PYTHIA  8.303
Public Member Functions | Public Attributes | Friends | List of all members
DireHistory Class Reference

#include <DireHistory.h>

Public Member Functions

 DireHistory (int depthIn, double scalein, Event statein, DireClustering c, MergingHooksPtr mergingHooksPtrIn, BeamParticle beamAIn, BeamParticle beamBIn, ParticleData *particleDataPtrIn, Info *infoPtrIn, PartonLevel *showersIn, shared_ptr< DireTimes > fsrIn, shared_ptr< DireSpace > isrIn, DireWeightContainer *psweightsIn, CoupSM *coupSMPtrIn, bool isOrdered, bool isAllowed, double clusterProbIn, double clusterCouplIn, double prodOfProbsIn, double prodOfProbsFullIn, DireHistory *mothin)
 
 ~DireHistory ()
 The destructor deletes each child.
 
bool projectOntoDesiredHistories ()
 Function to project paths onto desired paths. More...
 
double weightTREE (PartonLevel *trial, AlphaStrong *asFSR, AlphaStrong *asISR, AlphaEM *aemFSR, AlphaEM *aemISR, double RN)
 
double weightMOPS (PartonLevel *trial, AlphaStrong *as, AlphaEM *aem, double RN)
 
vector< double > weightMEM (PartonLevel *trial, AlphaStrong *as, AlphaEM *aem, double RN)
 
double weightMEC ()
 
double weightLOOP (PartonLevel *trial, double RN)
 
double weightFIRST (PartonLevel *trial, AlphaStrong *asFSR, AlphaStrong *asISR, AlphaEM *aemFSR, AlphaEM *aemISR, double RN, Rndm *rndmPtr)
 Return O()-term of CKKWL-weight for NL3 merging. More...
 
double weight_UMEPS_TREE (PartonLevel *trial, AlphaStrong *asFSR, AlphaStrong *asISR, AlphaEM *aemFSR, AlphaEM *aemISR, double RN)
 For UMEPS: More...
 
double weight_UMEPS_SUBT (PartonLevel *trial, AlphaStrong *asFSR, AlphaStrong *asISR, AlphaEM *aemFSR, AlphaEM *aemISR, double RN)
 Function to return weight of virtual correction events for NLO merging. More...
 
double weight_UNLOPS_TREE (PartonLevel *trial, AlphaStrong *asFSR, AlphaStrong *asISR, AlphaEM *aemFSR, AlphaEM *aemISR, double RN, int depthIn=-1)
 For unitary NL3: More...
 
double weight_UNLOPS_SUBT (PartonLevel *trial, AlphaStrong *asFSR, AlphaStrong *asISR, AlphaEM *aemFSR, AlphaEM *aemISR, double RN, int depthIn=-1)
 
double weight_UNLOPS_LOOP (PartonLevel *trial, AlphaStrong *asFSR, AlphaStrong *asISR, AlphaEM *aemFSR, AlphaEM *aemISR, double RN, int depthIn=-1)
 
double weight_UNLOPS_SUBTNLO (PartonLevel *trial, AlphaStrong *asFSR, AlphaStrong *asISR, AlphaEM *aemFSR, AlphaEM *aemISR, double RN, int depthIn=-1)
 
double weight_UNLOPS_CORRECTION (int order, PartonLevel *trial, AlphaStrong *asFSR, AlphaStrong *asISR, AlphaEM *aemFSR, AlphaEM *aemISR, double RN, Rndm *rndmPtr)
 Function to calculate O()-term of CKKWL-weight for NLO merging. More...
 
bool foundAllowedHistories ()
 Function to check if any allowed histories were found.
 
bool foundOrderedHistories ()
 Function to check if any ordered histories were found.
 
bool foundCompleteHistories ()
 Function to check if any ordered histories were found.
 
void getStartingConditions (const double RN, Event &outState)
 Function to set the state with complete scales for evolution. More...
 
bool getClusteredEvent (const double RN, int nSteps, Event &outState)
 Function to get the state with complete scales for evolution. More...
 
bool getFirstClusteredEventAboveTMS (const double RN, int nDesired, Event &process, int &nPerformed, bool updateProcess=true)
 Function to get the first reclustered state above the merging scale. More...
 
int nClusterings ()
 
Event lowestMultProc (const double RN)
 Function to get the lowest multiplicity reclustered event. More...
 
double getPDFratio (int side, bool forSudakov, bool useHardPDF, int flavNum, double xNum, double muNum, int flavDen, double xDen, double muDen)
 Calculate and return pdf ratio. More...
 
void printHistory (const double RN)
 Function to print the history that would be chosen from the number RN. More...
 
void printStates ()
 Function to print the states in a history, starting from the hard process. More...
 

Public Attributes

Event state
 
int generation
 Index for generation.
 
map< double, DireHistory * > goodBranches
 
map< double, DireHistory * > badBranches
 

Friends

class Pythia
 Make Pythia class friend.
 
class DireMerging
 Make Merging class friend.
 
class DireTimes
 
class DireSpace
 

Detailed Description

Declaration of MyHistory class

A MyHistory object represents an event in a given step in the CKKW-L clustering procedure. It defines a tree-like recursive structure, where the root node represents the state with n jets as given by the matrix element generator, and is characterized by the member variable mother being null. The leaves on the tree corresponds to a fully clustered paths where the original n-jets has been clustered down to the Born-level state. Also states which cannot be clustered down to the Born-level are possible - these will be called incomplete. The leaves are characterized by the vector of children being empty.

Constructor & Destructor Documentation

DireHistory ( int  depthIn,
double  scalein,
Event  statein,
DireClustering  c,
MergingHooksPtr  mergingHooksPtrIn,
BeamParticle  beamAIn,
BeamParticle  beamBIn,
ParticleData particleDataPtrIn,
Info infoPtrIn,
PartonLevel showersIn,
shared_ptr< DireTimes fsrIn,
shared_ptr< DireSpace isrIn,
DireWeightContainer psweightsIn,
CoupSM coupSMPtrIn,
bool  isOrdered = true,
bool  isAllowed = true,
double  clusterProbIn = 1.0,
double  clusterCouplIn = 1.0,
double  prodOfProbsIn = 1.0,
double  prodOfProbsFullIn = 1.0,
DireHistory mothin = 0 
)

The only constructor. Default arguments are used when creating the initial history node. The depth is the maximum number of clusterings requested. scalein is the scale at which the statein was clustered (should be set to the merging scale for the initial history node. beamAIn and beamBIn are needed to calcutate PDF ratios, particleDataIn to have access to the correct masses of particles. If isOrdered is true, the previous clusterings has been ordered. is the PDF ratio for this clustering (=1 for FSR clusterings). probin is the accumulated probabilities for the previous clusterings, and \ mothin is the previous history node (null for the initial node).

Declaration of DireHistory class The only constructor. Default arguments are used when creating the initial history node. The depth is the maximum number of clusterings requested. scalein is the scale at which the statein was clustered (should be set to the merging scale for the initial history node. beamAIn and beamBIn are needed to calcutate PDF ratios, particleDataIn to have access to the correct masses of particles. If isOrdered is true, the previous clusterings has been ordered. is the PDF ratio for this clustering (=1 for FSR clusterings). probin is the accumulated probabilities for the previous clusterings, and \ mothin is the previous history node (null for the initial node).

Initialize.

Remember how many steps in total were supposed to be taken.

Initialise beam particles

Update probability with PDF ratio

Remember reclustered radiator in lower multiplicity state

Check if more steps should be taken.

Update generation index.

If this is not the fully clustered state, try to find possible QCD clusterings.

Check if fixed-order ME calculation for this state exists.

Calculate ME

Check if more steps should be taken.

If no clusterings were found, the recursion is done and we register this node.

Multiply with hard process matrix element.

Divide out all couplings in hard process matrix element.

If the coupling exactly vanishes, force weight to zero.

Additional ordering requirement between shower starting scale and scale of first emission.

Now we sort the possible clusterings so that we try the smallest scale first.

Check if reclustering follows ordered sequence.

If this path is not ordered in pT and we already have found an ordered path, then we don't need to continue along this path, unless we have not yet found an allowed path.

Check if reclustered state should be disallowed.

Skip clustering with vanishing probability.

Skip if this branch is already strongly suppressed.

Perform the clustering and recurse and construct the next history node.

Register as valid history if no children allowed.

Multiply with hard process matrix element.

Divide out all couplings in hard process matrix element.

If the coupling exactly vanishes, force weight to zero.

Member Function Documentation

bool getClusteredEvent ( const double  RN,
int  nSteps,
Event outState 
)

Function to get the state with complete scales for evolution.

Function to set the state with complete scales for evolution.

Select history

Set scales in the states to the scales pythia would have set (Only needed if not done before in calculation of weights or setting of starting conditions)

If the history does not allow for nSteps clusterings (e.g. because the history is incomplete), return false

Return event with nSteps-1 additional partons (i.e. recluster the last splitting) and copy the output state

Done.

bool getFirstClusteredEventAboveTMS ( const double  RN,
int  nDesired,
Event process,
int &  nPerformed,
bool  updateProcess = true 
)

Function to get the first reclustered state above the merging scale.

Do reclustering (looping) steps. Remember process scale.

Get number of clustering steps.

Set scales in the states to the scales pythia would have set.

Recluster until reclustered event is above the merging scale.

Initialise temporary output of reclustering.

Recluster once more.

If reclustered event does not exist, exit.

Continue loop if reclustered event has unresolved partons.

Update the hard process.

Failed to produce output state.

Update to the actual number of steps.

Save MPI starting scale

Done

double getPDFratio ( int  side,
bool  forSudakov,
bool  useHardPDF,
int  flavNum,
double  xNum,
double  muNum,
int  flavDen,
double  xDen,
double  muDen 
)

Calculate and return pdf ratio.

Do nothing for e+e- beams

Now calculate PDF ratio if necessary

Get mother and daughter pdfs

Use hard process PDFs (i.e. PDFs NOT used in ISR, FSR or MPI).

Use rescaled PDFs in the presence of multiparton interactions

Cut out charm threshold.

Return ratio of pdfs

Done

void getStartingConditions ( const double  RN,
Event outState 
)

Function to set the state with complete scales for evolution.

Select the history

Set scales in the states to the scales pythia would have set

Get number of clustering steps.

Update the lowest order process.

Save information on last splitting, to allow the next emission in the shower to have smaller rapidity with respect to the last ME splitting.

Copy the output state.

Save MPI starting scale.

Event lowestMultProc ( const double  RN)
inline

Function to get the lowest multiplicity reclustered event.

Return lowest multiplicity state

int nClusterings ( )

Function to return the depth of the history (i.e. the number of reclustered splittings)

Function to return the depth of the history (i.e. the number of reclustered splittings) NO INPUT OUTPUT int : Depth of history

void printHistory ( const double  RN)

Function to print the history that would be chosen from the number RN.

Function to print the history that would be chosen from the random number RN. Mainly for debugging.

Done

void printStates ( )

Function to print the states in a history, starting from the hard process.

Function to print the states in a history, starting from the hard process. Mainly for debugging.

Print.

Recurse

Done

bool projectOntoDesiredHistories ( )

Function to project paths onto desired paths.

Function to project all possible paths onto only the desired paths.

In MOPS, discard states that yield clusterings below the shower cut-off.

Loop through good branches and set the set of "good" children in mother nodes.

Set good sisters.

Multiply couplings and ME corrections to probability.

Loop through the good branches and set their correct probabilities, if necessary.

paths.begin()->second->mother->setProbabilities(); Set probabilities from next-to-lowest multi –> highest multi. If lowest multi == highest multi, no need to set probabilities.

Set probabilities from next-to-lowest multi –> highest multi. If lowest multi == highest multi, no need to set probabilities.

Multiply couplings and ME corrections to probability.

Trim to desirable histories.

return foundGood;

double weight_UMEPS_SUBT ( PartonLevel trial,
AlphaStrong asFSR,
AlphaStrong asISR,
AlphaEM aemFSR,
AlphaEM aemISR,
double  RN 
)

Function to return weight of virtual correction events for NLO merging.

Read alpha_S in ME calculation and maximal scale (eCM)

Select a path of clusterings

Set scales in the states to the scales pythia would have set

Get weight.

Do trial shower, calculation of alpha_S ratios, PDF ratios

MPI no-emission probability.

Set hard process renormalisation scale to default Pythia value.

For pure QCD dijet events, evaluate the coupling of the hard process at a more reasonable pT, rather than evaluation at a fixed arbitrary scale.

Reset to a running coupling. Here we choose FSR for simplicity.

For prompt photon events, evaluate the coupling of the hard process at a more reasonable pT, rather than evaluation at a fixed arbitrary scale.

Reset to a running coupling. In prompt photon always ISR.

Done

double weight_UMEPS_TREE ( PartonLevel trial,
AlphaStrong asFSR,
AlphaStrong asISR,
AlphaEM aemFSR,
AlphaEM aemISR,
double  RN 
)

For UMEPS:

No difference to CKKW-L. Recycle CKKW-L function.

double weight_UNLOPS_CORRECTION ( int  order,
PartonLevel trial,
AlphaStrong asFSR,
AlphaStrong asISR,
AlphaEM aemFSR,
AlphaEM aemISR,
double  RN,
Rndm rndmPtr 
)

Function to calculate O()-term of CKKWL-weight for NLO merging.

Dummy statement to avoid compiler warnings.

Already done if no correction should be calculated

Read alpha_S in ME calculation and maximal scale (eCM)

double aemME = infoPtr->alphaEM();

Pick path of clusterings

Set scales in the states to the scales pythia would have set

Get the lowest order k-factor and add first two terms in expansion

If using Bbar, which includes a tree-level part, subtract an additional one, i.e. the O(^0) contribution as well

If only O(^0)-term is to be calculated, done already.

Start by adding the O(^1)-term of the k-factor.

Calculate sum of O(^1)-terms of the ckkw-l weight WITHOUT the O(^1)-term of the last no-emission probability. Get first term in expansion of alpha_s ratios.

Add logarithm from expansion to weight.

Generate true average, not only one-point.

Get average number of emissions.

Add average number of emissions off reconstructed states to weight.

Get first term in expansion of PDF ratios.

Add integral of DGLAP shifted PDF ratios from expansion to wt.

If O(^1)-term + O(^1)-term is to be calculated, done.

So far, no calculation of O(^2)-term

double weight_UNLOPS_LOOP ( PartonLevel trial,
AlphaStrong asFSR,
AlphaStrong asISR,
AlphaEM aemFSR,
AlphaEM aemISR,
double  RN,
int  depthIn = -1 
)

No difference to default NL3

double weight_UNLOPS_SUBT ( PartonLevel trial,
AlphaStrong asFSR,
AlphaStrong asISR,
AlphaEM aemFSR,
AlphaEM aemISR,
double  RN,
int  depthIn = -1 
)

Select a path of clusterings

Set scales in the states to the scales pythia would have set

So far, no reweighting

Read alpha_S in ME calculation and maximal scale (eCM)

Only allow two clusterings if all intermediate states above the merging scale.

Get weights: alpha_S ratios and PDF ratios

Do trial shower, calculation of alpha_S ratios, PDF ratios

MPI no-emission probability.

Set weight

Done

double weight_UNLOPS_SUBTNLO ( PartonLevel trial,
AlphaStrong asFSR,
AlphaStrong asISR,
AlphaEM aemFSR,
AlphaEM aemISR,
double  RN,
int  depthIn = -1 
)

Select a path of clusterings

Set scales in the states to the scales pythia would have set

So far, no reweighting

Only reweighting with MPI no-emission probability

Done

double weight_UNLOPS_TREE ( PartonLevel trial,
AlphaStrong asFSR,
AlphaStrong asISR,
AlphaEM aemFSR,
AlphaEM aemISR,
double  RN,
int  depthIn = -1 
)

For unitary NL3:

Read alpha_S in ME calculation and maximal scale (eCM)

Select a path of clusterings

Set scales in the states to the scales pythia would have set

Get weight.

Do trial shower, calculation of alpha_S ratios, PDF ratios

MPI no-emission probability.

Set hard process renormalisation scale to default Pythia value.

For pure QCD dijet events, evaluate the coupling of the hard process at a more reasonable pT, rather than evaluation at a fixed arbitrary scale.

Reset to a running coupling. Here we choose FSR for simplicity.

For prompt photon events, evaluate the coupling of the hard process at a more reasonable pT, rather than evaluation at a fixed arbitrary scale.

Reset to a running coupling. In prompt photon always ISR.

Done

double weightFIRST ( PartonLevel trial,
AlphaStrong asFSR,
AlphaStrong asISR,
AlphaEM aemFSR,
AlphaEM aemISR,
double  RN,
Rndm rndmPtr 
)

Return O()-term of CKKWL-weight for NL3 merging.

Function to calculate O()-term of CKKWL-weight for NLO merging.

Dummy statement to avoid compiler warnings.

Read alpha_S in ME calculation and maximal scale (eCM)

Pick path of clusterings

Set scales in the states to the scales pythia would have set

Get the lowest order k-factor and add first two terms in expansion

If using Bbar, which includes a tree-level part, subtract an additional one, i.e. the O(^0) contribution as well

Calculate sum of O(alpha) terms

Get starting scale for trial showers.

Count emissions: New variant Generate true average, not only one-point

Get number of emissions

Done

double weightLOOP ( PartonLevel trial,
double  RN 
)

For default NL3: Return weight of virtual correction and subtractive for NL3 merging

Function to return weight of virtual correction and subtractive events for NL3 merging

Select a path of clusterings

Set scales in the states to the scales pythia would have set

So far, no reweighting

Only reweighting with MPI no-emission probability

Done

vector< double > weightMEM ( PartonLevel trial,
AlphaStrong as,
AlphaEM aem,
double  RN 
)

Read alpha_S in ME calculation and maximal scale (eCM)

Select a path of clusterings

Set scales in the states to the scales pythia would have set

Calculate no-emission probability with trial shower.

Calculate PDF ratios.

Calculate coupling ratios.

double weightMOPS ( PartonLevel trial,
AlphaStrong as,
AlphaEM aem,
double  RN 
)

In the initial history node, select one of the paths according to the probabilities. This function should be called for the initial history node. IN trialShower* : Previously initialised trialShower object, to perform trial showering and as repository of pointers to initialise alphaS PartonSystems* : PartonSystems object needed to initialise shower objects OUT double : (Sukadov) , (alpha_S ratios) , (PDF ratios)

Read alpha_S in ME calculation and maximal scale (eCM)

Select a path of clusterings

Set scales in the states to the scales pythia would have set

Keep only unordered paths, since all ordered paths have been corrected with matrix element corrections.

Calculate no-emission probability with trial shower.

Calculate PDF ratios.

Calculate coupling ratios.

MPI no-emission probability

Done

double weightTREE ( PartonLevel trial,
AlphaStrong asFSR,
AlphaStrong asISR,
AlphaEM aemFSR,
AlphaEM aemISR,
double  RN 
)

For CKKW-L, NL3 and UMEPS: In the initial history node, select one of the paths according to the probabilities. This function should be called for the initial history node. IN trialShower* : Previously initialised trialShower object, to perform trial showering and as repository of pointers to initialise alphaS PartonSystems* : PartonSystems object needed to initialise shower objects OUT double : (Sukadov) , (alpha_S ratios) , (PDF ratios)

In the initial history node, select one of the paths according to the probabilities. This function should be called for the initial history node. IN trialShower* : Previously initialised trialShower object, to perform trial showering and as repository of pointers to initialise alphaS PartonSystems* : PartonSystems object needed to initialise shower objects OUT double : (Sukadov) , (alpha_S ratios) , (PDF ratios)

Read alpha_S in ME calculation and maximal scale (eCM)

Select a path of clusterings

Set scales in the states to the scales pythia would have set

Get weight.

Do trial shower, calculation of alpha_S ratios, PDF ratios

MPI no-emission probability

double mpiwt = selected->weightEmissions( trial, -1, 0, njetsMaxMPI, maxScale );

Set hard process renormalisation scale to default Pythia value.

For pure QCD dijet events, evaluate the coupling of the hard process at a more reasonable pT, rather than evaluation at a fixed arbitrary scale.

Reset to a running coupling. Here we choose FSR for simplicity.

Reset to a running coupling. Here we choose FSR for simplicity.

For W clustering, correct the .

Reset to a running coupling. Here we choose FSR for simplicity.

For prompt photon events, evaluate the coupling of the hard process at a more reasonable pT, rather than evaluation at a fixed arbitrary scale.

Reset to a running coupling. In prompt photon always ISR.

For DIS, set the hard process scale to Q2.

Done

Member Data Documentation

map<double,DireHistory *> goodBranches

The different allowed paths after projection, indexed with the (incremental) corresponding probability. This map is empty unless this is the initial step (mother == 0).

Event state

The state of the event correponding to this step in the reconstruction.


The documentation for this class was generated from the following files: