def test_pdf_eval():
tf_sess = tf.Session()
backends = [
numpy_backend(poisson_from_normal=True),
pytorch_backend(),
tensorflow_backend(session=tf_sess),
mxnet_backend()
]
values = []
for b in backends:
pyhf.set_backend(b)
source = {
"binning": [2, -0.5, 1.5],
"bindata": {
"data": [120.0, 180.0],
"bkg": [100.0, 150.0],
"bkgsys_up": [102, 190],
"bkgsys_dn": [98, 100],
"sig": [30.0, 95.0]
}
}
spec = {
'channels': [{
'name':
'singlechannel',
'samples': [{
'name':
'signal',
'data':
source['bindata']['sig'],
'modifiers': [{
'name': 'mu',
'type': 'normfactor',
'data': None
}]
}, {
'name':
'background',
'data':
source['bindata']['bkg'],
'modifiers': [{
'name': 'bkg_norm',
'type': 'histosys',
'data': {
'lo_data': source['bindata']['bkgsys_dn'],
'hi_data': source['bindata']['bkgsys_up']
}
}]
}]
}]
}
pdf = pyhf.hfpdf(spec)
data = source['bindata']['data'] + pdf.config.auxdata
v1 = pdf.logpdf(pdf.config.suggested_init(), data)
values.append(pyhf.tensorlib.tolist(v1)[0])
assert np.std(values) < 1e-6
def test_import_prepHistFactory():
schema = json.load(open('validation/spec.json'))
parsed_xml = pyhf.readxml.parse('validation/xmlimport_input/config/example.xml',
'validation/xmlimport_input/')
# build the spec, strictly checks properties included
spec = {'channels': parsed_xml['channels']}
jsonschema.validate(spec, schema)
pdf = pyhf.hfpdf(spec, poiname='SigXsecOverSM')
data = [binvalue for k in pdf.spec['channels'] for binvalue
in parsed_xml['data'][k['name']]] + pdf.config.auxdata
channels = {channel['name'] for channel in pdf.spec['channels']}
samples = {channel['name']: [sample['name'] for sample in channel['samples']] for channel in pdf.spec['channels']}
assert data == [122.0, 112.0, 0, 0, 1.0, 1.0, 0.0]
###
### signal overallsys
### bkg1 overallsys (stat ignored)
### bkg2 stateror (2 bins)
### bkg2 overallsys
assert 'channel1' in channels
assert 'signal' in samples['channel1']
assert 'background1' in samples['channel1']
assert 'background2' in samples['channel1']
assert pdf.expected_actualdata(
pdf.config.suggested_init()).tolist() == [120.0, 110.0]
pars = pdf.config.suggested_init()
pars[pdf.config.par_slice('SigXsecOverSM')] = [2.0]
assert pdf.expected_data(
pars, include_auxdata=False).tolist() == [140, 120]
def test_pdf_integration_shapesys():
schema = json.load(open('validation/spec.json'))
source = json.load(open('validation/data/2bin_histosys_example2.json'))
spec = {
'channels': [{
'name':
'singlechannel',
'samples': [{
'name':
'signal',
'data':
source['bindata']['sig'],
'modifiers': [{
'name': 'mu',
'type': 'normfactor',
'data': None
}]
}, {
'name':
'background',
'data':
source['bindata']['bkg'],
'modifiers': [{
'name': 'bkg_norm',
'type': 'shapesys',
'data': [10, 10]
}]
}]
}]
}
jsonschema.validate(spec, schema)
pdf = pyhf.hfpdf(spec)
pars = [None, None]
pars[pdf.config.par_slice('mu')], pars[pdf.config.par_slice(
'bkg_norm')] = [[0.0], [1.0, 1.0]]
assert pdf.expected_data(pars,
include_auxdata=False).tolist() == [100, 150]
pars[pdf.config.par_slice('mu')], pars[pdf.config.par_slice(
'bkg_norm')] = [[0.0], [1.1, 1.0]]
assert pdf.expected_data(
pars, include_auxdata=False).tolist() == [100 * 1.1, 150]
pars[pdf.config.par_slice('mu')], pars[pdf.config.par_slice(
'bkg_norm')] = [[0.0], [1.0, 1.1]]
assert pdf.expected_data(
pars, include_auxdata=False).tolist() == [100, 150 * 1.1]
pars[pdf.config.par_slice('mu')], pars[pdf.config.par_slice(
'bkg_norm')] = [[0.0], [1.1, 0.9]]
assert pdf.expected_data(
pars, include_auxdata=False).tolist() == [100 * 1.1, 150 * 0.9]
pars[pdf.config.par_slice('mu')], pars[pdf.config.par_slice(
'bkg_norm')] = [[0.0], [0.9, 1.1]]
assert pdf.expected_data(
pars, include_auxdata=False).tolist() == [100 * 0.9, 150 * 1.1]
def test_pdf_integration_normsys(backend):
pyhf.set_backend(backend)
if isinstance(pyhf.tensorlib, pyhf.tensor.tensorflow_backend):
tf.reset_default_graph()
pyhf.tensorlib.session = tf.Session()
schema = json.load(open('validation/spec.json'))
source = json.load(open('validation/data/2bin_histosys_example2.json'))
spec = {
'channels': [{
'name':
'singlechannel',
'samples': [{
'name':
'signal',
'data':
source['bindata']['sig'],
'modifiers': [{
'name': 'mu',
'type': 'normfactor',
'data': None
}]
}, {
'name':
'background',
'data':
source['bindata']['bkg'],
'modifiers': [{
'name': 'bkg_norm',
'type': 'normsys',
'data': {
'lo': 0.9,
'hi': 1.1
}
}]
}]
}]
}
jsonschema.validate(spec, schema)
pdf = pyhf.hfpdf(spec)
pars = [None, None]
pars[pdf.config.par_slice('mu')], pars[pdf.config.par_slice(
'bkg_norm')] = [[0.0], [0.0]]
assert np.allclose(
pyhf.tensorlib.tolist(pdf.expected_data(pars, include_auxdata=False)),
[100, 150])
pars[pdf.config.par_slice('mu')], pars[pdf.config.par_slice(
'bkg_norm')] = [[0.0], [1.0]]
assert np.allclose(
pyhf.tensorlib.tolist(pdf.expected_data(pars, include_auxdata=False)),
[100 * 1.1, 150 * 1.1])
pars[pdf.config.par_slice('mu')], pars[pdf.config.par_slice(
'bkg_norm')] = [[0.0], [-1.0]]
assert np.allclose(
pyhf.tensorlib.tolist(pdf.expected_data(pars, include_auxdata=False)),
[100 * 0.9, 150 * 0.9])
def test_add_unknown_modifier():
spec = {
'channels': [{
'name':
'channe',
'samples': [
{
'modifiers': [{
'name': 'a_name',
'type': 'this_should_not_exist',
'data': None
}]
},
]
}]
}
with pytest.raises(pyhf.exceptions.InvalidModifier):
pyhf.hfpdf(spec)
def test_pdf_integration_staterror():
spec = {
'channels': [
{
'name':
'firstchannel',
'samples': [{
'name':
'mu',
'data': [10., 10.],
'modifiers': [{
'name': 'mu',
'type': 'normfactor',
'data': None
}]
}, {
'name':
'bkg1',
'data': [50.0, 70.0],
'modifiers': [{
'name': 'stat_firstchannel',
'type': 'staterror',
'data': [12., 12.]
}]
}, {
'name':
'bkg2',
'data': [30.0, 20.],
'modifiers': [{
'name': 'stat_firstchannel',
'type': 'staterror',
'data': [5., 5.]
}]
}, {
'name': 'bkg3',
'data': [20.0, 15.0],
'modifiers': []
}]
},
]
}
pdf = pyhf.hfpdf(spec)
par = pdf.config.par_slice('stat_firstchannel')
mod = pdf.config.modifier('stat_firstchannel')
assert mod.uncertainties == [[12., 12.], [5., 5.]]
assert mod.nominal_counts == [[50., 70.], [30., 20.]]
computed = pyhf.tensorlib.tolist(mod.pdf([1.0, 1.0], [1.0, 1.0]))
expected = pyhf.tensorlib.tolist(
pyhf.tensorlib.normal([1.0, 1.0],
mu=[1.0, 1.0],
sigma=[13. / 80., 13. / 90.]))
for c, e in zip(computed, expected):
assert c == e
def test_optim(source, spec, mu, backend):
pdf = pyhf.hfpdf(spec)
data = source['bindata']['data'] + pdf.config.auxdata
init_pars = pdf.config.suggested_init()
par_bounds = pdf.config.suggested_bounds()
pyhf.set_backend(backend)
optim = pyhf.optimizer
if isinstance(pyhf.tensorlib, pyhf.tensor.tensorflow_backend):
tf.reset_default_graph()
pyhf.tensorlib.session = tf.Session()
result = optim.unconstrained_bestfit(
pyhf.loglambdav, data, pdf, init_pars, par_bounds)
assert pyhf.tensorlib.tolist(result)
result = optim.constrained_bestfit(
pyhf.loglambdav, mu, data, pdf, init_pars, par_bounds)
assert pyhf.tensorlib.tolist(result)
def test_validation(setup):
source = setup['source']
pdf = pyhf.hfpdf(setup['spec'])
if 'channels' in source:
data = []
for c in pdf.spec['channels']:
data += source['channels'][c['name']]['bindata']['data']
data = data + pdf.config.auxdata
else:
data = source['bindata']['data'] + pdf.config.auxdata
if 'auxdata' in setup['expected']['config']:
assert len(pdf.config.auxdata) == \
setup['expected']['config']['auxdata']
assert len(pdf.config.suggested_init()) == \
setup['expected']['config']['init_pars']
assert len(pdf.config.suggested_bounds()) == \
setup['expected']['config']['par_bounds']
validate_runOnePoint(pdf, data, setup['mu'], setup['expected']['result'])
def test_import_histosys():
schema = json.load(open('validation/spec.json'))
parsed_xml = pyhf.readxml.parse(
'validation/xmlimport_input2/config/example.xml',
'validation/xmlimport_input2')
# build the spec, strictly checks properties included
spec = {'channels': parsed_xml['channels']}
jsonschema.validate(spec, schema)
pdf = pyhf.hfpdf(spec, poiname='SigXsecOverSM')
data = [
binvalue for k in pdf.spec['channels']
for binvalue in parsed_xml['data'][k['name']]
] + pdf.config.auxdata
channels = {channel['name']: channel for channel in pdf.spec['channels']}
samples = {
channel['name']: [sample['name'] for sample in channel['samples']]
for channel in pdf.spec['channels']
}
assert channels['channel2']['samples'][0]['modifiers'][0][
'type'] == 'histosys'
def test_pdf_integration_histosys():
schema = json.load(open('validation/spec.json'))
source = json.load(open('validation/data/2bin_histosys_example2.json'))
spec = {
'channels': [{
'name':
'singlechannel',
'samples': [{
'name':
'signal',
'data':
source['bindata']['sig'],
'modifiers': [{
'name': 'mu',
'type': 'normfactor',
'data': None
}]
}, {
'name':
'background',
'data':
source['bindata']['bkg'],
'modifiers': [{
'name': 'bkg_norm',
'type': 'histosys',
'data': {
'lo_data': source['bindata']['bkgsys_dn'],
'hi_data': source['bindata']['bkgsys_up']
}
}]
}]
}]
}
jsonschema.validate(spec, schema)
pdf = pyhf.hfpdf(spec)
pars = [None, None]
pars[pdf.config.par_slice('mu')], pars[pdf.config.par_slice(
'bkg_norm')] = [[0.0], [1.0]]
assert pdf.expected_data(pars,
include_auxdata=False).tolist() == [102, 190]
pars[pdf.config.par_slice('mu')], pars[pdf.config.par_slice(
'bkg_norm')] = [[0.0], [2.0]]
assert pdf.expected_data(pars,
include_auxdata=False).tolist() == [104, 230]
pars[pdf.config.par_slice('mu')], pars[pdf.config.par_slice(
'bkg_norm')] = [[0.0], [-1.0]]
assert pdf.expected_data(pars, include_auxdata=False).tolist() == [98, 100]
pars[pdf.config.par_slice('mu')], pars[pdf.config.par_slice(
'bkg_norm')] = [[0.0], [-2.0]]
assert pdf.expected_data(pars, include_auxdata=False).tolist() == [96, 50]
pars[pdf.config.par_slice('mu')], pars[pdf.config.par_slice(
'bkg_norm')] = [[1.0], [1.0]]
assert pdf.expected_data(
pars, include_auxdata=False).tolist() == [102 + 30, 190 + 95]
pars[pdf.config.par_slice('mu')], pars[pdf.config.par_slice(
'bkg_norm')] = [[1.0], [-1.0]]
assert pdf.expected_data(
pars, include_auxdata=False).tolist() == [98 + 30, 100 + 95]
