executes()

template<typename Func >

RuleBuilder & MayaFlux::Yantra::ComputationGrammar::RuleBuilder::executes ( Func &&  executor )

inline

Sets the executor function for this rule.

Template Parameters

Func	Function type (usually a lambda)

Parameters

executor

Function that performs the computation

Returns

Reference to this builder for method chaining

The executor function receives the input data and execution context, and returns the result of the computation. This is where the actual work of the rule is performed.

If with_gpu_backend() was called, the executor wrapper attaches a ShaderExecutionContext to the operation before invoking the user lambda. The user lambda receives the already-GPU-backed operation transparently and serves as the CPU fallback if GPU setup fails.

Definition at line 416 of file ComputeGrammar.hpp.

        {
            if (m_rule.gpu_spec.has_value()) {
                auto spec = *m_rule.gpu_spec;
                m_rule.executor = [func = std::forward<Func>(executor), spec = std::move(spec)](
                                      const std::any& input, const ExecutionContext& ctx) -> std::any {
                    const auto cfg = config_from_spec(spec);
                    const auto bindings = bindings_from_spec(spec);
 
                    if (cfg.shader_id != Portal::Graphics::INVALID_SHADER) {
                        auto gpu_exec = std::make_shared<ShaderExecutionContext<>>(cfg, bindings);
                        ExecutionContext patched = ctx;
                        patched.execution_metadata["_gpu_exec"] = gpu_exec;
                        return func(input, patched);
                    }
                    return func(input, ctx);
                };
            } else {
                m_rule.executor = [func = std::forward<Func>(executor)](
                                      const std::any& input, const ExecutionContext& ctx) -> std::any {
                    return func(input, ctx);
                };
            }
            return *this;
        }

References MayaFlux::Yantra::bindings_from_spec(), MayaFlux::Yantra::config_from_spec(), MayaFlux::Yantra::ExecutionContext::execution_metadata, MayaFlux::Yantra::ComputationGrammar::Rule::executor, MayaFlux::Yantra::ComputationGrammar::Rule::gpu_spec, and input.

Here is the call graph for this function: