rpymat 0.1.9

• Updated Python version registry for MATLAB
• conda environment forces to install pip
• Added libomp fix: users will no longer encounter the error about double-initializing OpenMP by setting KMP_DUPLICATE_LIB_OK to TRUE as a compromise. However, manually run fix_omp_conflict is recommended.
• Added fix_omp_conflict to creates symbolic link on MacOS and replace the conda built-in version: setting to run manually. This fix requires users to run fix_omp_conflict whenever R is updated, and the OpenMP ABI versions must coincide between R’s and conda.

rpymat 0.1.8

• Supports multiple environments for incompatible packages
• Fixed windows dll modification issue
• Added function to accept and reject conda term-of-service
• Force configuring conda no longer download miniconda of already installed

rpymat 0.1.7

• Set longer time-out when installing conda
• Allow to read xlsx from pandas (if user choose to install)
• Fixed a command issue on linux where dot-dot cannot be used on file path to get the parent directory; using R’s built-in dirname instead
• Fixed a jupyter dry-run issue
• Allowed to set base URL for jupyter
• Allow to manually override jupyter token with JUPYTER_TOKEN environment variable instead of randomly-generated tokens. This setting is mainly designed for docker.

rpymat 0.1.6

• Fixed an installation problem introduced in version 0.1.5
• Added utility functions for basic Python types
• Added global variables in run_pyscript

rpymat 0.1.5

• Added choose_directory, choose_fileopen, choose_filesave to select files using either base R or Python (via tcl-tk but with better interface on osx)
• Ported reticulate functions to convert between R and Python objects
• Allowed to set conda to installed version via system environment R_RPYMAT_CONDA_EXE and R_RPYMAT_CONDA_PREFIX; this change will allow conda-forge library r-rpymat to use already-installed conda
• Speed-up ensure_rpymat by caching the configuration

rpymat 0.1.4

• Muffled conda information when running Python code or scripts
• Ported reticulate::repl_python function
• Allowed users to list installed packages
• Allowed jupyter to run in background in RStudio jobs
• Removed optional jupyter add-on packages from the installation script

rpymat 0.1.3

• Fixed command-line script failing to load conda
• Allowed to use the miniconda installed at user’s home directory
• Use slash instead of backslash on Windows system to avoid escaping issues

rpymat 0.1.2

• Fixed BLAS issues on some Unix systems
• Added Windows support
• Added support to install, configure, launch, and stop Jupyterlab
• Added run_command to invoke enhanced system command under conda environment

rpymat 0.1.1

• Added a NEWS.md file to track changes to the package.