class TestSpectrumFitResult:
def setup(self):
pha = gammapy_extra.filename("datasets/hess-crab4_pha/pha_obs23592.fits")
self.obs = SpectrumObservation.read(pha)
self.best_fit_model = models.PowerLaw(index=2 * u.Unit(''),
amplitude=1e-11 * u.Unit('cm-2 s-1 TeV-1'),
reference=1 * u.TeV)
self.npred = self.obs.predicted_counts(self.best_fit_model).data.value
self.covar_axis = ['index', 'amplitude']
self.covar = np.diag([0.1 ** 2, 1e-12 ** 2])
self.fit_range = [0.1, 50] * u.TeV
self.fit_result = SpectrumFitResult(
model=self.best_fit_model,
covariance=self.covar,
covar_axis=self.covar_axis,
fit_range=self.fit_range,
statname='wstat',
statval=42,
npred=self.npred,
obs=self.obs,
)
def test_basic(self):
assert 'PowerLaw' in str(self.fit_result)
assert 'index' in self.fit_result.to_table().colnames
@requires_dependency('matplotlib')
def test_plot(self):
self.fit_result.plot()
def test_io(self, tmpdir):
self.fit_result.to_yaml(tmpdir / 'test.yaml')
read_result = SpectrumFitResult.from_yaml(tmpdir / 'test.yaml')
test_e = 12.5 * u.TeV
assert_quantity_allclose(self.fit_result.model(test_e),
read_result.model(test_e))
def test_model_with_uncertainties(self):
actual = self.fit_result.model_with_uncertainties.parameters.index.s
desired = np.sqrt(self.covar[0][0])
assert actual == desired
