isabella232 / rkernel

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:

1. Run the Rwrapper
2. Connect using gRPC protocol
3. Send the init message. The init messages contains your project directory and the location of R files.
4. Once the Rwrapper receives the init message, it will source the R files.
5. 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