janus::Function Class Reference

Wrapper around casadi::Function providing Eigen-native IO. More...

#include <Function.hpp>

Public Member Functions
	Function (const std::string &name, const std::vector< SymbolicArg > &inputs, const std::vector< SymbolicArg > &outputs)
	Construct a new Function object.
	Function (const std::vector< SymbolicArg > &inputs, const std::vector< SymbolicArg > &outputs)
	Constructor with auto-generated name.
template<typename... Args>
auto	operator() (Args &&...args) const
	Evaluate function with arbitrary arguments (scalars or Eigen matrices) Handles both numeric (double/MatrixXd) and symbolic (MX/SymbolicMatrix) inputs.
template<typename... Args>
auto	eval (Args &&...args) const
	Evaluate function and return first output.
std::vector< NumericMatrix >	operator() (std::vector< double > &args) const
	Evaluate with a mutable vector of doubles.
std::vector< NumericMatrix >	operator() (const std::vector< double > &args) const
	Evaluate with a const vector of doubles.
const casadi::Function &	casadi_function () const
	Access the underlying CasADi function.
Function	map (int n, MapParallelization parallelization=MapParallelization::Parallel) const
	Create a batched version of this function using CasADi's map primitive.
Function	map (int n, const std::string &parallelization) const
	Create a batched version using a direct CasADi backend name.
Function	map (int n, MapParallelization parallelization, int max_num_threads) const
	Create a batched version of this function with an explicit thread cap.
Function	map (int n, const std::string &parallelization, int max_num_threads) const
	Create a batched version with explicit backend name and thread cap.

Detailed Description

Wrapper around casadi::Function providing Eigen-native IO.

See also

SymbolicArg, make_function

Constructor & Destructor Documentation

Function() [1/2]

janus::Function::Function ( const std::string & name, const std::vector< SymbolicArg > & inputs, const std::vector< SymbolicArg > & outputs )

inline

Construct a new Function object.

Parameters

name	Name of the function
inputs	Vector of symbolic arguments (scalars or matrices) describing the inputs
outputs	Vector of symbolic arguments (scalars or matrices) describing the outputs

Function() [2/2]

janus::Function::Function ( const std::vector< SymbolicArg > & inputs, const std::vector< SymbolicArg > & outputs )

inline

Constructor with auto-generated name.

Parameters

inputs	Vector of symbolic arguments (scalars or matrices) describing the inputs
outputs	Vector of symbolic arguments (scalars or matrices) describing the outputs

Member Function Documentation

casadi_function()

const casadi::Function & janus::Function::casadi_function ( ) const

inline

Access the underlying CasADi function.

Returns

const reference to casadi::Function

eval()

template<typename... Args>

auto janus::Function::eval ( Args &&... args ) const

inline

Evaluate function and return first output.

Template Parameters

Args	Argument types

Parameters

args	Variadic arguments matching the function inputs

Returns

First output matrix

Examples

/home/runner/work/janus/janus/include/janus/core/Sparsity.hpp.

map() [1/4]

Function janus::Function::map ( int n, const std::string & parallelization ) const

inline

Create a batched version using a direct CasADi backend name.

Parameters

n	Number of batch samples
parallelization	Backend string ("openmp", "serial", or "unroll")

Returns

Batched Janus function wrapper

map() [2/4]

Function janus::Function::map ( int n, const std::string & parallelization, int max_num_threads ) const

inline

Create a batched version with explicit backend name and thread cap.

Parameters

n	Number of batch samples
parallelization	Backend string ("openmp", "serial", or "unroll")
max_num_threads	Maximum worker threads

Returns

Batched Janus function wrapper

map() [3/4]

Function janus::Function::map ( int n, MapParallelization parallelization, int max_num_threads ) const

inline

Create a batched version of this function with an explicit thread cap.

Parameters

n	Number of batch samples
parallelization	Mapping backend selection
max_num_threads	Maximum number of worker threads for supported backends

Returns

Function Batched Janus function wrapper

map() [4/4]

Function janus::Function::map ( int n, MapParallelization parallelization = MapParallelization::Parallel ) const

inline

Create a batched version of this function using CasADi's map primitive.

For a function f(a, b) -> (y), map(n) returns F(A, B) -> (Y) where each input/output is the horizontal concatenation of n samples: A = horzcat([a0, a1, ..., a_(n-1)]), Y = horzcat([y0, y1, ..., y_(n-1)]).

The default backend is MapParallelization::Parallel, which requests CasADi's OpenMP mapper for parallel batch evaluation. CasADi falls back to serial if it was built without OpenMP support.

Use this for batch evaluation over multiple independent samples while preserving CasADi's derivative propagation through the mapped graph.

Parameters

n	Number of batch samples
parallelization	Mapping backend selection

Returns

Function Batched Janus function wrapper

operator()() [1/3]

template<typename... Args>

auto janus::Function::operator() ( Args &&... args ) const

inline

Evaluate function with arbitrary arguments (scalars or Eigen matrices) Handles both numeric (double/MatrixXd) and symbolic (MX/SymbolicMatrix) inputs.

Template Parameters

Args	Argument types

Parameters

args	Variadic arguments matching the function inputs

Returns

Vector of result matrices (NumericMatrix or SymbolicMatrix)

operator()() [2/3]

std::vector< NumericMatrix > janus::Function::operator() ( const std::vector< double > & args ) const

inline

Evaluate with a const vector of doubles.

Parameters

args	Input values

Returns

Vector of numeric result matrices

operator()() [3/3]

std::vector< NumericMatrix > janus::Function::operator() ( std::vector< double > & args ) const

inline

Evaluate with a mutable vector of doubles.

Parameters

args	Input values

Returns

Vector of numeric result matrices

---

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

• include/janus/core/Function.hpp