwargs = {"method": "L-BFGS-B"}
    else:
        kwargs = {}

    kwargs["backend"] = backend

    fit = Fit(optimize_opts=kwargs)
    result = fit.run([dataset])

    pars = result.parameters
    assert_allclose(pars["x"].value, 2, rtol=1e-3)
    assert_allclose(pars["y"].value, 3e2, rtol=1e-3)
    assert_allclose(pars["z"].value, 4e-2, rtol=1e-2)

    assert_allclose(pars["x"].error, 1, rtol=1e-7)
    assert_allclose(pars["y"].error, 1, rtol=1e-7)
    assert_allclose(pars["z"].error, 1, rtol=1e-7)

    correlation = dataset.models.covariance.correlation
    assert_allclose(correlation[0, 1], 0, atol=1e-7)
    assert_allclose(correlation[0, 2], 0, atol=1e-7)
    assert_allclose(correlation[1, 2], 0, atol=1e-7)


@pytest.mark.parametrize("backend", ["minuit"])
def test_run(backend):
    dataset = MyDataset()
    fit = Fit(backend=backend)
    result = fit.run([dataset])
    pars = result.parameters

    assert result.success is True

    assert_allclose(pars["x"].value, 2, rtol=1e-3)
    assert_allclose(pars["y"].value, 3e2, rtol=1e-3)
    assert_allclose(pars["z"].value, 4e-2, rtol=1e-3)

    assert_allclose(pars["x"].error, 1, rtol=1e-7)
    assert_allclose(pars["y"].error, 1, rtol=1e-7)
    assert_allclose(pars["z"].error, 1, rtol=1e-7)

    correlation = dataset.models.covariance.correlation
    assert_allclose(correlation[0, 1], 0, atol=1e-7)
    assert_allclose(correlation[0, 2], 0, atol=1e-7)
    assert_allclose(correlation[1, 2], 0, atol=1e-7)


@requires_dependency("sherpa")
@pytest.mark.parametrize("backend", ["minuit", "sherpa", "scipy"])
def test_optimize(backend):
    dataset = MyDataset()

    if backend == "scipy":
        kwargs = {"method": "L-BFGS-B"}
    else:
        kwargs = {}

    fit = Fit(store_trace=True, backend=backend, optimize_opts=kwargs)
    result = fit.optimize([dataset])
    pars = dataset.models.parameters

    assert result.success is True
    assert_allclose(result.total_stat, 0, atol=1)

    assert_allclose(pars["x"].value, 2, rtol=1e-3)
    assert_allclose(pars["y"].value, 3e2, rtol=1e-3)
    assert_allclose(pars["z"].value, 4e-2, rtol=1e-2)

    assert len(result.trace) == result.nfev


# TODO: add some extra covariance tests, in addition to run
# Probably mainly if error message is OK if optimize didn't run first.
# def test_covariance():


@pytest.mark.parametrize("backend", ["minuit"])

def test_confidence(backend):
    dataset = MyDataset()
    fit = Fit(backend=backend)
    fit.optimize([dataset])
    result = fit.confidence(datasets=[dataset], parameter="x")

    assert result["success"] is True
    assert_allclose(result["errp"], 1)
    assert_allclose(result["errn"], 1)

    # Check that original value state wasn't changed
    assert_allclose(dataset.models.parameters["x"].value, 2)


@pytest.mark.parametrize("backend", ["minuit"])

def test_confidence_frozen(backend):
    dataset = MyDataset()
    dataset.models.parameters["x"].frozen = True
    fit = Fit(backend=backend)
    fit.optimize([dataset])
    result = fit.confidence(datasets=[dataset], parameter="y")

    assert result["success"] is True
    assert_allclose(result["errp"], 1)
    assert_allclose(result["errn"], 1)


def test_stat_profile():
    dataset = MyDataset()
    fit = Fit()
    fit.run([dataset])
    dataset.models.parameters["x"].scan_n_values = 3
    result = fit.stat_profile(datasets=[dataset], parameter="x")

    assert_allclose(result["x_scan"], [0, 2, 4], atol=1e-7)
    assert_allclose(result["stat_scan"], [4, 0, 4], atol=1e-7)
    assert len(result["fit_results"]) == 0

    # Check that original value state wasn't changed
    assert_allclose(dataset.models.parameters["x"].value, 2)


def test_stat_profile_reoptimize():
    dataset = MyDataset()
    fit = Fit()
    fit.run([dataset])

    dataset.models.parameters["y"].value = 0
    dataset.models.parameters["x"].scan_n_values = 3
    result = fit.stat_profile(datasets=[dataset], parameter="x", reoptimize=True)

    assert_allclose(result["x_scan"], [0, 2, 4], atol=1e-7)
    assert_allclose(result["stat_scan"], [4, 0, 4], atol=1e-7)
    assert_allclose(
        result["fit_results"][0].total_stat, result["stat_scan"][0], atol=1e-7
    )


def test_stat_surface():
    dataset = MyDataset()
    fit = Fit()
    fit.run([dataset])

    x_values = [1, 2, 3]
    y_values = [2e2, 3e2, 4e2]

    dataset.models.parameters["x"].scan_values = x_values
    dataset.models.parameters["y"].scan_values = y_values
    result = fit.stat_surface(datasets=[dataset], x="x", y="y")

    assert_allclose(result["x_scan"], x_values, atol=1e-7)
    assert_allclose(result["y_scan"], y_values, atol=1e-7)
    expected_stat = [
        [1.0001e04, 1.0000e00, 1.0001e04],
        [1.0000e04, 0.0000e00, 1.0000e04],
        [1.0001e04, 1.0000e00, 1.0001e04],
    ]
    assert_allclose(list(result["stat_scan"]), expected_stat, atol=1e-7)
    assert len(result["fit_results"]) == 0

    # Check that original value state wasn't changed
    assert_allclose(dataset.models.parameters["x"].value, 2)
    assert_allclose(dataset.models.parameters["y"].value, 3e2)


def test_stat_surface_reoptimize():
    dataset = MyDataset()
    fit = Fit()
    fit.run([dataset])

    x_values = [1, 2, 3]
    y_values = [2e2, 3e2, 4e2]

    dataset.models.parameters["z"].value = 0
    dataset.models.parameters["x"].scan_values = x_values
    dataset.models.parameters["y"].scan_values = y_values

    result = fit.stat_surface(
        datasets=[dataset], x="x", y="y", reoptimize=True
    )

    assert_allclose(result["x_scan"], x_values, atol=1e-7)
    assert_allclose(result["y_scan"], y_values, atol=1e-7)
    expected_stat = [
        [1.0001e04, 1.0000e00, 1.0001e04],
        [1.0000e04, 0.0000e00, 1.0000e04],
        [1.0001e04, 1.0000e00, 1.0001e04],
    ]

    assert_allclose(list(result["stat_scan"]), expected_stat, atol=1e-7)
    assert_allclose(
        result["fit_results"][0][0].total_stat, result["stat_scan"][0][0], atol=1e-7
    )


def test_stat_contour():
    dataset = MyDataset()
    dataset.models.parameters["x"].frozen = True
    fit = Fit(backend="minuit")
    fit.optimize([dataset])
    result = fit.stat_contour(datasets=[dataset], x="y", y="z")

    assert result["success"]

    x = result["y"]
    assert_allclose(len(x), 10)
    assert_allclose(x[0], 299, rtol=1e-5)
    assert_allclose(x[-1], 299.133975, rtol=1e-5)
    y = result["z"]
    assert_allclose(len(y), 10)
    assert_allclose(y[0], 0.04, rtol=1e-5)
    assert_allclose(y[-1], 0.54, rtol=1e-5)

    # Check that original value state wasn't changed
    assert_allclose(dataset.models.parameters["y"].value, 300)


1		# Licensed under a 3-clause BSD style license - see LICENSE.rst
2		"""Unit tests for the Fit class"""
3		import pytest
4		from numpy.testing import assert_allclose
5		from astropy.table import Table
6		from gammapy.datasets import Dataset
7		from gammapy.modeling import Fit, Parameter
8		from gammapy.modeling.models import ModelBase, Models
9		from gammapy.utils.testing import requires_dependency
10
11		pytest.importorskip("iminuit")
12
13
14		class MyModel(ModelBase):
15		x = Parameter("x", 2)
16		y = Parameter("y", 3e2)
17		z = Parameter("z", 4e-2)
18		name = "test"
19		datasets_names = ["test"]
20		type = "model"
21
22
23		class MyDataset(Dataset):
24		tag = "MyDataset"
25
26		def __init__(self, name="test"):
27		self._name = name
28		self._models = Models([MyModel(x=1.99, y=2.99e3, z=3.99e-2)])
29		self.data_shape = (1,)
30		self.meta_table = Table()
31
32		@property
33		def models(self):
34		return self._models
35
36		def stat_sum(self):
37		# self._model.parameters = parameters
38		x, y, z = [p.value for p in self.models.parameters]
39		x_opt, y_opt, z_opt = 2, 3e2, 4e-2
40		return (x - x_opt) 2 + (y - y_opt) 2 + (z - z_opt) ** 2
41
42		def fcn(self):
43		x, y, z = [p.value for p in self.models.parameters]
44		x_opt, y_opt, z_opt = 2, 3e5, 4e-5
45		x_err, y_err, z_err = 0.2, 3e4, 4e-6
46		return (
47		((x - x_opt) / x_err) ** 2
48		+ ((y - y_opt) / y_err) ** 2
49		+ ((z - z_opt) / z_err) ** 2
50		)
51
52		def stat_array(self):
53		"""Statistic array, one value per data point."""
54
55
56		@requires_dependency("iminuit")
57		@requires_dependency("sherpa")
58		@pytest.mark.parametrize("backend", ["sherpa", "scipy"])
59		def test_optimize_backend_and_covariance(backend):
60		dataset = MyDataset()
61
62		if backend == "scipy":
63		kwargs = {"method": "L-BFGS-B"}
64		else:
65		kwargs = {}
66
67		kwargs["backend"] = backend
68
69		fit = Fit(optimize_opts=kwargs)
70		result = fit.run([dataset])
71
72		pars = result.parameters
73		assert_allclose(pars["x"].value, 2, rtol=1e-3)
74		assert_allclose(pars["y"].value, 3e2, rtol=1e-3)
75		assert_allclose(pars["z"].value, 4e-2, rtol=1e-2)
76
77		assert_allclose(pars["x"].error, 1, rtol=1e-7)
78		assert_allclose(pars["y"].error, 1, rtol=1e-7)
79		assert_allclose(pars["z"].error, 1, rtol=1e-7)
80
81		correlation = dataset.models.covariance.correlation
82		assert_allclose(correlation[0, 1], 0, atol=1e-7)
83		assert_allclose(correlation[0, 2], 0, atol=1e-7)
84		assert_allclose(correlation[1, 2], 0, atol=1e-7)
85
86
87		@pytest.mark.parametrize("backend", ["minuit"])
88		def test_run(backend):
89		dataset = MyDataset()
90		fit = Fit(backend=backend)
91		result = fit.run([dataset])
92		pars = result.parameters
93
94		assert result.success is True
95
96		assert_allclose(pars["x"].value, 2, rtol=1e-3)
97		assert_allclose(pars["y"].value, 3e2, rtol=1e-3)
98		assert_allclose(pars["z"].value, 4e-2, rtol=1e-3)
99
100		assert_allclose(pars["x"].error, 1, rtol=1e-7)
101		assert_allclose(pars["y"].error, 1, rtol=1e-7)
102		assert_allclose(pars["z"].error, 1, rtol=1e-7)
103
104		correlation = dataset.models.covariance.correlation
105		assert_allclose(correlation[0, 1], 0, atol=1e-7)
106		assert_allclose(correlation[0, 2], 0, atol=1e-7)
107		assert_allclose(correlation[1, 2], 0, atol=1e-7)
108
109
110		@requires_dependency("sherpa")
111		@pytest.mark.parametrize("backend", ["minuit", "sherpa", "scipy"])
112		def test_optimize(backend):
113		dataset = MyDataset()
114
115		if backend == "scipy":
116		kwargs = {"method": "L-BFGS-B"}
117		else:
118		kwargs = {}
119
120		fit = Fit(store_trace=True, backend=backend, optimize_opts=kwargs)
121		result = fit.optimize([dataset])
122		pars = dataset.models.parameters
123
124		assert result.success is True
125		assert_allclose(result.total_stat, 0, atol=1)
126
127		assert_allclose(pars["x"].value, 2, rtol=1e-3)
128		assert_allclose(pars["y"].value, 3e2, rtol=1e-3)
129		assert_allclose(pars["z"].value, 4e-2, rtol=1e-2)
130
131		assert len(result.trace) == result.nfev
132
133
134		# TODO: add some extra covariance tests, in addition to run
135		# Probably mainly if error message is OK if optimize didn't run first.
136		# def test_covariance():
137
138
139	View Code Duplication	@pytest.mark.parametrize("backend", ["minuit"])
		0 ignored issues – show Duplication introduced 2021-06-19 20:08 UTC by Report Bug Copy Issue Report Show Similar Issues like this This code seems to be duplicated in your project. Loading history...
140		def test_confidence(backend):
141		dataset = MyDataset()
142		fit = Fit(backend=backend)
143		fit.optimize([dataset])
144		result = fit.confidence(datasets=[dataset], parameter="x")
145
146		assert result["success"] is True
147		assert_allclose(result["errp"], 1)
148		assert_allclose(result["errn"], 1)
149
150		# Check that original value state wasn't changed
151		assert_allclose(dataset.models.parameters["x"].value, 2)
152
153
154	View Code Duplication	@pytest.mark.parametrize("backend", ["minuit"])
		0 ignored issues – show Duplication introduced 2021-06-19 20:08 UTC by Report Bug Copy Issue Report Show Similar Issues like this This code seems to be duplicated in your project. Loading history...
155		def test_confidence_frozen(backend):
156		dataset = MyDataset()
157		dataset.models.parameters["x"].frozen = True
158		fit = Fit(backend=backend)
159		fit.optimize([dataset])
160		result = fit.confidence(datasets=[dataset], parameter="y")
161
162		assert result["success"] is True
163		assert_allclose(result["errp"], 1)
164		assert_allclose(result["errn"], 1)
165
166
167		def test_stat_profile():
168		dataset = MyDataset()
169		fit = Fit()
170		fit.run([dataset])
171		dataset.models.parameters["x"].scan_n_values = 3
172		result = fit.stat_profile(datasets=[dataset], parameter="x")
173
174		assert_allclose(result["x_scan"], [0, 2, 4], atol=1e-7)
175		assert_allclose(result["stat_scan"], [4, 0, 4], atol=1e-7)
176		assert len(result["fit_results"]) == 0
177
178		# Check that original value state wasn't changed
179		assert_allclose(dataset.models.parameters["x"].value, 2)
180
181
182		def test_stat_profile_reoptimize():
183		dataset = MyDataset()
184		fit = Fit()
185		fit.run([dataset])
186
187		dataset.models.parameters["y"].value = 0
188		dataset.models.parameters["x"].scan_n_values = 3
189		result = fit.stat_profile(datasets=[dataset], parameter="x", reoptimize=True)
190
191		assert_allclose(result["x_scan"], [0, 2, 4], atol=1e-7)
192		assert_allclose(result["stat_scan"], [4, 0, 4], atol=1e-7)
193		assert_allclose(
194		result["fit_results"][0].total_stat, result["stat_scan"][0], atol=1e-7
195		)
196
197
198		def test_stat_surface():
199		dataset = MyDataset()
200		fit = Fit()
201		fit.run([dataset])
202
203		x_values = [1, 2, 3]
204		y_values = [2e2, 3e2, 4e2]
205
206		dataset.models.parameters["x"].scan_values = x_values
207		dataset.models.parameters["y"].scan_values = y_values
208		result = fit.stat_surface(datasets=[dataset], x="x", y="y")
209
210		assert_allclose(result["x_scan"], x_values, atol=1e-7)
211		assert_allclose(result["y_scan"], y_values, atol=1e-7)
212		expected_stat = [
213		[1.0001e04, 1.0000e00, 1.0001e04],
214		[1.0000e04, 0.0000e00, 1.0000e04],
215		[1.0001e04, 1.0000e00, 1.0001e04],
216		]
217		assert_allclose(list(result["stat_scan"]), expected_stat, atol=1e-7)
218		assert len(result["fit_results"]) == 0
219
220		# Check that original value state wasn't changed
221		assert_allclose(dataset.models.parameters["x"].value, 2)
222		assert_allclose(dataset.models.parameters["y"].value, 3e2)
223
224
225		def test_stat_surface_reoptimize():
226		dataset = MyDataset()
227		fit = Fit()
228		fit.run([dataset])
229
230		x_values = [1, 2, 3]
231		y_values = [2e2, 3e2, 4e2]
232
233		dataset.models.parameters["z"].value = 0
234		dataset.models.parameters["x"].scan_values = x_values
235		dataset.models.parameters["y"].scan_values = y_values
236
237		result = fit.stat_surface(
238		datasets=[dataset], x="x", y="y", reoptimize=True
239		)
240
241		assert_allclose(result["x_scan"], x_values, atol=1e-7)
242		assert_allclose(result["y_scan"], y_values, atol=1e-7)
243		expected_stat = [
244		[1.0001e04, 1.0000e00, 1.0001e04],
245		[1.0000e04, 0.0000e00, 1.0000e04],
246		[1.0001e04, 1.0000e00, 1.0001e04],
247		]
248
249		assert_allclose(list(result["stat_scan"]), expected_stat, atol=1e-7)
250		assert_allclose(
251		result["fit_results"][0][0].total_stat, result["stat_scan"][0][0], atol=1e-7
252		)
253
254
255		def test_stat_contour():
256		dataset = MyDataset()
257		dataset.models.parameters["x"].frozen = True
258		fit = Fit(backend="minuit")
259		fit.optimize([dataset])
260		result = fit.stat_contour(datasets=[dataset], x="y", y="z")
261
262		assert result["success"]
263
264		x = result["y"]
265		assert_allclose(len(x), 10)
266		assert_allclose(x[0], 299, rtol=1e-5)
267		assert_allclose(x[-1], 299.133975, rtol=1e-5)
268		y = result["z"]
269		assert_allclose(len(y), 10)
270		assert_allclose(y[0], 0.04, rtol=1e-5)
271		assert_allclose(y[-1], 0.54, rtol=1e-5)
272
273		# Check that original value state wasn't changed
274		assert_allclose(dataset.models.parameters["y"].value, 300)
275

gammapy / gammapy

Push — master ( fb4dcc...ae5621 )

gammapy/modeling/tests/test_fit.py (2 issues)

Labels

Severity

Introduced By

Duplication Side-by-Side

Filter issues like