def test_calculate_std(example_spec, example_spec_multibin):
model = pyhf.Workspace(example_spec).model()
parameters = np.asarray([1.05, 0.95])
uncertainty = np.asarray([0.1, 0.1])
corr_mat = np.asarray([[1.0, 0.2], [0.2, 1.0]])
total_std = model_tools.calculate_std(model, parameters, uncertainty, corr_mat)
expected_std = [[17.4320561320614]]
assert np.allclose(ak.to_list(total_std), expected_std)
# pre-fit
parameters = np.asarray([1.0, 1.0])
uncertainty = np.asarray([0.0495665682, 0.0])
diag_corr_mat = np.diag([1.0, 1.0])
total_std = model_tools.calculate_std(model, parameters, uncertainty, diag_corr_mat)
expected_std = [[5.572758655480406]] # the staterror
assert np.allclose(ak.to_list(total_std), expected_std)
# multiple channels, bins, staterrors
model = pyhf.Workspace(example_spec_multibin).model()
parameters = np.asarray([0.9, 1.05, 1.3, 0.95])
uncertainty = np.asarray([0.1, 0.05, 0.3, 0.1])
corr_mat = np.asarray(
[
[1.0, 0.1, 0.2, 0.1],
[0.1, 1.0, 0.2, 0.3],
[0.2, 0.2, 1.0, 0.3],
[0.1, 0.3, 0.3, 1.0],
]
)
total_std = model_tools.calculate_std(model, parameters, uncertainty, corr_mat)
expected_std = [[12.889685799118613, 2.6730057987217317], [3.469221814759039]]
for i_reg in range(2):
assert np.allclose(ak.to_list(total_std[i_reg]), expected_std[i_reg])
def test_rename_outfile(tmpdir, script_runner):
temp = tmpdir.join("parsed_output.json")
command = 'pyhf xml2json validation/xmlimport_input/config/example.xml --basedir validation/xmlimport_input/ --output-file {0:s} --hide-progress'.format(
temp.strpath
)
ret = script_runner.run(*shlex.split(command))
tempout = tmpdir.join("rename_output.json")
command = 'pyhf rename -m staterror_channel1 staterror_channelone --measurement GammaExample GamEx {0:s} --output-file {1:s}'.format(
temp.strpath, tempout.strpath
)
ret = script_runner.run(*shlex.split(command))
assert ret.success
spec = json.loads(temp.read())
ws = pyhf.Workspace(spec)
assert 'GammaExample' in ws.measurement_names
assert 'GamEx' not in ws.measurement_names
assert 'staterror_channel1' in ws.parameters
assert 'staterror_channelone' not in ws.parameters
renamed_spec = json.loads(tempout.read())
renamed_ws = pyhf.Workspace(renamed_spec)
assert 'GammaExample' not in renamed_ws.measurement_names
assert 'GamEx' in renamed_ws.measurement_names
assert 'staterror_channel1' not in renamed_ws.parameters
assert 'staterror_channelone' in renamed_ws.parameters
def wsMaker(self):
"""
Apply each region patch (self.patches) to his associated json (self.inputJsons) to obtain the complete workspaces
:returns: the list of patched workspaces
"""
if self.patches == None:
return None
if self.nWS == 1:
try:
return [pyhf.Workspace(jsonpatch.apply_patch(self.inputJsons[0], self.patches[0]))]
except (pyhf.exceptions.InvalidSpecification, KeyError) as e:
logger.error("The json file is corrupted:\n{}".format(e))
return None
else:
workspaces = []
for js, patch in zip(self.inputJsons, self.patches):
wsDict = jsonpatch.apply_patch(js, patch)
try:
ws = pyhf.Workspace(wsDict)
except (pyhf.exceptions.InvalidSpecification, KeyError) as e:
logger.error("Json file number {} is corrupted:\n{}".format(self.inputJsons.index(json), e))
return None
workspaces.append(ws)
return workspaces
def test_model_and_data(example_spec):
model, data = model_tools.model_and_data(example_spec)
assert model.spec["channels"] == example_spec["channels"]
assert model.config.modifier_settings == {
"normsys": {"interpcode": "code4"},
"histosys": {"interpcode": "code4p"},
}
assert data == [691, 1.0]
# requesting Asimov dataset
model, data = model_tools.model_and_data(pyhf.Workspace(example_spec), asimov=True)
assert model.spec["channels"] == example_spec["channels"]
assert model.config.modifier_settings == {
"normsys": {"interpcode": "code4"},
"histosys": {"interpcode": "code4p"},
}
assert data == [112.429786, 1.0]
# test handing a workspace instead of JSON
model, data = model_tools.model_and_data(pyhf.Workspace(example_spec))
assert model.spec["channels"] == example_spec["channels"]
assert model.config.modifier_settings == {
"normsys": {"interpcode": "code4"},
"histosys": {"interpcode": "code4p"},
}
assert data == [691, 1.0]
# without auxdata
model, data = model_tools.model_and_data(example_spec, include_auxdata=False)
assert data == [691]
def test_calculate_stdev(example_spec, example_spec_multibin):
model = pyhf.Workspace(example_spec).model()
parameters = np.asarray([1.05, 0.95])
uncertainty = np.asarray([0.1, 0.1])
corr_mat = np.asarray([[1.0, 0.2], [0.2, 1.0]])
total_stdev = model_utils.calculate_stdev(model, parameters, uncertainty,
corr_mat)
expected_stdev = [[8.03767016]]
assert np.allclose(ak.to_list(total_stdev), expected_stdev)
# pre-fit
parameters = np.asarray([1.0, 1.0])
uncertainty = np.asarray([0.0495665682, 0.0])
diag_corr_mat = np.diag([1.0, 1.0])
total_stdev = model_utils.calculate_stdev(model, parameters, uncertainty,
diag_corr_mat)
expected_stdev = [[2.56951880]] # the staterror
assert np.allclose(ak.to_list(total_stdev), expected_stdev)
# multiple channels, bins, staterrors
model = pyhf.Workspace(example_spec_multibin).model()
parameters = np.asarray([0.9, 1.05, 1.3, 0.95])
uncertainty = np.asarray([0.1, 0.05, 0.3, 0.1])
corr_mat = np.asarray([
[1.0, 0.1, 0.2, 0.1],
[0.1, 1.0, 0.2, 0.3],
[0.2, 0.2, 1.0, 0.3],
[0.1, 0.3, 0.3, 1.0],
])
total_stdev = model_utils.calculate_stdev(model, parameters, uncertainty,
corr_mat)
expected_stdev = [[8.056054, 1.670629], [2.775377]]
for i_reg in range(2):
assert np.allclose(total_stdev[i_reg], expected_stdev[i_reg])
def test_get_prefit_uncertainties(example_spec, example_spec_multibin,
example_spec_shapefactor):
model = pyhf.Workspace(example_spec).model()
uncertainties = model_tools.get_prefit_uncertainties(model)
assert np.allclose(uncertainties, [0.0, 0.0])
model = pyhf.Workspace(example_spec_multibin).model()
uncertainties = model_tools.get_prefit_uncertainties(model)
assert np.allclose(uncertainties, [0.175, 0.375, 0.0, 0.2])
model = pyhf.Workspace(example_spec_shapefactor).model()
uncertainties = model_tools.get_prefit_uncertainties(model)
assert np.allclose(uncertainties, [0.0, 0.0, 0.0])
def test_get_prefit_uncertainties(example_spec, example_spec_multibin,
example_spec_shapefactor):
model = pyhf.Workspace(example_spec).model()
unc = model_utils.get_prefit_uncertainties(model)
assert np.allclose(unc, [0.0, 0.0]) # fixed parameter and normfactor
model = pyhf.Workspace(example_spec_multibin).model()
unc = model_utils.get_prefit_uncertainties(model)
assert np.allclose(unc, [0.2, 0.4, 0.0, 0.125])
model = pyhf.Workspace(example_spec_shapefactor).model()
unc = model_utils.get_prefit_uncertainties(model)
assert np.allclose(unc, [0.0, 0.0, 0.0])
def test_unconstrained_parameter_count(example_spec, example_spec_shapefactor):
model = pyhf.Workspace(example_spec).model()
assert model_utils.unconstrained_parameter_count(model) == 1
model = pyhf.Workspace(example_spec_shapefactor).model()
assert model_utils.unconstrained_parameter_count(model) == 3
# fixed parameters are skipped in counting
example_spec_shapefactor["measurements"][0]["config"]["parameters"].append(
{
"name": "Signal strength",
"fixed": True
})
model = pyhf.Workspace(example_spec_shapefactor).model()
assert model_utils.unconstrained_parameter_count(model) == 2
def run_bkg(self):
import json
import copy
import pyhf
with open('{region}/BkgOnly.json'.format(region=self.region)) as fname:
spec = json.load(fname)
spec = copy.copy(spec)
#switch the parameter of interest to the lumi from any signal strength
spec["measurements"][0]["config"]["poi"] = "lumi"
#load the workspace
ws = pyhf.Workspace(spec)
#load the model
model = ws.model(
measurement_name="NormalMeasurement",
modifier_settings={
"normsys": {"interpcode": "code4"},
"histosys": {"interpcode": "code4p"},
},
)
data = ws.data(model)
self.model = model
self.data = data
self.next(self.fit_bkg)
def test_wspace_unexpected_keyword_argument(simplemodels_model_data):
model, data = simplemodels_model_data
workspace = pyhf.Workspace.build(model, data)
spec = dict(workspace)
with pytest.raises(pyhf.exceptions.Unsupported):
pyhf.Workspace(spec, abc=True)
def model_and_data(
spec: Dict[str, Any],
asimov: bool = False,
with_aux: bool = True) -> Tuple[pyhf.pdf.Model, List[float]]:
"""Returns model and data for a ``pyhf`` workspace specification.
Args:
spec (Dict[str, Any]): a ``pyhf`` workspace specification
asimov (bool, optional): whether to return the Asimov dataset, defaults to False
with_aux (bool, optional): whether to also return auxdata, defaults to True
Returns:
Tuple[pyhf.pdf.Model, List[float]]:
- a HistFactory-style model in ``pyhf`` format
- the data (plus auxdata if requested) for the model
"""
workspace = pyhf.Workspace(spec)
model = workspace.model(modifier_settings={
"normsys": {
"interpcode": "code4"
},
"histosys": {
"interpcode": "code4p"
},
}) # use HistFactory InterpCode=4
if not asimov:
data = workspace.data(model, with_aux=with_aux)
else:
data = build_Asimov_data(model, with_aux=with_aux)
return model, data
def validate(ws: Dict[str, Any]) -> None:
"""Validates a workspace with ``pyhf``.
Args:
ws (Dict[str, Any]): the workspace to validate
"""
pyhf.Workspace(ws)
def validate(ws):
"""validate a workspace
Args:
ws (dict): pyhf-compatible HistFactory workspace
"""
pyhf.Workspace(ws)
def test_combine_outfile(tmpdir, script_runner):
temp_1 = tmpdir.join("parsed_output.json")
temp_2 = tmpdir.join("renamed_output.json")
command = f'pyhf xml2json validation/xmlimport_input/config/example.xml --basedir validation/xmlimport_input/ --output-file {temp_1.strpath:s} --hide-progress'
ret = script_runner.run(*shlex.split(command))
rename_channels = {'channel1': 'channel2'}
rename_measurements = {
'ConstExample': 'OtherConstExample',
'LogNormExample': 'OtherLogNormExample',
'GaussExample': 'OtherGaussExample',
'GammaExample': 'OtherGammaExample',
}
_opts_channels = ''.join(
' -c ' + ' '.join(item) for item in rename_channels.items()
)
_opts_measurements = ''.join(
' --measurement ' + ' '.join(item) for item in rename_measurements.items()
)
command = f"pyhf rename {temp_1.strpath:s} {_opts_channels:s} {_opts_measurements:s} --output-file {temp_2.strpath:s}"
ret = script_runner.run(*shlex.split(command))
tempout = tmpdir.join("combined_output.json")
command = f'pyhf combine {temp_1.strpath:s} {temp_2.strpath:s} --output-file {tempout.strpath:s}'
ret = script_runner.run(*shlex.split(command))
assert ret.success
combined_spec = json.loads(tempout.read())
combined_ws = pyhf.Workspace(combined_spec)
assert combined_ws.channels == ['channel1', 'channel2']
assert len(combined_ws.measurement_names) == 8
def test_model_and_data(example_spec):
model, data = model_tools.model_and_data(example_spec)
assert model.spec["channels"] == example_spec["channels"]
assert model.config.modifier_settings == {
"normsys": {
"interpcode": "code4"
},
"histosys": {
"interpcode": "code4p"
},
}
assert data == [691, 1.0]
# requesting Asimov dataset
# TODO: request asimov dataset by setting asimove=True
# TODO: should return [112.429786, 1.0]
# TODO: Need to test overloaded method as well
# test handing a workspace instead of JSON
model, data = model_tools.model_and_data(pyhf.Workspace(example_spec))
assert model.spec["channels"] == example_spec["channels"]
assert model.config.modifier_settings == {
"normsys": {
"interpcode": "code4"
},
"histosys": {
"interpcode": "code4p"
},
}
assert data == [691, 1.0]
# without auxdata
model, data = model_tools.model_and_data(example_spec, with_aux=False)
assert data == [691]
def test_patchset_apply(datadir):
patchset = pyhf.PatchSet(
json.load(open(datadir.join('example_patchset.json'))))
ws = pyhf.Workspace(json.load(open(datadir.join('example_bkgonly.json'))))
with mock.patch('pyhf.patchset.PatchSet.verify') as m:
assert m.call_count == 0
assert patchset.apply(ws, 'patch_channel1_signal_syst1')
assert m.call_count == 1
def test_build_Asimov_data(example_spec):
ws = pyhf.Workspace(example_spec)
model = ws.model()
assert model_utils.build_Asimov_data(model) == [51.839756, 1]
# without auxdata
assert model_utils.build_Asimov_data(model, with_aux=False) == [51.839756]
# respect nominal settings for normfactors
example_spec["measurements"][0]["config"]["parameters"].append({
"name":
"Signal strength",
"inits": [2.0]
})
ws = pyhf.Workspace(example_spec)
model = ws.model()
assert model_utils.build_Asimov_data(model, with_aux=False) == [103.679512]
def test_workspace_without_validation(mocker, simplemodels_model_data):
model, data = simplemodels_model_data
mocker.patch('pyhf.utils.validate')
ws = pyhf.Workspace.build(model, data, validate=False)
assert pyhf.utils.validate.called is False
pyhf.Workspace(dict(ws), validate=False)
assert pyhf.utils.validate.called is False
def test_get_asimov_parameters(example_spec, example_spec_shapefactor):
model = pyhf.Workspace(example_spec).model()
pars = model_utils.get_asimov_parameters(model)
assert np.allclose(pars, [1.0, 1.0])
model = pyhf.Workspace(example_spec_shapefactor).model()
pars = model_utils.get_asimov_parameters(model)
assert np.allclose(pars, [1.0, 1.0, 1.0])
# respect nominal settings for normfactors
example_spec["measurements"][0]["config"]["parameters"].append({
"name":
"Signal strength",
"inits": [2.0]
})
model = pyhf.Workspace(example_spec).model()
pars = model_utils.get_asimov_parameters(model)
assert np.allclose(pars, [1.0, 2.0])
def compare_nuisance(root_workspace, pyhf_json):
# Get the root nuisance params
infile = ROOT.TFile.Open(root_workspace)
workspace = infile.Get("combined")
mc = workspace.obj("ModelConfig")
def exhaust_argset(s):
it = s.fwdIterator()
while True:
n = it.next()
if not n:
break
yield n
pars = [
x.GetName() for x in exhaust_argset(mc.GetNuisanceParameters())
] + [x.GetName() for x in exhaust_argset(mc.GetParametersOfInterest())]
# Replace some strings to match root nuisance param names to pyhf naming scheme
pars_root = [
sub.replace("alpha_", "").replace("gamma_stat_", "staterror_").replace(
"gamma_stat_", "staterror_").replace("lumi",
"Lumi").replace("_bin", "")
for sub in pars
]
# Get pyhf nuisance params
ws = pyhf.Workspace(json.load(open(pyhf_json)))
model = ws.model()
pars_pyhf = []
for k, v in model.config.par_map.items():
sl = v["slice"]
npars = sl.stop - sl.start
if npars > 1 or "staterror" in k:
for i in range(npars):
pars_pyhf.append(f"{k}_{i}")
else:
pars_pyhf.append(k)
# Compare the nuisance params
nuisance_dict = {"root": pars_root, "pyhf": pars_pyhf}
unique_dict = {"root": [], "pyhf": []}
unique_dict["pyhf"] = set(nuisance_dict["pyhf"]) - set(
nuisance_dict["root"])
unique_dict["root"] = set(nuisance_dict["root"]) - set(
nuisance_dict["pyhf"])
print("Nuisance params unique to pyhf:")
for param in unique_dict["pyhf"]:
print(param)
print("\nNuisance params unique to root:")
for param in unique_dict["root"]:
print(param)
def test_prune_outfile(tmpdir, script_runner):
temp = tmpdir.join("parsed_output.json")
command = f'pyhf xml2json validation/xmlimport_input/config/example.xml --basedir validation/xmlimport_input/ --output-file {temp.strpath:s} --hide-progress'
ret = script_runner.run(*shlex.split(command))
tempout = tmpdir.join("prune_output.json")
command = f'pyhf prune -m staterror_channel1 --measurement GammaExample {temp.strpath:s} --output-file {tempout.strpath:s}'
ret = script_runner.run(*shlex.split(command))
assert ret.success
spec = json.loads(temp.read())
ws = pyhf.Workspace(spec)
assert 'GammaExample' in ws.measurement_names
assert 'staterror_channel1' in ws.parameters
pruned_spec = json.loads(tempout.read())
pruned_ws = pyhf.Workspace(pruned_spec)
assert 'GammaExample' not in pruned_ws.measurement_names
assert 'staterror_channel1' not in pruned_ws.parameters
def test_workspace_poiless(datadir):
"""
Test that a workspace with a measurement with empty POI string is treated as POI-less
"""
spec = json.load(open(datadir.join("poiless.json")))
ws = pyhf.Workspace(spec)
model = ws.model()
assert model.config.poi_name is None
assert model.config.poi_index is None
def test__get_channel_bounds_indices(example_spec, example_spec_multibin):
model = pyhf.Workspace(example_spec).model()
indices = model_tools._get_channel_bounds_indices(model)
assert indices == []
model = pyhf.Workspace(example_spec_multibin).model()
indices = model_tools._get_channel_bounds_indices(model)
assert indices == [2]
# add extra channel to model to test three channels (two indices needed)
three_channel_model = copy.deepcopy(example_spec_multibin)
extra_channel = copy.deepcopy(three_channel_model["channels"][0])
extra_channel["name"] = "region_3"
extra_channel["samples"][0]["modifiers"][0]["name"] = "staterror_region_3"
three_channel_model["channels"].append(extra_channel)
three_channel_model["observations"].append({"data": [35, 8], "name": "region_3"})
model = pyhf.Workspace(three_channel_model).model()
indices = model_tools._get_channel_bounds_indices(model)
assert indices == [2, 3]
def test__yields_per_bin(example_spec_multibin, example_spec_with_background):
# multiple channels
model = pyhf.Workspace(example_spec_multibin).model()
yields = [[[25.0, 5.0]], [[8.0]]]
total_stdev = [[5.0, 2.0], [1.0]]
data = [[35, 8], [10]]
yield_table = tabulate._yields_per_bin(model, yields, total_stdev, data)
assert yield_table == [
{
"sample": "Signal",
"region_1\nbin 1": "25.00",
"region_1\nbin 2": "5.00",
"region_2\nbin 1": "8.00",
},
{
"sample": "total",
"region_1\nbin 1": "25.00 \u00B1 5.00",
"region_1\nbin 2": "5.00 \u00B1 2.00",
"region_2\nbin 1": "8.00 \u00B1 1.00",
},
{
"sample": "data",
"region_1\nbin 1": "35.00",
"region_1\nbin 2": "8.00",
"region_2\nbin 1": "10.00",
},
]
# multiple samples
model = pyhf.Workspace(example_spec_with_background).model()
yields = [[[150.0], [50.0]]]
total_stdev = [[8.60]]
data = [[160]]
yield_table = tabulate._yields_per_bin(model, yields, total_stdev, data)
assert yield_table == [
{"sample": "Background", "Signal Region\nbin 1": "150.00"},
{"sample": "Signal", "Signal Region\nbin 1": "50.00"},
{"sample": "total", "Signal Region\nbin 1": "200.00 \u00B1 8.60"},
{"sample": "data", "Signal Region\nbin 1": "160.00"},
]
def test__yields_per_channel(example_spec_multibin, example_spec_with_background):
# multiple channels
model = pyhf.Workspace(example_spec_multibin).model()
yields = [[30], [8.0]]
total_stdev = [5.39, 1.0]
data = [43, 10]
yield_table = tabulate._yields_per_channel(model, yields, total_stdev, data)
assert yield_table == [
{
"sample": "Signal",
"region_1": "30.00",
"region_2": "8.00",
},
{
"sample": "total",
"region_1": "30.00 \u00B1 5.39",
"region_2": "8.00 \u00B1 1.00",
},
{
"sample": "data",
"region_1": "43.00",
"region_2": "10.00",
},
]
# multiple samples
model = pyhf.Workspace(example_spec_with_background).model()
yields = [[150.0, 50]]
total_stdev = [8.60]
data = [160]
yield_table = tabulate._yields_per_channel(model, yields, total_stdev, data)
assert yield_table == [
{"sample": "Background", "Signal Region": "150.00"},
{"sample": "Signal", "Signal Region": "50.00"},
{"sample": "total", "Signal Region": "200.00 \u00B1 8.60"},
{"sample": "data", "Signal Region": "160.00"},
]
def test_wspace_immutable(simplemodels_model_data):
model, data = simplemodels_model_data
workspace = pyhf.Workspace.build(model, data)
spec = dict(workspace)
ws = pyhf.Workspace(spec)
model = ws.model()
before = model.config.suggested_init()
spec["measurements"][0]["config"]["parameters"][0]["inits"] = [1.5]
model = ws.model()
after = model.config.suggested_init()
assert before == after
def fit(spec):
workspace = pyhf.Workspace(spec)
model = workspace.model()
data = workspace.data(model)
pyhf.set_backend("numpy", pyhf.optimize.minuit_optimizer(verbose=True))
result = pyhf.infer.mle.fit(data, model, return_uncertainties=True)
bestfit = result[:, 0]
uncertainty = result[:, 1]
labels = get_parameter_names(model)
print_results(bestfit, uncertainty, labels)
return bestfit, uncertainty, labels
def test_calculate_stdev(example_spec, example_spec_multibin):
model = pyhf.Workspace(example_spec).model()
parameters = np.asarray([1.05, 0.95])
uncertainty = np.asarray([0.1, 0.1])
corr_mat = np.asarray([[1.0, 0.2], [0.2, 1.0]])
total_stdev_bin, total_stdev_chan = model_utils.calculate_stdev(
model, parameters, uncertainty, corr_mat)
assert np.allclose(total_stdev_bin, [[8.03150606]])
assert np.allclose(total_stdev_chan, [8.03150606])
# pre-fit
parameters = np.asarray([1.0, 1.0])
uncertainty = np.asarray([0.0495665682, 0.0])
diag_corr_mat = np.diag([1.0, 1.0])
total_stdev_bin, total_stdev_chan = model_utils.calculate_stdev(
model, parameters, uncertainty, diag_corr_mat)
assert np.allclose(total_stdev_bin, [[2.56754823]]) # the staterror
assert np.allclose(total_stdev_chan, [2.56754823])
# multiple channels, bins, staterrors
model = pyhf.Workspace(example_spec_multibin).model()
parameters = np.asarray([0.9, 1.05, 1.3, 0.95])
uncertainty = np.asarray([0.1, 0.05, 0.3, 0.1])
corr_mat = np.asarray([
[1.0, 0.1, 0.2, 0.1],
[0.1, 1.0, 0.2, 0.3],
[0.2, 0.2, 1.0, 0.3],
[0.1, 0.3, 0.3, 1.0],
])
total_stdev_bin, total_stdev_chan = model_utils.calculate_stdev(
model, parameters, uncertainty, corr_mat)
expected_stdev_bin = [[8.056054, 1.670629], [2.775377]]
expected_stdev_chan = [9.585327, 2.775377]
for i_reg in range(2):
assert np.allclose(total_stdev_bin[i_reg], expected_stdev_bin[i_reg])
assert np.allclose(total_stdev_chan[i_reg], expected_stdev_chan[i_reg])
def test_import_normfactor_bounds():
parsed_xml = pyhf.readxml.parse(
'validation/xmlimport_input2/config/example.xml',
'validation/xmlimport_input2')
ws = pyhf.Workspace(parsed_xml)
assert ('SigXsecOverSM', 'normfactor') in ws.modifiers
parameters = [
p for p in ws.get_measurement(
measurement_name='GaussExample')['config']['parameters']
if p['name'] == 'SigXsecOverSM'
]
assert len(parameters) == 1
parameter = parameters[0]
assert parameter['bounds'] == [[0, 10]]
def test_reproducible_model_spec():
ws = {
"channels": [{
"name":
"SR",
"samples": [{
"data": [
10.0,
],
"modifiers": [
{
"data": None,
"name": "mu",
"type": "normfactor"
},
],
"name":
"Signal",
}],
}],
"measurements": [{
"config": {
"parameters": [{
"bounds": [[0, 5]],
"inits": [1],
"name": "mu"
}],
"poi": "mu",
},
"name": "minimal_example",
}],
"observations": [{
"data": [12],
"name": "SR"
}],
"version":
"1.0.0",
}
workspace = pyhf.Workspace(ws)
model_from_ws = workspace.model()
assert model_from_ws.spec['parameters'] == [{
'bounds': [[0, 5]],
'inits': [1],
'name': 'mu'
}]
assert pyhf.Model(model_from_ws.spec)
