Skip to contents

bascule

bascule.Rmd

Installing the package

The bascule R package can be installed directly from GitHub using devtools:

devtools::install_github("caravagnalab/bascule")

Once installed, load the package in your R session:

Installing Python dependencies

The bascule package relies on the Python module pybascule. The recommended approach is to install these Python dependencies inside a dedicated conda environment.

This can be done from R using the reticulate package through the following steps.

  • if Miniconda is not already available on your system, you can install it using through reticulate:
install_miniconda(path = miniconda_path(), update = TRUE, force = FALSE)
  • create a new conda environment with Python version 3.10:
reticulate::conda_create(envname="bascule", python_version="3.10")
  • activate the new conda environment within the current R session:
reticulate::use_condaenv("bascule")
  • install the pybascule Python module using pip inside the conda environment:
reticulate::conda_install(envname="bascule", packages="pybascule", pip=TRUE)

Fitting a bascule object

Before fitting a bascule object with the fit() or the fit_clustering() functions, make sure that the correct conda environment is active so that the pybascule module can be accessed:

reticulate::use_condaenv("bascule")

Functions that rely on pybascule include an additional argument, py, which allows you to explicitly pass the imported Python module. This ensures that the correct version of pybascule is used.

reticulate::use_condaenv("bascule")
py = reticulate::import("pybascule")
x = fit(..., py=py)
# or
x = fit_clustering(..., py=py)