Title:  C++ Header Files for Stan 

Description:  The C++ header files of the Stan project are provided by this package, but it contains little R code or documentation. The main reference is the vignette. There is a shared object containing part of the 'CVODES' library, but its functionality is not accessible from R. 'StanHeaders' is primarily useful for developers who want to utilize the 'LinkingTo' directive of their package's DESCRIPTION file to build on the Stan library without incurring unnecessary dependencies. The Stan project develops a probabilistic programming language that implements full or approximate Bayesian statistical inference via Markov Chain Monte Carlo or 'variational' methods and implements (optionally penalized) maximum likelihood estimation via optimization. The Stan library includes an advanced automatic differentiation scheme, 'templated' statistical and linear algebra functions that can handle the automatically 'differentiable' scalar types (and doubles, 'ints', etc.), and a parser for the Stan language. The 'rstan' package provides userfacing R functions to parse, compile, test, estimate, and analyze Stan models. 
Authors:  Ben Goodrich [cre, aut], Joshua Pritikin [ctb], Andrew Gelman [aut], Bob Carpenter [aut], Matt Hoffman [aut], Daniel Lee [aut], Michael Betancourt [aut], Marcus Brubaker [aut], Jiqiang Guo [aut], Peter Li [aut], Allen Riddell [aut], Marco Inacio [aut], Mitzi Morris [aut], Jeffrey Arnold [aut], Rob Goedman [aut], Brian Lau [aut], Rob Trangucci [aut], Jonah Gabry [aut], Alp Kucukelbir [aut], Robert Grant [aut], Dustin Tran [aut], Michael Malecki [aut], Yuanjun Gao [aut], Hamada S. Badr [aut] , Trustees of Columbia University [cph], Lawrence Livermore National Security [cph] (CVODES), The Regents of the University of California [cph] (CVODES), Southern Methodist University [cph] (CVODES) 
Maintainer:  Ben Goodrich <[email protected]> 
License:  BSD_3_clause + file LICENSE 
Version:  2.35.0.9000 
Built:  20240828 13:34:46 UTC 
Source:  https://github.com/standev/rstan 
Output the compiler or linker flags required to build with the StanHeaders package
CxxFlags(as_character = FALSE) LdFlags(as_character = FALSE)
CxxFlags(as_character = FALSE) LdFlags(as_character = FALSE)
as_character 
A logical scalar that defaults to 
These functions are currently not exported and are typically called from a Makevars or a Makevars.win file of another package as follows:
PKG_CXXFLAGS += $(shell "${R_HOME}/bin/Rscript" e "StanHeaders:::CxxFlags()")
PKG_LDLIBS += $(shell "${R_HOME}/bin/Rscript" e "StanHeaders:::LdFlags()")
If as_character
is TRUE
, then these functions return
a character vector of length one. Otherwise, (which is the default) these
functions return NULL
invisibly after outputing the compiler
or linker flags to the screen.
Call a function defined in the Stan Math Library from R using this wrapper around
cppFunction
.
stanFunction(function_name, ..., env = parent.frame(), rebuild = FALSE, cacheDir = getOption("rcpp.cache.dir", tempdir()), showOutput = verbose, verbose = getOption("verbose"))
stanFunction(function_name, ..., env = parent.frame(), rebuild = FALSE, cacheDir = getOption("rcpp.cache.dir", tempdir()), showOutput = verbose, verbose = getOption("verbose"))
function_name 
A 
... 
Further arguments that are passed to 
env , rebuild , cacheDir , showOutput , verbose

The same as in 
The stanFunction
function essentially compiles and
evaluates a C++ function of the form
auto function_name(...) { return stan::math::function_name(...); }
It is essential to pass all arguments to function_name
through the ...
in order for the C++ wrapper to know what the argument types are. The mapping
between R types and Stan types is
R type  Stan type 
double 
real 
integer 
int 
complex 
complex 
vector 
vector or complex_vector 
matrix(*, nrow = 1) 
row_vector or complex_row_vector 
matrix 
matrix or complex_matrix

and, in addition, lists of the aforementioned R types map to arrays of Stan types and thus must not be ragged if they are nested. The Stan version of the function is called with arguments specified by position, i.e. in the order that they appear in the .... However, the R wrapper function has arguments whose names are the same as the names passed through the ....
The result of function_name
evaluated at the arguments
that are passed through the ..., which could be of various
R types. It also has the side effect of defining a function
named function_name
in the environment given by the
env
argument that can subsequently be called with
inputs of the same type (but not necessarily the same value)
that were passed through the ....
files < dir(system.file("include", "stan", "math", "prim", package = "StanHeaders"), pattern = "hpp$", recursive = TRUE) functions < sub("\\.hpp$", "", sort(unique(basename(files[dirname(files) != "."])))) length(functions) # you could call most of these Stan functions ## Not run: log(sum(exp(exp(1)), exp(pi))) # true value stanFunction("log_sum_exp", x = exp(1), y = pi) args(log_sum_exp) # now exists in .GlobalEnv log_sum_exp(x = pi, y = exp(1)) # but log_sum_exp() was not defined for a vector or matrix x < c(exp(1), pi) try(log_sum_exp(x)) stanFunction("log_sum_exp", x = x) # now it is # log_sum_exp() is now also defined for a matrix log_sum_exp(as.matrix(x)) log_sum_exp(t(as.matrix(x))) log_sum_exp(rbind(x, x)) # but log_sum_exp() was not defined for a list try(log_sum_exp(as.list(x))) stanFunction("log_sum_exp", x = as.list(x)) # now it is # in rare cases, passing a nested list is needed stanFunction("dims", x = list(list(1:3))) # functions of complex arguments work stanFunction("eigenvalues", # different ordering than base:eigen() x = matrix(complex(real = 1:9, imaginary = pi), nrow = 3, ncol = 3)) # nullary functions work but are not that interesting stanFunction("negative_infinity") # PRNG functions work by adding a seed argument stanFunction("lkj_corr_rng", K = 3L, eta = 1) args(lkj_corr_rng) # has a seed argument ## End(Not run)
files < dir(system.file("include", "stan", "math", "prim", package = "StanHeaders"), pattern = "hpp$", recursive = TRUE) functions < sub("\\.hpp$", "", sort(unique(basename(files[dirname(files) != "."])))) length(functions) # you could call most of these Stan functions ## Not run: log(sum(exp(exp(1)), exp(pi))) # true value stanFunction("log_sum_exp", x = exp(1), y = pi) args(log_sum_exp) # now exists in .GlobalEnv log_sum_exp(x = pi, y = exp(1)) # but log_sum_exp() was not defined for a vector or matrix x < c(exp(1), pi) try(log_sum_exp(x)) stanFunction("log_sum_exp", x = x) # now it is # log_sum_exp() is now also defined for a matrix log_sum_exp(as.matrix(x)) log_sum_exp(t(as.matrix(x))) log_sum_exp(rbind(x, x)) # but log_sum_exp() was not defined for a list try(log_sum_exp(as.list(x))) stanFunction("log_sum_exp", x = as.list(x)) # now it is # in rare cases, passing a nested list is needed stanFunction("dims", x = list(list(1:3))) # functions of complex arguments work stanFunction("eigenvalues", # different ordering than base:eigen() x = matrix(complex(real = 1:9, imaginary = pi), nrow = 3, ncol = 3)) # nullary functions work but are not that interesting stanFunction("negative_infinity") # PRNG functions work by adding a seed argument stanFunction("lkj_corr_rng", K = 3L, eta = 1) args(lkj_corr_rng) # has a seed argument ## End(Not run)