EIC code displayed by LXR master/​acts/​Fatras/​include/​ActsFatras/​Kernel/​InteractionList.hpp	Source navigation Diff markup Identifier search General search
	Version: master test Revert   Change

File indexing completed on 2025-10-18 08:22:02

 // This file is part of the ACTS project.
 //
 // Copyright (C) 2016 CERN for the benefit of the ACTS project
 //
 // This Source Code Form is subject to the terms of the Mozilla Public
 // License, v. 2.0. If a copy of the MPL was not distributed with this
 // file, You can obtain one at https://mozilla.org/MPL/2.0/.
 
 #pragma once
 
 #include "Acts/Material/MaterialSlab.hpp"
 #include "ActsFatras/EventData/Particle.hpp"
 
 #include <bitset>
 #include <tuple>
 #include <type_traits>
 #include <utility>
 
 namespace ActsFatras {
 namespace detail {
 
 /// Retrieve index of the first matching type in the tuple.
 ///
 /// Taken from https://stackoverflow.com/a/18063608.
 template <class T, class Tuple>
 struct TupleIndexOf;
 template <class T, class... Types>
 struct TupleIndexOf<T, std::tuple<T, Types...>> {
   static constexpr std::size_t value = 0u;
 };
 template <class T, class U, class... Types>
 struct TupleIndexOf<T, std::tuple<U, Types...>> {
   static constexpr std::size_t value =
       1u + TupleIndexOf<T, std::tuple<Types...>>::value;
 };
 
 // Construct an index sequence for a subset of the tuple elements.
 //
 // Whether an element is part of the subset is defined by the predicate
 // template type. It must take the element type as its only template parameter
 // and must provide a static `value` member value. If the value evaluates to
 // `true`, then the corresponding index will be part of the index sequence.
 //
 // Example: The tuple contains four elements, where all but the third one (i=2)
 // should be selected. This leads to the following recursive expansion
 // where the index sequence of the subset is filled from the front.
 //
 //        TupleFilterImpl<..., kCounter=4>          // select index=3
 //     -> TupleFilterImpl<..., kCounter=3, 3>       // skip   index=2
 //     -> TupleFilterImpl<..., kCounter=2, 3>       // select index=1
 //     -> TupleFilterImpl<..., kCounter=1, 1, 3>    // select index=0
 //     -> TupleFilterImpl<..., kCounter=0, 0, 1, 3> // terminate
 //
 template <template <typename> typename predicate_t, typename tuple_t,
           std::size_t kCounter, std::size_t... kIndices>
 struct TupleFilterImpl {
   static constexpr auto kIndex = kCounter - 1u;
   static constexpr bool kElementSelection =
       predicate_t<std::tuple_element_t<kIndex, tuple_t>>::value;
   // recursive type if the element would be selected
   using SelectElement = typename TupleFilterImpl<predicate_t, tuple_t, kIndex,
                                                  kIndex, kIndices...>::Type;
   // recursive type if the element would be skipped
   using SkipElement =
       typename TupleFilterImpl<predicate_t, tuple_t, kIndex, kIndices...>::Type;
   // select recursive type based on the selector decision
   using Type =
       std::conditional_t<kElementSelection, SelectElement, SkipElement>;
 };
 template <template <typename> typename predicate_t, typename tuple_t,
           std::size_t... kIndices>
 struct TupleFilterImpl<predicate_t, tuple_t, 0u, kIndices...> {
   using Type = std::index_sequence<kIndices...>;
 };
 template <template <typename> typename predicate_t, typename tuple_t>
 using TupleFilter = typename TupleFilterImpl<predicate_t, tuple_t,
                                              std::tuple_size_v<tuple_t>>::Type;
 
 /// Check if the given type is a point-like process.
 ///
 /// Only checks for the existence of the templated `generatePathLimits` method
 template <typename process_t>
 concept PointLikeProcessConcept = requires(
     const process_t& p, std::uniform_int_distribution<unsigned int>& rng,
     const Particle& prt) {
   { p.generatePathLimits(rng, prt) } -> std::same_as<std::pair<double, double>>;
 };
 
 template <typename process_t>
 concept ContinuousProcessConcept = !PointLikeProcessConcept<process_t>;
 
 template <typename process_t>
 struct PointLikeProcessTrait {
   static constexpr bool value = PointLikeProcessConcept<process_t>;
 };
 
 template <typename process_t>
 struct ContinuousProcessTrait {
   static constexpr bool value = ContinuousProcessConcept<process_t>;
 };
 
 template <typename processes_t>
 using ContinuousIndices = TupleFilter<ContinuousProcessTrait, processes_t>;
 template <typename processes_t>
 using PointLikeIndices = TupleFilter<PointLikeProcessTrait, processes_t>;
 
 }  // namespace detail
 
 /// Compile-time set of interaction processes for the simulation.
 ///
 /// Two different type of interaction processes are supported: continuous and
 /// point-like interactions.
 ///
 /// Continuous processes scale with the passed material. They tpyically
 /// describe effective results of a large number of small interactions such as
 /// multiple scattering or ionisation. Continuous process types **must** provide
 /// a call operator with the following signature:
 ///
 ///     template <typename generator_t>
 ///     bool
 ///     operator()(
 ///         generator_t& rng,
 ///         const Acts::MaterialSlab& slab,
 ///         Particle& particle,
 ///         std::vector<Particle>& generatedParticles) const
 ///
 /// If multiple continuous processes are defined, they are executed serially in
 /// the order in which they are given.
 ///
 /// For point-like processes, the passed material only affects the probability
 /// with which they occur but not the interaction itself, e.g. photon conversion
 /// into electron pairs. They are simulated by first drawing a limit on the
 /// material paths and then executing the interaction with the shortest limit
 /// when the drawn amount of material has been passed. Point-like process
 /// types **must** provide the following two member functions:
 ///
 ///     // generate X0/L0 limits
 ///     template <typename generator_t>
 ///     std::pair<Scalar, Scalar>
 ///     generatePathLimits(
 ///         generator& rng,
 ///         const Particle& particle) const
 ///
 ///     // run the process simulation
 ///     template <typename generator_t>
 ///     bool
 ///     run(
 ///         generator_t& rng,
 ///         Particle& particle,
 ///         std::vector<Particle>& generatedParticles) const
 ///
 /// For both continuous and point-like interactions, the output particle is
 /// modified in-place (if needed) and the return value indicates a break
 /// condition in the simulation, i.e. the particle is dead (true) or alive
 /// (false) after the interaction.
 ///
 /// @note If an interaction destroys the incoming particle, the process
 ///   simulation should indicate this via the break condition only and not
 ///   by reducing the particle momentum to zero. The incoming particle should
 ///   retain its initial kinematic state; the final kinematic state before
 ///   destruction is typically of more interest to the user and this simplifies
 ///   validation.
 ///
 /// The physics processes are extendable by the user to accommodate their
 /// specific requirements. While the set of available physics processes must be
 /// configured at compile-time, within that set, processes can again be
 /// selectively disabled at run-time. By default all processes are applied.
 template <typename... processes_t>
 class InteractionList {
   using Mask = std::bitset<sizeof...(processes_t)>;
   using Processes = std::tuple<processes_t...>;
   using ContinuousIndices = detail::ContinuousIndices<Processes>;
   using PointLikeIndices = detail::PointLikeIndices<Processes>;
 
  public:
   /// Point-like interaction selection.
   struct Selection {
     /// X0 (radiation length) limit for interaction selection
     double x0Limit = std::numeric_limits<double>::infinity();
     /// L0 (absorption length) limit for interaction selection
     double l0Limit = std::numeric_limits<double>::infinity();
     /// X0-based process index for point-like interactions
     std::size_t x0Process = std::numeric_limits<std::size_t>::max();
     /// L0-based process index for point-like interactions
     std::size_t l0Process = std::numeric_limits<std::size_t>::max();
   };
 
   /// Disable a specific process identified by index.
   /// @param process Index of the process to disable
   void disable(std::size_t process) { m_mask.set(process); }
   /// Disable a specific process identified by type.
   ///
   /// @note Disables only the first element, if multiple elements of the same
   ///   type exist.
   template <typename process_t>
   void disable() {
     m_mask.set(detail::TupleIndexOf<process_t, Processes>::value);
   }
 
   /// Access a specific process identified by index.
   /// @return Reference to the process at the specified index
   template <std::size_t kProcess>
   std::tuple_element_t<kProcess, Processes>& get() {
     return std::get<kProcess>(m_processes);
   }
   /// Access a specific process identified by type.
   ///
   /// @warning This function only works if all configured processes have
   ///   different types.
   /// @return Reference to the process of the specified type
   template <typename process_t>
   process_t& get() {
     return std::get<process_t>(m_processes);
   }
 
   /// Simulate the combined effects from all continuous interactions.
   ///
   /// @tparam generator_t must be a RandomNumberEngine
   /// @param[in]     rng       is the random number generator
   /// @param[in]     slab      is the passed material
   /// @param[in,out] particle  is the particle being updated
   /// @param[out]    generated is the container of generated particles
   /// @return Break condition, i.e. whether a process stopped the propagation
   template <typename generator_t>
   bool runContinuous(generator_t& rng, const Acts::MaterialSlab& slab,
                      Particle& particle,
                      std::vector<Particle>& generated) const {
     return runContinuousImpl(rng, slab, particle, generated,
                              ContinuousIndices());
   }
 
   /// Arm the point-like interactions by generating limits and select processes.
   ///
   /// @tparam generator_t must be a RandomNumberEngine
   /// @param[in] rng      is the random number generator
   /// @param[in] particle is the initial particle state
   /// @return X0/L0 limits for the particle and the process index that should be
   ///   executed once the limit has been reached.
   template <typename generator_t>
   Selection armPointLike(generator_t& rng, const Particle& particle) const {
     Selection selection;
     armPointLikeImpl(rng, particle, selection, PointLikeIndices());
     return selection;
   }
 
   /// Simulate the effects from a single point-like interaction.
   ///
   /// @tparam generator_t must be a RandomNumberEngine
   /// @param[in]     rng          is the random number generator
   /// @param[in]     processIndex is the index of the process to be executed
   /// @param[in,out] particle     is the particle being updated
   /// @param[out]    generated    is the container of generated particles
   /// @return Break condition, i.e. whether a process killed the particle
   ///
   /// The process index is expected to originate from a previous
   /// `armPointLike(...)` call, but this is not enforced. How to select the
   /// correct process requires more information that is not available here.
   template <typename generator_t>
   bool runPointLike(generator_t& rng, std::size_t processIndex,
                     Particle& particle,
                     std::vector<Particle>& generated) const {
     return runPointLikeImpl(rng, processIndex, particle, generated,
                             PointLikeIndices());
   }
 
  private:
   // allow processes to be masked. defaults to zeros -> no masked processes
   Mask m_mask;
   Processes m_processes;
 
   // for the `runContinuous` call, we need to iterate over all available
   // processes and apply the ones that implement the continuous process
   // interface. this is done using an index-based compile-time recursive call.
   template <typename generator_t, std::size_t kI0, std::size_t... kIs>
   bool runContinuousImpl(generator_t& rng, const Acts::MaterialSlab& slab,
                          Particle& particle, std::vector<Particle>& generated,
                          std::index_sequence<kI0, kIs...> /*indices*/) const {
     const auto& process = std::get<kI0>(m_processes);
     // only call process if it is not masked
     if (!m_mask[kI0] && process(rng, slab, particle, generated)) {
       // exit early in case the process signals an abort
       return true;
     }
     return runContinuousImpl(rng, slab, particle, generated,
                              std::index_sequence<kIs...>());
   }
   template <typename generator_t>
   bool runContinuousImpl(generator_t& /*rng*/,
                          const Acts::MaterialSlab& /*slab*/,
                          Particle& /*particle*/,
                          std::vector<Particle>& /*generated*/,
                          std::index_sequence<> /*indices*/) const {
     return false;
   }
 
   // for the `armPointLike` call, we need to iterate over all available
   // processes and select the ones that generate the smallest limits. this is
   // done using an index-based compile-time recursive call.
   template <typename generator_t, std::size_t kI0, std::size_t... kIs>
   void armPointLikeImpl(generator_t& rng, const Particle& particle,
                         Selection& selection,
                         std::index_sequence<kI0, kIs...> /*indices*/) const {
     // only arm the process if it is not masked
     if (!m_mask[kI0]) {
       auto [x0Limit, l0Limit] =
           std::get<kI0>(m_processes).generatePathLimits(rng, particle);
       if (x0Limit < selection.x0Limit) {
         selection.x0Limit = x0Limit;
         selection.x0Process = kI0;
       }
       if (l0Limit < selection.l0Limit) {
         selection.l0Limit = l0Limit;
         selection.l0Process = kI0;
       }
     }
     // continue with the remaining processes
     armPointLikeImpl(rng, particle, selection, std::index_sequence<kIs...>());
   }
   template <typename generator_t>
   void armPointLikeImpl(generator_t& /*rng*/, const Particle& /*particle*/,
                         Selection& /*selection*/,
                         std::index_sequence<> /*indices*/) const {}
 
   // for the `runPointLike` call we need to call just one process. since we can
   // not select a tuple element with a runtime index, we need to iterate over
   // all processes with a compile-time recursive function until we reach the
   // requested one.
   template <typename generator_t, std::size_t kI0, std::size_t... kIs>
   bool runPointLikeImpl(generator_t& rng, std::size_t processIndex,
                         Particle& particle, std::vector<Particle>& generated,
                         std::index_sequence<kI0, kIs...> /*indices*/) const {
     if (kI0 == processIndex) {
       if (m_mask[kI0]) {
         // the selected process is masked. since nothing is executed the
         // particle continues to be alive; not a break condition.
         return false;
       }
       return std::get<kI0>(m_processes).run(rng, particle, generated);
     }
     // continue the iteration with the remaining processes
     return runPointLikeImpl(rng, processIndex, particle, generated,
                             std::index_sequence<kIs...>());
   }
   template <typename generator_t>
   bool runPointLikeImpl(generator_t& /*rng*/, std::size_t /*processIndex*/,
                         Particle& /*particle*/,
                         std::vector<Particle>& /*generated*/,
                         std::index_sequence<> /*indices*/) const {
     // the requested process index is outside the possible range. **do not**
     // treat this as an error to simplify the case of an empty physics lists or
     // a default process index.
     return false;
   }
 };
 
 }  // namespace ActsFatras

This page was automatically generated by the 2.3.7 LXR engine. The LXR team