def test_multi_bin_single_dim():
instructions_mc = [dict(n_events=24, x=0.5),
dict(n_events=56, x=1.5)]
data, n_mc = make_data(instructions_mc)
conf = test_conf(events_per_day=42,
analysis_space=[['x', [0, 1, 5]]], default_source_class=FixedSampleSource, data=data)
lf = BinnedLogLikelihood(conf)
lf.add_rate_parameter('s0')
instructions_data = [dict(n_events=18, x=0.5),
dict(n_events=70, x=1.5)]
data, _ = make_data(instructions_data)
lf.set_data(data)
mus = [42 / n_mc * instructions_mc[i]['n_events']
for i in range(len(instructions_mc))]
seen = [instructions_data[i]['n_events']
for i in range(len(instructions_data))]
assert almost_equal(lf(),
np.sum([stats.poisson(mu).logpmf(seen_in_bin)
for mu, seen_in_bin in zip(mus, seen)]),
1e-6)
def test_BeestonBarlow_second_source():
instructions_mc_calibration = [dict(n_events=16, x=0.5),
dict(n_events=30, x=1.5),
dict(n_events=32, x=2.5),
dict(n_events=27, x=3.5),]
data_calibration, n_mc = make_data(instructions_mc_calibration)
instructions_mc_other = [dict(n_events=5, x=0.5),
dict(n_events=7, x=1.5),
dict(n_events=1, x=2.5),
dict(n_events=3, x=3.5),]
data_other, n_mc = make_data(instructions_mc_other)
# in sum, 105 calibration/mc evts.
conf = test_conf(default_source_class=FixedSampleSource,
analysis_space=[['x', [0, 1, 2, 3, 4]]],
dummy=1)
conf['sources']=[{'name': 's0','events_per_day':105/5.,'data': data_calibration},
{'name': 's1','events_per_day':16. ,"data": data_other}]
likelihood_config = {'model_statistical_uncertainty_handling': 'bb_single',
'bb_single_source': 0}
lf = BinnedLogLikelihood(conf, likelihood_config=likelihood_config)
lf.add_shape_parameter('dummy', (0, 1))
lf.prepare()
assert lf.n_model_events is not None
# Make events:
instructions_mc = [dict(n_events=3, x=0.5),
dict(n_events=5, x=1.5),
dict(n_events=2, x=2.5),
dict(n_events=7, x=3.5),]
data, _ = make_data(instructions_mc)
lf.set_data(data)
assert lf.n_model_events is not None
A_minimized = np.array([14.24, 26.8070, 28.08, 26.21])
A_BB = beeston_barlow_root2(np.array([16, 30, 32, 27]), 0.2, np.array([5, 7, 1, 3]), np.array([3, 5, 2, 7]))
npt.assert_almost_equal(A_minimized,A_BB,decimal=2)
dbin = np.array([3, 5, 2, 7])
U_bin = np.array([5, 7, 1, 3])
print("A_BB*0.2: %s" % str(A_BB*0.2))
assert almost_equal(lf(), np.sum(stats.poisson(0.2*A_BB+U_bin).logpmf(dbin)))
def test_single_bin():
conf = test_conf(mc=True, analysis_space=[['x', [-40, 40]]])
lf = BinnedLogLikelihood(conf)
lf.add_rate_parameter('s0')
lf.prepare()
# Make a single event at x=0
lf.set_data(np.zeros(1, dtype=[('x', np.float), ('source', np.int)]))
assert lf() == stats.poisson(1000).logpmf(1)
assert lf(s0_rate_multiplier=5.4) == stats.poisson(5400).logpmf(1)
def test_twobin_mc():
conf = test_conf(mc=True, analysis_space=[['x', [-40, 0, 40]]])
lf = BinnedLogLikelihood(conf)
lf.add_rate_parameter('s0')
lf.prepare()
# Make 100 events at x=1
lf.set_data(np.ones(100, dtype=[('x', np.float), ('source', np.int)]))
assert almost_equal(
lf(),
stats.poisson(500).logpmf(100) + stats.poisson(500).logpmf(0), 1e-2)
def test_BeestonBarlowSingleBin():
instructions_mc = [dict(n_events=32, x=0.5)]
data, n_mc = make_data(instructions_mc)
conf = test_conf(default_source_class=FixedSampleSource,
events_per_day=32/5,
analysis_space=[['x', [0, 1]]],
data=data)
likelihood_config = {'model_statistical_uncertainty_handling': 'bb_single',
'bb_single_source': 0}
lf = BinnedLogLikelihood(conf, likelihood_config=likelihood_config)
lf.prepare()
assert lf.n_model_events is not None
# Make a single event at x=0
lf.set_data(np.zeros(2, dtype=[('x', np.float), ('source', np.int)]))
assert lf.n_model_events is not None
assert almost_equal(28.0814209, beeston_barlow_root2(np.array([32]), 0.2, np.array([1]), np.array([2])))
# A = beeston_barlow_root2(np.array([32]), 0.2, np.array([0]), np.array([2]))
A = (2+32)/(1+0.2)
assert almost_equal(lf(), stats.poisson(0.2*A).logpmf(2))
def test_multi_bin_single_dim():
instructions_mc = [dict(n_events=24, x=0.5), dict(n_events=56, x=1.5)]
data, n_mc = make_data(instructions_mc)
conf = test_conf(events_per_day=42,
analysis_space=[['x', [0, 1, 5]]],
default_source_class=FixedSampleSource,
data=data)
lf = BinnedLogLikelihood(conf)
lf.add_rate_parameter('s0')
instructions_data = [dict(n_events=18, x=0.5), dict(n_events=70, x=1.5)]
data, _ = make_data(instructions_data)
lf.set_data(data)
mus = [
42 / n_mc * instructions_mc[i]['n_events']
for i in range(len(instructions_mc))
]
seen = [
instructions_data[i]['n_events'] for i in range(len(instructions_data))
]
assert almost_equal(
lf(),
np.sum([
stats.poisson(mu).logpmf(seen_in_bin)
for mu, seen_in_bin in zip(mus, seen)
]), 1e-6)
def test_multi_bin():
instructions_mc = [dict(n_events=24, x=0.5, y=0.5),
dict(n_events=56, x=1.5, y=0.5),
dict(n_events=6, x=0.5, y=2),
dict(n_events=14, x=1.5, y=2)]
data, n_mc = make_data(instructions_mc)
conf = test_conf(events_per_day=42, default_source_class=FixedSampleSource, data=data,
analysis_space=[['x', [0, 1, 5]], ['y', [0, 1, 4]]])
lf = BinnedLogLikelihood(conf)
lf.add_rate_parameter('s0')
lf.add_shape_parameter('strlen_multiplier', {1: 'x', 2: 'hi', 3:'wha'},base_value=1)
lf.prepare()
instructions_data = [dict(n_events=18, x=0.5, y=0.5),
dict(n_events=70, x=1.5, y=0.5),
dict(n_events=4, x=0.5, y=2),
dict(n_events=10, x=1.5, y=2)]
data, _ = make_data(instructions_data)
lf.set_data(data)
mus = [42 / n_mc * instructions_mc[i]['n_events']
for i in range(len(instructions_mc))]
seen = [instructions_data[i]['n_events']
for i in range(len(instructions_data))]
assert almost_equal(lf(strlen_multiplier=1),
np.sum([stats.poisson(mu).logpmf(seen_in_bin)
for mu, seen_in_bin in zip(mus, seen)]))
print("Expected events: ", lf.base_model.expected_events())
with pytest.raises(NotImplementedError):
lf(compute_pdf=True,strlen_multiplier=2), np.sum([stats.poisson(2 * mu).logpmf(seen_in_bin) for mu, seen_in_bin in zip(mus, seen)])
assert almost_equal(lf(compute_pdf=False,strlen_multiplier=2),
np.sum([stats.poisson(2 * mu).logpmf(seen_in_bin)
for mu, seen_in_bin in zip(mus, seen)]))
assert almost_equal(lf(strlen_multiplier=2.3),
np.sum([stats.poisson(2.3*mu).logpmf(seen_in_bin)
for mu, seen_in_bin in zip(mus, seen)]))
def test_single_bin():
conf = test_conf(mc=True, analysis_space=[['x', [-40,40 ]]])
lf = BinnedLogLikelihood(conf)
lf.add_rate_parameter('s0')
lf.prepare()
# Make a single event at x=0
lf.set_data(np.zeros(1,
dtype=[('x', np.float), ('source', np.int)]))
assert lf() == stats.poisson(1000).logpmf(1)
assert lf(s0_rate_multiplier=5.4) == stats.poisson(5400).logpmf(1)
def test_twobin_mc():
conf = test_conf(mc=True, analysis_space=[['x', [-40, 0, 40]]])
lf = BinnedLogLikelihood(conf)
lf.add_rate_parameter('s0')
lf.prepare()
# Make 100 events at x=1
lf.set_data(np.ones(100,
dtype=[('x', np.float), ('source', np.int)]))
assert almost_equal(lf(),
stats.poisson(500).logpmf(100) + stats.poisson(500).logpmf(0),
1e-2)
def test_BeestonBarlowMultiBin():
instructions_mc = [dict(n_events=16, x=0.5),
dict(n_events=30, x=1.5),
dict(n_events=32, x=2.5),
dict(n_events=27, x=3.5),]
data, n_mc = make_data(instructions_mc)
conf = test_conf(default_source_class=FixedSampleSource,
events_per_day=105/5,
analysis_space=[['x', [0, 1, 2, 3, 4]]],
data=data)
likelihood_config = {'model_statistical_uncertainty_handling': 'bb_single',
'bb_single_source': 0}
lf = BinnedLogLikelihood(conf, likelihood_config=likelihood_config)
lf.prepare()
assert lf.n_model_events is not None
# Make events:
instructions_mc = [dict(n_events=3, x=0.5),
dict(n_events=5, x=1.5),
dict(n_events=2, x=2.5),
dict(n_events=7, x=3.5),]
data, _ = make_data(instructions_mc)
lf.set_data(data)
assert lf.n_model_events is not None
A_minimized = np.array([15.833,29.166,28.333,28.333])
A_BB = beeston_barlow_root2(np.array([16,30,32,27]), 0.2, np.array([0.]), np.array([3,5,2,7]))
npt.assert_almost_equal(A_minimized,A_BB,decimal=2);
dbin = np.array([3,5,2,7])
assert almost_equal(lf(), np.sum(stats.poisson(0.2*A_BB).logpmf(dbin)))
def test_multi_bin():
instructions_mc = [
dict(n_events=24, x=0.5, y=0.5),
dict(n_events=56, x=1.5, y=0.5),
dict(n_events=6, x=0.5, y=2),
dict(n_events=14, x=1.5, y=2)
]
data, n_mc = make_data(instructions_mc)
conf = test_conf(events_per_day=42,
default_source_class=FixedSampleSource,
data=data,
analysis_space=[['x', [0, 1, 5]], ['y', [0, 1, 4]]])
lf = BinnedLogLikelihood(conf)
lf.add_rate_parameter('s0')
lf.add_shape_parameter('strlen_multiplier', {
1: 'x',
2: 'hi',
3: 'wha'
},
base_value=1)
lf.prepare()
instructions_data = [
dict(n_events=18, x=0.5, y=0.5),
dict(n_events=70, x=1.5, y=0.5),
dict(n_events=4, x=0.5, y=2),
dict(n_events=10, x=1.5, y=2)
]
data, _ = make_data(instructions_data)
lf.set_data(data)
mus = [
42 / n_mc * instructions_mc[i]['n_events']
for i in range(len(instructions_mc))
]
seen = [
instructions_data[i]['n_events'] for i in range(len(instructions_data))
]
assert almost_equal(
lf(strlen_multiplier=1),
np.sum([
stats.poisson(mu).logpmf(seen_in_bin)
for mu, seen_in_bin in zip(mus, seen)
]))
print("Expected events: ", lf.base_model.expected_events())
with pytest.raises(NotImplementedError):
lf(compute_pdf=True, strlen_multiplier=2), np.sum([
stats.poisson(2 * mu).logpmf(seen_in_bin)
for mu, seen_in_bin in zip(mus, seen)
])
assert almost_equal(
lf(compute_pdf=False, strlen_multiplier=2),
np.sum([
stats.poisson(2 * mu).logpmf(seen_in_bin)
for mu, seen_in_bin in zip(mus, seen)
]))
assert almost_equal(
lf(strlen_multiplier=2.3),
np.sum([
stats.poisson(2.3 * mu).logpmf(seen_in_bin)
for mu, seen_in_bin in zip(mus, seen)
]))