Simple interpolation functions designed to be used from C. There is
essentially no R support in this package, save code designed to be used
by the package’s own testing. cinterpolate
is designed to
be used in package code only and modification for use outside of a
package is not explicitly supported.
cinterpolate
is now on CRAN and can be installed
with
install.packages("cinterpolate")
After that, install with
install.packages(
"orderly2",
repos = c("https://mrc-ide.r-universe.dev", "https://cloud.r-project.org"))
If you install from source, this package requires a fortran compiler because it uses LAPACK and BLAS, despite being only C code.
apt-get install gfortran
depending on your platform#include <cinterpolate/cinterpolate.h>
Allocate an object to perform interpolation with. nx
and
x
are the size and values of the function to be
interpolated, while ny
is the number of functions to be
simultaneously (but independently) interpolateed and y
is
the values for these. If ny > 1
, then the values of
y
are as an R matrix with nx
rows and
ny
columns (the first nx
values are the first
function, the second nx
are the second, and so on). The
final two are booleans: fail_on_extrapolate
throws an error
if extrapolation is requested, and auto_free
uses
R_alloc()
rather than Calloc
allowing
automatic cleanup.
void *obj = cinterpolate_alloc(type, nx, ny, x, y,
, auto_free); fail_on_extrapolate
With the interpolation function, an input value of xout
and given some storage yout
of length ny
,
determine f(xout)
with
(xout, obj, yout); cinterpolate_eval
Once done the object must be freed with
cinterpolate_free(obj);
Somewhere in the package you must include cinterpolate.c
as
#include <cinterpolate/cinterpolate.h>
but this must be included only once or linking will
fail. If you only use the interpolation in one place, you can use
cinterpolate.c
rather than cinterpolate.h
. If
you use interpolation in more than once place, I recommend a file
src/cinterpolate.c
containing
#include <cinterpolate/cinterpolate.c>
Be sure to include in your DESCRIPTION
a line
LinkingTo: cinterpolate
cinterpolate
takes advantage of BLAS for matrix
multiplication, so you may need a Makevars that includes
PKG_LIBS = ${LAPACK_LIBS} ${BLAS_LIBS} ${FLIBS}
See inst/example
for a full
example of using cinterpolate
within a package
MIT + file LICENSE © Imperial College of Science, Technology and Medicine
Please note that this project is released with a Contributor Code of Conduct. By participating in this project you agree to abide by its terms.