My proposal for Automatically-Generated R Bindings was selected as an official project under Google Summer of Code 2020. I worked with the organisation mlpack, under the mentorship of Dirk Eddelbuettel (@eddelbuettel), James Balamuta (@coatless) and Ryan Curtin(@rcurtin).
mlpack is an intuitive, fast and flexible C++ machine learning library with bindings to other languages. It is meant to be a machine learning analog to LAPACK, and aims to implement a wide array of machine learning methods and functions as a "swiss army knife" for machine learning researchers. In addition to its powerful C++ interface, mlpack also provides command-line programs, Python, Julia and Go bindings.
R is a programming language and free software environment for statistical computing and graphics supported by the R Foundation for Statistical Computing. The R language is widely used among statisticians and data miners for developing statistical software and data analysis. In this summer we aimed to implement R-bindings for mlpack, as with Python, Julia, and Go bindings for mlpack, adding a R interface would further enhance the library.
For the past 3 months, I have been implementing R binding generator. The generator produces two files for every binding, and it also generate proper documentation for every binding in man/ folder:
A .R file: This file is the R interface for R users. It uses Rcpp in order to sharing data and communicate with C++. When generated, the .R file is build in the mlpack/binding/R/mlpack/src directory.
A .cpp file: This file is used as a C++ interface. This file includes the function which can be called by R.
Additionally, a method specific library is created the method_main.cpp in order for R to call the mlpack_main() function.
Here's an example usage:
R version 4.0.2 (2020-06-22) -- "Taking Off Again"
Copyright (C) 2020 The R Foundation for Statistical Computing
Platform: x86_64-pc-linux-gnu (64-bit)
R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.
Natural language support but running in an English locale
R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.
Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for an HTML browser interface to help.
Type 'q()' to quit R.
> library(mlpack)
> x <- matrix(rexp(1000, rate=.1), nrow=10)
> pca(x, verbose=TRUE, new_dimensionality=5)
[INFO ] Performing PCA on dataset...
[INFO ] 72.0016% of variance retained (5 dimensions).
$output
[,1] [,2] [,3] [,4] [,5]
[1,] 111.667141 -11.65468 12.0705689 -8.472432 19.570236
[2,] -20.948960 14.12169 -18.4961187 27.159226 -7.702679
[3,] -61.017751 -81.38778 -17.9509279 -20.468096 16.520563
[4,] -18.417396 29.84052 41.0875320 -8.433435 48.420332
[5,] 18.359201 -28.26317 -0.3566754 -58.028579 -7.920467
[6,] -39.205519 29.39570 75.0991434 1.771608 -7.283641
[7,] -5.009963 45.03162 -44.8698549 -1.245450 15.697076
[8,] -6.844800 32.32820 -62.9257916 7.394938 5.893395
[9,] 4.144613 14.69571 4.6391430 -18.892109 -73.807397
[10,] 17.273434 -44.10780 11.7029811 79.214330 -9.387417In order to get documentation about a method, you can use:
> ?mlpack::pcaor
> library(mlpack)
> ?pcaHere is an example for the auto-generated R file:https://gist.github.com/4967c5939ba054bc868415d1e3e5c956
If you want to build/generate these bindings and run them, you can follow this procedure:
- Download R from https://cran.r-project.org and install all dependencies required by the bindings using:
Rscript -e "install.packages(c('Rcpp', 'RcppArmadillo', 'RcppEnsmallen', 'BH', 'roxygen2', 'testthat'))"- Configure mlpack with
-DBUILD_R_BINDINGS=ON(optionally-DBUILD_MARKDOWN_BINDINGS=ONif you want to see that output). - Build with make R, and if you want make test will test the R bindings also.
- Once the R bindings are built, then you can run them by (from the build directory)
cd src/mlpack/bindings/R/mlpack/&&R CMD INSTALL .then you can dolibrary(mlpack)and play with it.
Currently for Direct Installation you can use:
devtools::install_github('Yashwants19/RcppMLPACK')From mlpack's build system you can use:
make installor, Like discussed earlier after building R bindings for mlpack, you can run them using (from the build directory):
cd src/mlpack/bindings/R/mlpack/
R CMD INSTALLFinally, to test the binding, testthat is used. Simply use:
ctest -R r_binding_test --verboseForce CMake to show error when it didn't find python/modules.
(GSoC Week 11 -12) Refactor ProgramInfo() to separate out all the different information.
Fix ModelType Julia Documentation.
Fix pointer error in Go-bindings
🚀 Add Go bindings for some missed models.
You can follow my weekly progress on my medium blogs. Medium Articles by yashwantsingh.sngh
I have no words to describe how grateful I am to my mentors - Dirk Eddelbuettel, James Balamuta and Ryan Curtin - for all their support and for being extremely responsive and helpful with every one of my problems. Without their vote of confidence, this project would've been a lot harder and a lot less fun to do.
I am also thankful to mlpack and Google for the opportunity to work on this project, which helped me learn a lot in such a short period of time.