Skip to contents

Set up the Python environment for multimod

Source: R/setup_python.R
mm_setup_python.Rd

This is the installer: one Conda environment, shared by the whole stack. energyRt resolves its Python through get_multimod_python(), so there is one environment and one source of truth rather than a per-package env.

Usage

mm_setup_python(
  env_name = "multimod",
  packages = c("pyomo", "highspy", "pyarrow", "pandas"),
  conda_packages = "glpk",
  python_version = "3.14",
  ensure_miniconda = TRUE,
  reinstall = FALSE,
  pip = TRUE,
  ...
)

Arguments

env_name

Name of the Conda environment to manage. Defaults to "multimod" so the generated Pyomo artifacts have a predictable runtime.

packages

Character vector of Python packages to install once the environment exists. The default covers every route multimod writes: pyomo and highspy to build and solve, pyarrow and pandas to read the Arrow data save_model(format = "ipc") writes. Set to NULL or character(0) to skip installation.

conda_packages

Character vector installed with conda rather than pip. glpk by default: the MathProg route needs the glpsol executable, which pip does not ship.

python_version

Python version spec passed to reticulate::conda_create(). Use NULL for reticulate's default.

ensure_miniconda

Logical flag controlling whether reticulate::install_miniconda() should run when no Conda binary is detected on the system. Defaults to TRUE to minimize user setup.

reinstall

Logical flag; when TRUE the environment is recreated even if it already exists. Defaults to FALSE to preserve prior installs.

pip

Logical; when TRUE (default) packages are installed via pip inside the Conda environment (ensures highspy is available). Set to FALSE to install via conda instead.

...

Additional arguments forwarded to reticulate::install_miniconda() when ensure_miniconda = TRUE.

Value

Invisibly returns the environment name once configured.

Details

It bootstraps a Conda environment (named "multimod" by default) using the reticulate toolkit. It ensures reticulate::install_miniconda() has been invoked (unless disabled), creates the target environment when missing, installs the requested Python packages, and activates the environment for the current R session via reticulate::use_condaenv().

After setup, multimod uses this environment either when it is set with set_multimod_python() or when get_multimod_python() detects it through reticulate.

See also

get_multimod_python(), set_multimod_python()