R: Conda Tools

conda-tools {reticulate}

R Documentation

Conda Tools

Description

Tools for managing Python conda environments.

Usage

conda_list(conda = "auto")

conda_create(
  envname = NULL,
  packages = NULL,
  ...,
  forge = TRUE,
  channel = character(),
  environment = NULL,
  conda = "auto",
  python_version = miniconda_python_version()
)

conda_clone(envname, ..., clone = "base", conda = "auto")

conda_export(
  envname,
  file = if (json) "environment.json" else "environment.yml",
  json = FALSE,
  ...,
  conda = "auto"
)

conda_remove(envname, packages = NULL, conda = "auto")

conda_install(
  envname = NULL,
  packages,
  forge = TRUE,
  channel = character(),
  pip = FALSE,
  pip_options = character(),
  pip_ignore_installed = FALSE,
  conda = "auto",
  python_version = NULL,
  ...
)

conda_binary(conda = "auto")

conda_exe(conda = "auto")

conda_version(conda = "auto")

conda_update(conda = "auto")

conda_python(envname = NULL, conda = "auto", all = FALSE)

Arguments

`conda`	The path to a `conda` executable. Use `"auto"` to allow `reticulate` to automatically find an appropriate `conda` binary. See Finding Conda and `conda_binary()` for more details.
`envname`	The name of, or path to, a conda environment.
`packages`	A character vector, indicating package names which should be installed or removed. Use `python=<version>` to request the installation of a specific version of Python.
`...`	Optional arguments, reserved for future expansion.
`forge`	Boolean; include the conda-forge repository?
`channel`	An optional character vector of conda channels to include. When specified, the `forge` argument is ignored. If you need to specify multiple channels, including the conda forge, you can use `c("conda-forge", <other channels>)`.
`environment`	The path to an environment definition, generated via (for example) `conda_export()`, or via `conda env export`. When provided, the conda environment will be created using this environment definition, and other arguments will be ignored.
`python_version`	The version of Python to be installed. Set this if you'd like to change the version of Python associated with a particular conda environment.
`clone`	The name of the conda environment to be cloned.
`file`	The path where the conda environment definition will be written.
`json`	Boolean; should the environment definition be written as JSON? By default, conda exports environments as YAML.
`pip`	Boolean; use `pip` for package installation? By default, packages are installed from the active conda channels.
`pip_options`	An optional character vector of additional command line arguments to be passed to `pip`. Only relevant when `pip = TRUE`.
`pip_ignore_installed`	Ignore already-installed versions when using pip? (defaults to `FALSE`). Set this to `TRUE` so that specific package versions can be installed even if they are downgrades. The `FALSE` option is useful for situations where you don't want a pip install to attempt an overwrite of a conda binary package (e.g. SciPy on Windows which is very difficult to install via pip due to compilation requirements).
`all`	Boolean; report all instances of Python found?

Value

conda_list() returns an R data.frame, with name giving the name of the associated environment, and python giving the path to the Python binary associated with that environment.

conda_create() returns the path to the Python binary associated with the newly-created conda environment.

conda_clone() returns the path to Python within the newly-created conda environment.

conda_export() returns the path to the exported environment definition, invisibly.

Finding Conda

Most of reticulate's conda APIs accept a conda parameter, used to control the conda binary used in their operation. When conda = "auto", reticulate will attempt to automatically find a conda installation. The following locations are searched, in order:

The location specified by the reticulate.conda_binary R option,
The location specified by the RETICULATE_CONDA environment variable,
The miniconda_path() location (if it exists),
The program PATH,
A set of pre-defined locations where conda is typically installed.

To force reticulate to use a particular conda binary, we recommend setting:

options(reticulate.conda_binary = "/path/to/conda")

This can be useful if your conda installation lives in a location that reticulate is unable to automatically discover.

[Package reticulate version 1.25 Index]