"""
.. module:: theoryPrediction
:synopsis: Provides a class to encapsulate the results of the computation of
reference cross sections and related functions.
.. moduleauthor:: Andre Lessa <lessa.a.p@gmail.com>
.. autofunction:: _getElementsFrom
"""
from smodels.theory import clusterTools, crossSection, element
from smodels.theory.particleNames import elementsInStr
from smodels.theory.auxiliaryFunctions import cSim, cGtr
from smodels.tools.physicsUnits import TeV,fb
from smodels.theory.exceptions import SModelSTheoryError as SModelSError
from smodels.experiment.datasetObj import CombinedDataSet
from smodels.tools.smodelsLogging import logger
[docs]class TheoryPrediction(object):
"""
An instance of this class represents the results of the theory prediction
for an analysis.
:ivar analysis: holds the analysis (ULanalysis or EManalysis object)
to which the prediction refers
:ivar xsection: xsection of the theory prediction
(relevant cross section to be compared with the experimental limits).
For EM-type analyses, it corresponds to sigma*eff, for
UL-type analyses, eff is considered to be 1.
It is a XSection object.
:ivar conditions: list of values for the analysis conditions
(only for upper limit-type analysis, e.g. analysis=ULanalysis)
:ivar mass: mass of the cluster to which the theory prediction refers
(only for upper limit-type analysis, e.g. analysis=ULanalysis)
"""
def __init__(self):
self.analysis = None
self.xsection = None
self.conditions = None
self.mass = None
[docs] def dataId(self):
"""
Return ID of dataset
"""
return self.dataset.getID()
[docs] def analysisId(self):
"""
Return experimental analysis ID
"""
return self.dataset.globalInfo.id
[docs] def dataType(self):
"""
Return the type of dataset
"""
return self.dataset.getType()
[docs] def getUpperLimit(self, expected=False, deltas_rel=0.2):
"""
Get the upper limit on sigma*eff.
For UL-type results, use the UL map. For EM-Type returns
the corresponding dataset (signal region) upper limit.
For combined results, returns the upper limit on the
total sigma*eff (for all signal regions/datasets).
:param expected: return expected Upper Limit, instead of observed.
:param deltas_rel: relative uncertainty in signal (float). Default value is 20%.
:return: upper limit (Unum object)
"""
#First check if the upper-limit and expected upper-limit have already been computed.
#If not, compute it and store them.
if not hasattr(self, 'expectedUL') or not hasattr(self, 'upperLimit'):
if self.dataType() == 'efficiencyMap':
self.expectedUL = self.dataset.getSRUpperLimit(expected=True,deltas_rel=deltas_rel)
self.upperLimit = self.dataset.getSRUpperLimit(expected=False,deltas_rel=deltas_rel)
if self.dataType() == 'upperLimit':
self.expectedUL = self.dataset.getUpperLimitFor(mass=self.mass,
txnames=self.txnames,
expected=True)
self.upperLimit = self.dataset.getUpperLimitFor(mass=self.mass,
txnames=self.txnames,
expected=False)
if self.dataType() == 'combined':
lumi = self.expResult.globalInfo.lumi
#Create a list of signal events in each dataset/SR sorted according to datasetOrder
srNsigDict = dict([[pred.dataset.getID(),(pred.xsection.value*lumi).asNumber()] for pred in self.datasetPredictions])
srNsigs = [srNsigDict[dataID] if dataID in srNsigDict else 0. for dataID in self.dataset.globalInfo.datasetOrder]
self.expectedUL = self.dataset.getCombinedUpperLimitFor(srNsigs,expected=True,deltas_rel=deltas_rel)
self.upperLimit = self.dataset.getCombinedUpperLimitFor(srNsigs,expected=False,deltas_rel=deltas_rel)
#Return the expected or observed UL:
if expected:
return self.expectedUL
else:
return self.upperLimit
[docs] def getRValue(self, expected = False):
"""
Get the r value = theory prediction / experimental upper limit
"""
upperLimit = self.getUpperLimit(expected)
if upperLimit is None or upperLimit.asNumber(fb)==0.:
return None
else:
return (self.xsection.value/upperLimit).asNumber()
[docs] def computeStatistics(self,marginalize=False,deltas_rel=0.2):
"""
Compute the likelihood, chi2 and expected upper limit for this theory prediction.
The resulting values are stored as the likelihood and chi2
attributes.
:param marginalize: if true, marginalize nuisances. Else, profile them.
:param deltas_rel: relative uncertainty in signal (float). Default value is 20%.
"""
if self.dataType() == 'upperLimit':
self.likelihood = None
self.chi2 = None
elif self.dataType() == 'efficiencyMap':
lumi = self.dataset.globalInfo.lumi
nsig = (self.xsection.value*lumi).asNumber()
llhd = self.dataset.likelihood(nsig,marginalize=marginalize,deltas_rel=deltas_rel)
chi2 = self.dataset.chi2(nsig,marginalize=marginalize,deltas_rel=deltas_rel)
self.likelihood = llhd
self.chi2 = chi2
elif self.dataType() == 'combined':
lumi = self.expResult.globalInfo.lumi
#Create a list of signal events in each dataset/SR sorted according to datasetOrder
srNsigDict = dict([[pred.dataset.getID(),(pred.xsection.value*lumi).asNumber()] for pred in self.datasetPredictions])
srNsigs = [srNsigDict[dataID] if dataID in srNsigDict else 0. for dataID in self.dataset.globalInfo.datasetOrder]
self.likelihood = self.dataset.combinedLikelihood(srNsigs, marginalize=marginalize,deltas_rel=deltas_rel)
self.chi2 = self.dataset.totalChi2(srNsigs, marginalize=marginalize,deltas_rel=deltas_rel)
[docs] def getmaxCondition(self):
"""
Returns the maximum xsection from the list conditions
:returns: maximum condition xsection (float)
"""
if not self.conditions: return 0.
# maxcond = 0.
values = [ 0. ]
for value in self.conditions.values():
if value == 'N/A': return value
if value == None: continue
#print ( "value=",value,type(value),float(value) )
#maxcond = max(maxcond,float(value))
values.append ( float(value) )
return max(values)
# return maxcond
def __str__(self):
ret = "%s:%s" % ( self.analysisId(), self.xsection )
#self.computeStatistics()
#ret += ", llhd=%f." % self.likelihood
return ret
[docs] def describe ( self ):
if not hasattr ( self, "chi2" ):
self.computeStatistics()
# return a lengthy description
ret = "[theoryPrediction] analysis: %s\n" % self.analysisId()
ret += " prediction: %s\n" % self.xsection
# ret += " prediction (sigma): %s\n" % ( self.xsection.value / self.effectiveEff )
# ret += " effective efficiency: %s\n" % self.effectiveEff
ds = "None"
if type (self.dataset) == list:
ds = "multiple (%d combined)" % len(self.dataset)
else:
dataId = self.dataset.getID()
fname="" ## folder name
if not "combined" in dataId:
fname=" (%s)" % self.dataset.folderName()
ds = "%s%s" % ( dataId, fname )
ret += " datasets: %s\n" % ds
ret += " obs limit: %s\n" % self.getUpperLimit()
ret += " exp limit: %s\n" % self.getUpperLimit( expected=True )
ret += " obs r: %f\n" % ( self.getRValue(expected=False) )
ret += " exp r: %f\n" % ( self.getRValue(expected=True) )
ret += " chi2: %s\n" % ( self.chi2 )
return ret
[docs]class TheoryPredictionList(object):
"""
An instance of this class represents a collection of theory prediction
objects.
:ivar _theoryPredictions: list of TheoryPrediction objects
"""
def __init__(self, theoryPredictions=None):
"""
Initializes the list.
:parameter theoryPredictions: list of TheoryPrediction objects
"""
self._theoryPredictions = []
if theoryPredictions and isinstance(theoryPredictions,list):
self._theoryPredictions = theoryPredictions
[docs] def append(self,theoryPred):
self._theoryPredictions.append ( theoryPred )
def __str__(self):
if len ( self._theoryPredictions ) == 0:
return "no predictions."
ret = "%d predictions: " % len ( self._theoryPredictions )
ret += ", ".join ( [ str(s) for s in self._theoryPredictions ] )
return ret
def __iter__(self):
for theoryPrediction in self._theoryPredictions:
yield theoryPrediction
def __getitem__(self, index):
return self._theoryPredictions[index]
def __len__(self):
return len(self._theoryPredictions)
def __add__(self,theoPredList):
if isinstance(theoPredList,TheoryPredictionList):
res = TheoryPredictionList()
res._theoryPredictions = self._theoryPredictions + theoPredList._theoryPredictions
return res
else:
return None
def __radd__(self, theoPredList):
if theoPredList == 0:
return self
else:
return self.__add__(theoPredList)
[docs]def theoryPredictionsFor(expResult, smsTopList, maxMassDist=0.2,
useBestDataset=True, combinedResults=True,
marginalize=False,deltas_rel=0.2):
"""
Compute theory predictions for the given experimental result, using the list of
elements in smsTopList.
For each Txname appearing in expResult, it collects the elements and
efficiencies, combine the masses (if needed) and compute the conditions
(if exist).
:parameter expResult: expResult to be considered (ExpResult object)
:parameter smsTopList: list of topologies containing elements
(TopologyList object)
:parameter maxMassDist: maximum mass distance for clustering elements (float)
:parameter useBestDataset: If True, uses only the best dataset (signal region).
If False, returns predictions for all datasets (if combinedResults is False),
or only the combinedResults (if combinedResults is True).
:parameter combinedResults: add theory predictions that result from
combining datasets.
:parameter marginalize: If true, marginalize nuisances. If false, profile them.
:parameter deltas_rel: relative uncertainty in signal (float). Default value is 20%.
:returns: a TheoryPredictionList object containing a list of TheoryPrediction
objects
"""
dataSetResults = []
#Compute predictions for each data set (for UL analyses there is one single set)
for dataset in expResult.datasets:
predList = _getDataSetPredictions(dataset,smsTopList,maxMassDist)
if predList:
dataSetResults.append(predList)
if not dataSetResults: ## no results at all?
return None
elif len(dataSetResults) == 1: ## only a single dataset? Easy case.
result = dataSetResults[0]
for theoPred in result:
theoPred.expResult = expResult
theoPred.upperLimit = theoPred.getUpperLimit(deltas_rel=deltas_rel)
return result
#For results with more than one dataset, return all dataset predictions
#if useBestDataSet=False and combinedResults=False:
if not useBestDataset and not combinedResults:
allResults = sum(dataSetResults)
for theoPred in allResults:
theoPred.expResult = expResult
theoPred.upperLimit = theoPred.getUpperLimit(deltas_rel=deltas_rel)
return allResults
#Else include best signal region results, if asked for.
bestResults = TheoryPredictionList()
if useBestDataset:
bestResults.append(_getBestResult(dataSetResults))
#If combinedResults = True, also include the combined result (when available):
if combinedResults and len(dataSetResults) > 1:
combinedDataSetResult = _getCombinedResultFor(dataSetResults,
expResult,marginalize)
if combinedDataSetResult:
bestResults.append(combinedDataSetResult)
for theoPred in bestResults:
theoPred.expResult = expResult
theoPred.upperLimit = theoPred.getUpperLimit(deltas_rel=deltas_rel)
return bestResults
def _getCombinedResultFor(dataSetResults,expResult,marginalize=False):
"""
Compute the compbined result for all datasets, if covariance
matrices are available. Return a TheoryPrediction object
with the signal cross-section summed over all the signal regions
and the respective upper limit.
:param datasetPredictions: List of TheoryPrediction objects for each signal region
:param expResult: ExpResult object corresponding to the experimental result
:parameter marginalize: If true, marginalize nuisances. If false, profile them.
:return: TheoryPrediction object
"""
if len(dataSetResults) == 1:
return dataSetResults[0]
elif not expResult.hasCovarianceMatrix():
return None
txnameList = []
elementList = []
totalXsec = None
massList = []
PIDList = []
IDList = []
datasetPredictions = []
for predList in dataSetResults:
if len(predList) != 1:
raise SModelSError("Results with multiple datasets should have a single theory prediction (EM-type)!")
pred = predList[0]
datasetPredictions.append(pred)
txnameList += pred.txnames
elementList += pred.elements
if not totalXsec:
totalXsec = pred.xsection
else:
totalXsec += pred.xsection
if not pred.mass in massList:
massList.append(pred.mass)
for pidEntry in pred.PIDs:
if not pidEntry in PIDList:
PIDList.append(pidEntry)
IDList += pred.IDs
txnameList = list(set(txnameList))
IDList = list(set(IDList))
if len(massList) > 1:
mass = None
else:
mass = massList[0]
#Create a combinedDataSet object:
combinedDataset = CombinedDataSet(expResult)
combinedDataset._marginalize = marginalize
#Create a theory preidction object for the combined datasets:
theoryPrediction = TheoryPrediction()
theoryPrediction.dataset = combinedDataset
theoryPrediction.txnames = txnameList
theoryPrediction.xsection = totalXsec
theoryPrediction.datasetPredictions = datasetPredictions
theoryPrediction.conditions = None
theoryPrediction.elements = elementList
theoryPrediction.mass = mass
theoryPrediction.PIDs = PIDList
theoryPrediction.IDs = IDList
return theoryPrediction
def _getBestResult(dataSetResults):
"""
Returns the best result according to the expected upper limit
:param datasetPredictions: list of TheoryPredictionList objects
:return: best result (TheoryPrediction object)
"""
#In the case of UL analyses or efficiency-maps with a single signal region
#return the single result:
if len(dataSetResults) == 1:
return dataSetResults[0]
#For efficiency-map analyses with multipler signal regions,
#select the best one according to the expected upper limit:
bestExpectedR = 0.
bestXsec = 0.*fb
for predList in dataSetResults:
if len(predList) != 1:
logger.error("Multiple clusters should only exist for upper limit results!")
raise SModelSError()
dataset = predList[0].dataset
if dataset.getType() != 'efficiencyMap':
logger.error("Multiple data sets should only exist for efficiency map results!")
raise SModelSError()
pred = predList[0]
xsec = pred.xsection
expectedR = (xsec.value/dataset.getSRUpperLimit(0.05,True,False) ).asNumber()
if expectedR > bestExpectedR or (expectedR == bestExpectedR and xsec.value > bestXsec):
bestExpectedR = expectedR
bestPred = pred
bestXsec = xsec.value
return bestPred
def _getDataSetPredictions(dataset,smsTopList,maxMassDist):
"""
Compute theory predictions for a given data set.
For upper-limit results returns the list of theory predictions for the
experimental result. For efficiency-map results returns the list of theory
predictions for the signal region. Uses the list of elements in
smsTopList.
For each Txname appearing in dataset, it collects the elements and efficiencies,
combine the masses (if needed) and compute the conditions (if existing).
:parameter dataset: Data Set to be considered (DataSet object)
:parameter smsTopList: list of topologies containing elements
(TopologyList object)
:parameter maxMassDist: maximum mass distance for clustering elements (float)
:returns: a TheoryPredictionList object containing a list of TheoryPrediction
objects
"""
predictionList = TheoryPredictionList()
# Select elements belonging to expResult and apply efficiencies
elements = _getElementsFrom(smsTopList, dataset)
#Check dataset sqrts format:
if (dataset.globalInfo.sqrts/TeV).normalize()._unit:
ID = dataset.globalInfo.id
logger.error( "Sqrt(s) defined with wrong units for %s" % (ID) )
return False
#Remove unwanted cross sections
newelements = []
for el in elements:
el.weight = el.weight.getXsecsFor(dataset.globalInfo.sqrts)
if not el.weight: continue
newelements.append(el)
elements = newelements
if len(elements) == 0:
return None
# Combine elements according to their respective constraints and masses
# (For efficiencyMap analysis group all elements)
clusters = _combineElements(elements, dataset, maxDist=maxMassDist)
# Collect results and evaluate conditions
for cluster in clusters:
theoryPrediction = TheoryPrediction()
theoryPrediction.dataset = dataset
theoryPrediction.txnames = cluster.txnames
theoryPrediction.xsection = _evalConstraint(cluster)
theoryPrediction.conditions = _evalConditions(cluster)
theoryPrediction.elements = cluster.elements
theoryPrediction.mass = cluster.getAvgMass()
theoryPrediction.PIDs = cluster.getPIDs()
theoryPrediction.IDs = cluster.getIDs()
predictionList._theoryPredictions.append(theoryPrediction)
if len(predictionList) == 0:
return None
else:
return predictionList
def _getElementsFrom(smsTopList, dataset):
"""
Get elements that belong to any of the TxNames in dataset
(appear in any of constraints in the result).
Loop over all elements in smsTopList and returns a copy of the elements belonging
to any of the constraints (i.e. have efficiency != 0). The copied elements
have their weights multiplied by their respective efficiencies.
:parameter dataset: Data Set to be considered (DataSet object)
:parameter smsTopList: list of topologies containing elements
(TopologyList object)
:returns: list of elements (Element objects)
"""
elements = []
for txname in dataset.txnameList:
for top in smsTopList:
itop = txname._topologyList.index(top) #Check if the topology appear in txname
if itop is None: continue
for el in top.getElements():
newEl = txname.hasElementAs(el) #Check if element appears in txname
if not newEl: continue
el.covered = True
eff = txname.getEfficiencyFor(newEl.getMasses())
if not eff:
continue
el.tested = True
newEl.eff = eff
newEl.weight *= eff
newEl.txname = txname
elements.append(newEl) #Save element with correct branch ordering
#Remove duplicated elements:
allmothers = []
#First collect the list of all mothers:
for el in elements:
allmothers += [elMom[1].elID for elMom in el.motherElements]
elementsClean = []
for el in elements:
#Skip the element if it is a mother of another element in the list
if any((elMom is el.elID) for elMom in allmothers):
continue
elementsClean.append(el)
return elementsClean
def _combineElements(elements, dataset, maxDist):
"""
Combine elements according to the data set type.
If expResult == upper limit type, first group elements with different TxNames
and then into mass clusters.
If expResult == efficiency map type, group all elements into a single cluster.
:parameter elements: list of elements (Element objects)
:parameter expResult: Data Set to be considered (DataSet object)
:returns: list of element clusters (ElementCluster objects)
"""
clusters = []
if dataset.getType() == 'efficiencyMap':
cluster = clusterTools.groupAll(elements)
clusters.append(cluster)
elif dataset.getType() == 'upperLimit':
txnames = list(set([el.txname for el in elements]))
for txname in txnames:
txnameEls = []
for element in elements:
if not element.txname == txname:
continue
else: txnameEls.append(element)
txnameClusters = clusterTools.clusterElements(txnameEls, maxDist)
clusters += txnameClusters
else:
logger.warning("Unkown data type: %s. Data will be ignored."
% dataset.getType())
return clusters
def _evalConstraint(cluster):
"""
Evaluate the constraint inside an element cluster.
If the cluster refers to a specific TxName, sum all the elements' weights
according to the analysis constraint.
For efficiency map results, sum all the elements' weights.
:parameter cluster: cluster of elements (ElementCluster object)
:returns: cluster cross section
"""
if cluster.getDataType() == 'efficiencyMap':
return cluster.getTotalXSec()
elif cluster.getDataType() == 'upperLimit':
if len(cluster.txnames) != 1:
logger.error("An upper limit cluster should never contain more than one TxName")
raise SModelSError()
txname = cluster.txnames[0]
if not txname.constraint or txname.constraint == "not yet assigned":
return txname.constraint
exprvalue = _evalExpression(txname.constraint,cluster)
return exprvalue
else:
logger.error("Unknown data type %s" %(str(cluster.getDataType())))
raise SModelSError()
def _evalConditions(cluster):
"""
Evaluate the conditions (if any) inside an element cluster.
:parameter cluster: cluster of elements (ElementCluster object)
:returns: list of condition values (floats) if analysis type == upper limit. None, otherwise.
"""
conditionVals = {}
for txname in cluster.txnames:
if not txname.condition or txname.condition == "not yet assigned":
continue
#Make sure conditions is always a list
if isinstance(txname.condition,str):
conditions = [txname.condition]
elif isinstance(txname.condition,list):
conditions = txname.condition
else:
logger.error("Conditions should be a list or a string")
raise SModelSError()
# Loop over conditions
for cond in conditions:
exprvalue = _evalExpression(cond,cluster)
if isinstance(exprvalue,crossSection.XSection):
conditionVals[cond] = exprvalue.value
else:
conditionVals[cond] = exprvalue
if not conditionVals:
return None
else:
return conditionVals
def _evalExpression(stringExpr,cluster):
"""
Auxiliary method to evaluate a string expression using the weights of the elements in the cluster.
Replaces the elements in stringExpr (in bracket notation) by their weights and evaluate the
expression.
e.g. computes the total weight of string expressions such as "[[[e^+]],[[e^-]]]+[[[mu^+]],[[mu^-]]]"
or ratios of weights of string expressions such as "[[[e^+]],[[e^-]]]/[[[mu^+]],[[mu^-]]]"
and so on...
:parameter stringExpr: string containing the expression to be evaluated
:parameter cluster: cluster of elements (ElementCluster object)
:returns: xsection for the expression. Can be a XSection object, a float or not numerical (None,string,...)
"""
#Get cross section info from cluster (to generate zero cross section values):
infoList = cluster.elements[0].weight.getInfo()
#Get final state info
finalStates = cluster.elements[0].getFinalStates()
#Get weights for elements appearing in stringExpr
weightsDict = {}
evalExpr = stringExpr.replace("'","").replace(" ","")
for i,elStr in enumerate(elementsInStr(evalExpr)):
el = element.Element(elStr)
el.setFinalState(finalStates)
weightsDict['w%i'%i] = crossSection.XSectionList(infoList)
for el1 in cluster.elements:
if el1.particlesMatch(el):
weightsDict['w%i'%i].combineWith(el1.weight)
el.combineMotherElements(el1)
evalExpr = evalExpr.replace(elStr,'w%i'%i)
weightsDict.update({"Cgtr" : cGtr, "cGtr" : cGtr, "cSim" : cSim, "Csim" : cSim})
exprvalue = eval(evalExpr, weightsDict)
if type(exprvalue) == type(crossSection.XSectionList()):
if len(exprvalue) != 1:
logger.error("Evaluation of expression "+evalExpr+" returned multiple values.")
return exprvalue[0] #Return XSection object
return exprvalue