feature: Export fftw c++ methods for use in other packages

[jkennel]

I've been looking for some nice wrappers for fftw for quite some time and it looks like you have done a great job with it! I have a desire to use the functions inside of some C++ code and was wondering if you would be willing export them from your package. I could generate a pull request for this if you are interested. I followed the instructions here and it worked great. Thanks for the nice package!

[dipterix]

Hi, @jkennel, thanks. A pull request is of course welcome. In the meanwhile, I'll try to export and make sure the function generates expected results in the https://github.com/dipterix/ravetools/tree/test branch

• Export cpp functions to inst/includes so users can dynamically link to ravetools
• remove the ambiguous argument (inplace) from cpp functions and make sure FFTW flags are respected
• Make sure FFTW plans are correctly scheduled to avoid data pollution (also add unit tests)
  • r2c
  • c2r
  • c2c
  • mv-r2c
  • mv-c2r
  • mv-c2c
  • r2c-2D, 3D, XD
  • c2r-2D, 3D, XD
  • c2c-2D, 3D, XD

[dipterix]

Hi, @jkennel , sorry for the late reply. I was busy working on other projects these days so there is a delay on this issue. Currently I have exposed some C++ functions so that you can dynamically link to my package. I'll gradually release other functions (I have to test them to make sure they comply with the FFTW documentation and produce the correct results in most cases).

Here are some examples to link those functions in C++ code:

library(Rcpp)

# --- FFTW r2c ----------------
test_fftw_r2c <- cppFunction(
  depends = "ravetools",
  'SEXP test_fftw_r2c(SEXP data, int HermConj = 1, int fftwplanopt = 0, SEXP ret = R_NilValue) {
    SEXP re = PROTECT(ravetools::fftw_r2c(data, HermConj, fftwplanopt, ret));
    // Do something
    UNPROTECT(1);
    return re;
  }')

data <- rnorm(1000)
ret <- complex(1000)
invisible(test_fftw_r2c(data, ret = ret))
ret

# --- Filter signals, agrees with Matlab filter function ----------
test_filter <- cppFunction(
  depends = "ravetools",
  'SEXP test(SEXP b, SEXP a, SEXP x, SEXP z) {
    SEXP re = PROTECT(ravetools::cpp_filter(b, a, x, z));
    // Do something
    UNPROTECT(1);
    return re;
  }')

t <- seq(0, 1, by = 0.01)
x <- sin(2 * pi * t * 2.3)
bf <- signal::butter(2, c(0.15, 0.3))

test_filter(bf$b, bf$a, x, rep(0, 4))

I'm trying to update ravetools package on CRAN on the monthly basis. However if you want to use the most recent features, there is a auto-builder that compiles changes on the hourly basis:

# Enable repository from dipterix
options(repos = c(
  dipterix = 'https://dipterix.r-universe.dev',
  CRAN = 'https://cloud.r-project.org'))
# Download and install ravetools in R
install.packages('ravetools')

[jkennel]

This is great! Thank you for all your efforts on this!