Rkernel

Rkernel is a program responsible for handling various types of requests: code execution, code completions, debugging, and so on. Is also provides a response from the R interpreter.

The idea of Rkernel is pretty similar to gdb or Jupyter kernel but it aimed for R.

Building

1 Build gRPC

On Windows: build_grpc.bat
On GNU/Linux or MacOS: build_grpc.sh

2 Use `cmake` with your toolchain to build the project

Ensure that the cmake R_HOME variable is defined or suitable R version is present in the PATH variable. Use the following toolchains:

On Windows: MSVS 2017 and ninja (you probably have to use dll2lib.R for generating lib files and apply patches from Rcpp-patches)
On MacOS: clang 8+ and make
On GNU/Linux: gcc 7.0+ and make

To use other configurations, you may need to edit the CMakeLists.txt file.

Running

Rkernel consists of:

Rwrapper: responsible for running gRPC server and handling interaction with R interpreter.
R files: files that are located in the R directory of the project root.

For interacting with Rwrapper you should implement a gRPC client for the protocol defined in Rkernel-proto

To start interacting with Rwrapper:

Run the Rwrapper
Connect using gRPC protocol
Send the init message. The init messages contains your project directory and the location of R files.
Once the Rwrapper receives the init message, it will source the R files.
Use the message from the protocol for communication. For more details about the protocol, see the Rkernel-proto project. You could find an example of the client implementation here.

Licensing

Rkernel-proto is covered by Apache License 2.0
R/RSession is covered by AGPL-3.0
the rest of the project is covered by GPL-3.0