gridfire::policy::NetworkPolicy Class Reference

Abstract interface for policies that construct DynamicEngine networks from a seed composition. More...

#include <policy_abstract.h>

Inheritance diagram for gridfire::policy::NetworkPolicy:

Public Member Functions
virtual	~NetworkPolicy ()=default
virtual std::string	name () const =0
	Human-readable name for the policy.
virtual const std::set< fourdst::atomic::Species > &	get_seed_species () const =0
	Returns the seed species the policy requires to initialize the network.
virtual const reaction::ReactionSet &	get_seed_reactions () const =0
	Returns the set of seed reactions the policy requires.
virtual engine::DynamicEngine &	construct ()=0
	Construct and return a DynamicEngine instance (or engine view stack) satisfying the policy.
virtual NetworkPolicyStatus	get_status () const =0
	Returns the current verification/construction status of the policy.
virtual const std::vector< std::unique_ptr< engine::DynamicEngine > > &	get_engine_stack () const =0
virtual std::vector< engine::EngineTypes >	get_engine_types_stack () const =0
virtual const std::unique_ptr< partition::PartitionFunction > &	get_partition_function () const =0

Detailed Description

Abstract interface for policies that construct DynamicEngine networks from a seed composition.

A NetworkPolicy provides three main pieces of information:

• A name identifying the policy.
• A set of seed species and seed reactions required to initialize the network.
• A constructor method that returns a fully constructed DynamicEngine (or view stack) built to satisfy the policy.

Concrete implementations include MainSequencePolicy (see stellar_policy.h) and may throw policy-specific exceptions during construction (for example when required reactions or species are missing).

Example

// Example usage (see tests/graphnet_sandbox/main.cpp for a complete example):
// fourdst::composition::Composition comp = prepared composition
// gridfire::policy::LowMassMainSequencePolicy policy(comp);
// // construct() returns a reference to a DynamicEngine (could be a view stack)
// DynamicEngine &engine = policy.construct();
// // engine can now be passed to a solver (e.g. CVODESolverStrategy)

Constructor & Destructor Documentation

~NetworkPolicy()

virtual gridfire::policy::NetworkPolicy::~NetworkPolicy ( )

virtualdefault

Member Function Documentation

construct()

virtual engine::DynamicEngine & gridfire::policy::NetworkPolicy::construct ( )

pure virtual

Construct and return a DynamicEngine instance (or engine view stack) satisfying the policy.

Implementations typically build one or more engine layers (GraphEngine, MultiscalePartitioningEngineView, AdaptiveEngineView, etc.) and return a reference to the top-most DynamicEngine. The storage lifetime of the returned reference is implementation-defined (usually owned by the policy instance).

Returns

DynamicEngine& reference to a running/constructed engine ready for solver consumption.

Exceptions

gridfire::exceptions::MissingKeyReactionError	if required reactions are not present in the constructed network (see gridfire/exceptions/error_policy.h).
gridfire::exceptions::MissingSeedSpeciesError	if required seed species are missing from the initializing composition.
gridfire::exceptions::PolicyError	for other construction/verification failures.

Example

DynamicEngine &engine = policy.construct();
solver::CVODESolverStrategy solver(engine);
NetOut out = solver.evaluate(netIn, true);

Implemented in gridfire::policy::MainSequencePolicy, and PyNetworkPolicy.

get_engine_stack()

virtual const std::vector< std::unique_ptr< engine::DynamicEngine > > & gridfire::policy::NetworkPolicy::get_engine_stack ( ) const

pure virtual

Implemented in gridfire::policy::MainSequencePolicy, and PyNetworkPolicy.

get_engine_types_stack()

virtual std::vector< engine::EngineTypes > gridfire::policy::NetworkPolicy::get_engine_types_stack ( ) const

pure virtual

Implemented in gridfire::policy::MainSequencePolicy, and PyNetworkPolicy.

get_partition_function()

virtual const std::unique_ptr< partition::PartitionFunction > & gridfire::policy::NetworkPolicy::get_partition_function ( ) const

pure virtual

Implemented in gridfire::policy::MainSequencePolicy, and PyNetworkPolicy.

get_seed_reactions()

virtual const reaction::ReactionSet & gridfire::policy::NetworkPolicy::get_seed_reactions ( ) const

pure virtual

Returns the set of seed reactions the policy requires.

The ReactionSet describes reactions that must be present in the constructed network for the policy to be considered satisfied. Concrete policies often implement their reaction requirements by composing one or more ReactionChainPolicy instances (see chains.h).

Example

const reaction::ReactionSet &reactions = policy.get_seed_reactions();
// inspect reaction IDs or count
std::cout << "Policy requires " << reactions.size() << " reactions" << std::endl;

Implemented in gridfire::policy::MainSequencePolicy, and PyNetworkPolicy.

get_seed_species()

virtual const std::set< fourdst::atomic::Species > & gridfire::policy::NetworkPolicy::get_seed_species ( ) const

pure virtual

Returns the seed species the policy requires to initialize the network.

The returned set contains atomic species identifiers (fourdst::atomic::Species) which the policy expects to be present in the initial composition used to build the network.

Implementations should return a copy or an immutable reference to their internal set of required seed species.

Example

const auto seeds = policy.get_seed_species();
for (const auto &s : seeds) { std::cout << s.name() << std::endl; }

Implemented in gridfire::policy::MainSequencePolicy, and PyNetworkPolicy.

get_status()

virtual NetworkPolicyStatus gridfire::policy::NetworkPolicy::get_status ( ) const

pure virtual

Returns the current verification/construction status of the policy.

The status reports whether the policy has been initialized and whether the constructed network satisfies the policy's key requirements.

Example

NetworkPolicyStatus s = policy.getStatus();
if (s != NetworkPolicyStatus::INITIALIZED_VERIFIED) { // handle error }

Implemented in gridfire::policy::MainSequencePolicy, and PyNetworkPolicy.

name()

virtual std::string gridfire::policy::NetworkPolicy::name ( ) const

pure virtual

Human-readable name for the policy.

Returns

a std::string identifying the policy implementation (e.g. "MainSequencePolicy").

Example

std::string n = policy.name();
std::cout << "Using policy: " << n << std::endl;

Implemented in gridfire::policy::MainSequencePolicy, and PyNetworkPolicy.

---

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

• src/include/gridfire/policy/policy_abstract.h