def _integral_transformer_aux(points, comp, values=None):
'''
Auxiliar function to perform check of the FlatDistTransform class.
'''
tr = hep_spt.FlatDistTransform(points, values)
vals = tr.transform(comp)
# Check that the values are between 0 and 1
assert np.all(vals >= 0) and np.all(vals <= 1)
bins = 20
values, edges = np.histogram(vals, bins, range=(0, 1))
centers = hep_spt.cfe(edges)
p, residuals, _, _, _ = np.polyfit(centers, values, 0, full=True)
# Check the mean of the values (depends on the number of bins and on the
# length of the samples)
assert np.isclose(p, len(points) / float(bins))
chi2_ndof = residuals / (len(values) - 1.)
return chi2_ndof
def test_residual_asym():
'''
Test the "residual" function with asymmetric errors.
'''
size = 1000
data = np.random.normal(0, 2, size)
values, edges, ex, ey = hep_spt.errorbar_hist(data, uncert='freq')
centers = hep_spt.cfe(edges)
rv = norm.pdf(centers)
ref = float(size)*rv/rv.sum()
res, perr = hep_spt.residual(values, ey, ref)
assert perr.shape == (2, len(values))
assert np.allclose(res, values - ref)
def test_pull_asym():
'''
Test the "pull" function with asymmetric errors.
'''
size = 1000
data = np.random.normal(0, 2, size)
values, edges, ex, ey = hep_spt.errorbar_hist(data, uncert='freq')
centers = hep_spt.cfe(edges)
rv = norm.pdf(centers)
ref = float(size)*rv/rv.sum()
pull, perr = hep_spt.pull(values, ey, ref)
assert perr.shape == (2, len(values))
p = values - ref
assert np.all(pull[p >= 0] >= 0)
assert np.all(pull[p < 0] < 0)
def test_cfe():
'''
Test the "cfe" function.
'''
assert np.all(hep_spt.cfe(np.array([1, 3, 5, 7])) == [2, 4, 6])