These functions are part of QEDbase.jl's particle interface.
ParticleStateful is the implementation of QEDbase's QEDbase.AbstractParticleStateful interface. It represents a particle with a direction (as used in the context of scattering processes, QEDbase.Incoming, QEDbase.Outgoing, or QEDbase.UnknownDirection), a particle species (Electron, Positron, Photon, ...), and a 4-momentum vector.
ps = ParticleStateful(Incoming(), Electron(), rand(SFourMomentum))ParticleStateful: incoming electron
- momentum: [0.47703174926171565, 0.19460163162039423, 0.9151576843508751, 0.003738391918556183]
+ momentum: [0.5124300336542875, 0.7400217661008659, 0.2351508149844419, 0.48646193706840823]
The relevant accessor functions for the interface are implemented:
particle_direction(ps)incomingparticle_species(ps)electronmomentum(ps)4-element SFourMomentum with indices SOneTo(4):
- 0.47703174926171565
- 0.19460163162039423
- 0.9151576843508751
- 0.003738391918556183A PhaseSpacePoint is the combination of incoming and outgoing ParticleStatefuls. It also contains information about the scattering process, model, and phase space that it is created for.
psp = PhaseSpacePoint(
+ 0.5124300336542875
+ 0.7400217661008659
+ 0.2351508149844419
+ 0.48646193706840823A PhaseSpacePoint is the combination of incoming and outgoing ParticleStatefuls. It also contains information about the scattering process, model, and phase space that it is created for.
psp = PhaseSpacePoint(
Compton(), # scattering process
PerturbativeQED(), # physics model
PhasespaceDefinition( # phase space definition
@@ -33,11 +33,11 @@
model: perturbative QED
phasespace definition: spherical coordinates in electron rest frame
incoming particles:
- -> incoming electron: [0.8899378446541081, 0.004042153604023158, 0.397980274465378, 0.9610450439858487]
- -> incoming photon: [0.6637061172891727, 0.01462410761335109, 0.10090752719904517, 0.20802200980715757]
+ -> incoming electron: [0.19364840415165618, 0.24713276973982545, 0.9822832773674299, 0.5638418824042442]
+ -> incoming photon: [0.47325619402520813, 0.7797770507299893, 0.4210694557748539, 0.9689293691848972]
outgoing particles:
- -> outgoing electron: [0.0696772754527597, 0.7641012166594603, 0.5686140488697503, 0.2563191730085329]
- -> outgoing photon: [0.5659818523982115, 0.24166991346480982, 0.6278900175446048, 0.4592090607795496]
+ -> outgoing electron: [0.918573794862195, 0.7892778303161222, 0.6605951799243333, 0.9791779741049096]
+ -> outgoing photon: [0.9450018883849867, 0.28221344002700044, 0.13643028665632917, 0.20100520986150583]
This version of the constructor automatically creates ParticleStateful obejcts from the momenta, matching the particles of the process. In the case of Compton, this is means an incoming electron and photon, and outgoing electron and photon.
Automatic checks make sure that the number of 4-momenta given matches the necessary number of 4-momenta for the process (this adds 0 overhead at runtime because it is inferred from type information alone).
PhaseSpacePoint(
Compton(),
PerturbativeQED(),
@@ -65,11 +65,11 @@
model: perturbative QED
phasespace definition: spherical coordinates in electron rest frame
incoming particles:
- -> incoming electron: [0.6429611412493572, 0.7837700675283127, 0.2619477240927065, 0.4841195234855733]
- -> incoming photon: [0.643802296078773, 0.02287745011387987, 0.7548684149523159, 0.09984651940294875]
+ -> incoming electron: [0.5674869193331556, 0.2024874623404207, 0.921436568429739, 0.5504496842704256]
+ -> incoming photon: [0.38137571150850036, 0.9043507373617743, 0.47898427197873894, 0.13326362720199847]
outgoing particles:
- -> outgoing electron: [0.3271902398635579, 0.061495145276687424, 0.7367392725723305, 0.34471658447863296]
- -> outgoing photon: [0.5611987883342415, 0.7228200282378292, 0.8151759715010839, 0.43435743141580496]
+ -> outgoing electron: [0.45647920499745265, 0.013441466367441368, 0.3817364383240517, 0.9195676098173129]
+ -> outgoing photon: [0.7878063900891932, 0.265257473337417, 0.14226328024077395, 0.3535591839985033]
Similar to the constructor from momenta, this checks that the given ParticleStatefuls fit to the given process and throws otherwise. Again, since this can be infered from type information alone, it adds no overhead.
PhaseSpacePoint(
Compton(),
PerturbativeQED(),
@@ -84,19 +84,19 @@
),
)┌ Error: InvalidInputError("expected incoming electron but got incoming positron")
└ @ Main particles.md:162While these constructors check that the given types make sense and work together, they do not check whether the given momenta make a physical phase space point or that the incoming or outgoing particles have on-shell 4-momenta.
The phase space point provides some convenient accessors to the stateful particles within:
psp[Incoming(), 1] # the first incoming particleParticleStateful: incoming electron
- momentum: [0.6429611412493572, 0.7837700675283127, 0.2619477240927065, 0.4841195234855733]
+ momentum: [0.5674869193331556, 0.2024874623404207, 0.921436568429739, 0.5504496842704256]
psp[Outgoing(), 2] # the second outgoing particleParticleStateful: outgoing photon
- momentum: [0.5611987883342415, 0.7228200282378292, 0.8151759715010839, 0.43435743141580496]
-particles(psp, Incoming()) # all incoming particles as a tuple(incoming electron: [0.6429611412493572, 0.7837700675283127, 0.2619477240927065, 0.4841195234855733], incoming photon: [0.643802296078773, 0.02287745011387987, 0.7548684149523159, 0.09984651940294875])Momentum accessors:
momentum(psp, Incoming(), Electron(), 1) # the momentum of the first incoming electron4-element SFourMomentum with indices SOneTo(4):
- 0.6429611412493572
- 0.7837700675283127
- 0.2619477240927065
- 0.4841195234855733When only one particle of the species exists in the particle set, the 1 can be ommitted for convenience.
@assert ans == momentum(psp, Incoming(), Electron())This method throws when multiple (or zero) particles of the given direction and species exist in the phase space point.
When the index of the required momentum is known at compile time, a Val(N) can be used instead of N. This performs bounds checks at compile time and removes loops from the runtime execution
using BenchmarkTools
+ momentum: [0.7878063900891932, 0.265257473337417, 0.14226328024077395, 0.3535591839985033]
+particles(psp, Incoming()) # all incoming particles as a tuple(incoming electron: [0.5674869193331556, 0.2024874623404207, 0.921436568429739, 0.5504496842704256], incoming photon: [0.38137571150850036, 0.9043507373617743, 0.47898427197873894, 0.13326362720199847])Momentum accessors:
momentum(psp, Incoming(), Electron(), 1) # the momentum of the first incoming electron4-element SFourMomentum with indices SOneTo(4):
+ 0.5674869193331556
+ 0.2024874623404207
+ 0.921436568429739
+ 0.5504496842704256When only one particle of the species exists in the particle set, the 1 can be ommitted for convenience.
@assert ans == momentum(psp, Incoming(), Electron())This method throws when multiple (or zero) particles of the given direction and species exist in the phase space point.
When the index of the required momentum is known at compile time, a Val(N) can be used instead of N. This performs bounds checks at compile time and removes loops from the runtime execution
using BenchmarkTools
judge(
median(@benchmark momentum($psp, Incoming(), Photon(), Val(1))),
median(@benchmark momentum($psp, Incoming(), Photon(), 1)),
)BenchmarkTools.TrialJudgement:
- time: -89.72% => improvement (5.00% tolerance)
+ time: -87.64% => improvement (5.00% tolerance)
memory: -100.00% => improvement (1.00% tolerance)
This is only faster when N is actually known at compile time, for example when it is a literal integer or a function's type parameter. For dynamic values of N, prefer the Int variant or in case of loops, directly loop over the tuple of momenta.
Some more overloads for the momentum function exist, for a complete list please refer to its documentation: QEDbase.momentum, QEDbase.momenta.
Finally, process, model, and phase_space_definition can be used to request the object in question:
process(psp)one-photon Compton scattering
incoming: electron (all spins), photon (all polarizations)
@@ -116,10 +116,10 @@
)
in_sum(psp)4-element SFourMomentum with indices SOneTo(4):
- 0.9685983876831324
- 0.9507606576104933
- 1.5851315567486823
- 1.0679407986175447Every full PhaseSpacePoint is both an InPhaseSpacePoint and an OutPhaseSpacePoint, too. For example, the in_sum function defined above still works with a full PhaseSpacePoint:
psp = PhaseSpacePoint(
+ 0.6412279295996643
+ 1.2870828853705887
+ 1.3228101063204138
+ 1.1965772413832587Every full PhaseSpacePoint is both an InPhaseSpacePoint and an OutPhaseSpacePoint, too. For example, the in_sum function defined above still works with a full PhaseSpacePoint:
psp = PhaseSpacePoint(
Compton(),
PerturbativeQED(),
PhasespaceDefinition(SphericalCoordinateSystem(), ElectronRestFrame()),
@@ -128,10 +128,10 @@
)
in_sum(psp)4-element SFourMomentum with indices SOneTo(4):
- 0.6866388172481368
- 1.265007772859199
- 1.5841141730166584
- 1.072191214970697But an InPhaseSpacePoint is not an OutPhaseSpacePoint and vice versa. We cannot call in_sum on an OutPhaseSpacePoint:
psp = OutPhaseSpacePoint(
+ 1.6850747517134694
+ 0.8931521290356832
+ 1.311106328053985
+ 0.7693022922954273But an InPhaseSpacePoint is not an OutPhaseSpacePoint and vice versa. We cannot call in_sum on an OutPhaseSpacePoint:
psp = OutPhaseSpacePoint(
Compton(),
PerturbativeQED(),
PhasespaceDefinition(SphericalCoordinateSystem(), ElectronRestFrame()),
@@ -139,4 +139,4 @@
)
in_sum(psp)┌ Error: MethodError(Main.in_sum, (PhaseSpacePoint of one-photon Compton scattering,), 0x0000000000007b40)
-└ @ Main particles.md:280This page was generated using Literate.jl.
Settings
This document was generated with Documenter.jl version 1.7.0 on Tuesday 5 November 2024. Using Julia version 1.10.6.