def test_too_few_points_for_polyfit_warning():
"""Test that the correct warning is raised if data is not enough to fit."""
fac = PolyFactory(X_VALS, order=2)
fac._instack = [
{
"scale_factor": 1.0,
"shots": 100
},
{
"scale_factor": 2.0,
"shots": 100
},
]
fac._outstack = [1.0, 2.0]
with warns(
ExtrapolationWarning,
match=r"The extrapolation fit may be ill-conditioned.",
):
fac.reduce()
# test also the static "extrapolate" method.
with warns(
ExtrapolationWarning,
match=r"The extrapolation fit may be ill-conditioned.",
):
PolyFactory.extrapolate([1.0, 2.0], [1.0, 2.0], order=2)
def test_poly_extr():
"""Test of polynomial extrapolator."""
# test (order=1)
fac = PolyFactory(X_VALS, order=1)
fac.run_classical(f_lin)
assert np.isclose(fac.reduce(), f_lin(0, err=0), atol=CLOSE_TOL)
# test that, for some non-linear functions,
# order=1 is bad while order=2 is better.
seeded_f = apply_seed_to_func(f_non_lin, SEED)
fac = PolyFactory(X_VALS, order=1)
fac.run_classical(seeded_f)
assert not np.isclose(fac.reduce(), seeded_f(0, err=0), atol=NOT_CLOSE_TOL)
seeded_f = apply_seed_to_func(f_non_lin, SEED)
fac = PolyFactory(X_VALS, order=2)
fac.run_classical(seeded_f)
assert np.isclose(fac.reduce(), seeded_f(0, err=0), atol=CLOSE_TOL)
def test_too_few_points_for_polyfit_error():
"""Test that the correct error is raised if data is not enough to fit."""
fac = PolyFactory(X_VALS, order=2)
fac._instack = [
{
"scale_factor": 1.0,
"shots": 100
},
{
"scale_factor": 2.0,
"shots": 100
},
]
fac._outstack = [1.0, 2.0]
with raises(ValueError, match=r"Extrapolation order is too high."):
fac.reduce()
def test_opt_params_poly_factory(order):
"""Tests that optimal parameters are stored after calling the reduce method.
"""
fac = PolyFactory(scale_factors=np.linspace(1, 10, 10), order=order)
assert fac.opt_params == []
fac.iterate(apply_seed_to_func(f_non_lin, seed=SEED))
zne_value = fac.reduce()
assert len(fac.opt_params) == order + 1
assert np.isclose(fac.opt_params[-1], zne_value)
def test_poly_extr():
"""Test of polynomial extrapolator."""
# test (order=1)
fac = PolyFactory(X_VALS, order=1)
fac.run_classical(f_lin)
assert np.isclose(fac.reduce(), f_lin(0, err=0), atol=CLOSE_TOL)
# test that, for some non-linear functions,
# order=1 is bad while order=2 is better.
seeded_f = apply_seed_to_func(f_non_lin, SEED)
fac = PolyFactory(X_VALS, order=1)
fac.run_classical(seeded_f)
assert not np.isclose(fac.reduce(), seeded_f(0, err=0), atol=NOT_CLOSE_TOL)
seeded_f = apply_seed_to_func(f_non_lin, SEED)
fac = PolyFactory(X_VALS, order=2)
fac.run_classical(seeded_f)
zne_value = fac.reduce()
assert np.isclose(fac.reduce(), seeded_f(0, err=0), atol=CLOSE_TOL)
exp_vals = fac.get_expectation_values()
assert np.isclose(fac.extrapolate(X_VALS, exp_vals, order=2), zne_value)
assert np.isclose(
fac.extrapolate(X_VALS, exp_vals, order=2, full_output=True)[0],
zne_value,
)