Skip to contents

Package Management

Source: vignettes/pkg.Rmd
pkg.Rmd

Overview

The pkg module provides two functions:

Function Purpose
set_mirror() Configure CRAN / Bioconductor download mirrors
pkg_functions() List exported functions from any installed package

For installation, updates, and status checks, we recommend pak — it provides a more elegant and performant workflow with parallel downloads and a unified interface for CRAN, GitHub, and Bioconductor.

Note: All code examples in this vignette are static (eval = FALSE).

1 Mirror Configuration

set_mirror() — Set CRAN and Bioconductor mirrors

Configures the CRAN and/or Bioconductor download mirrors. Settings are respected by R’s native install functions and by pak. The previous settings are returned invisibly, making it easy to restore them.

# Set both mirrors to Tsinghua (default)
set_mirror()
#> v CRAN mirror set to: https://mirrors.tuna.tsinghua.edu.cn/CRAN
#> v Bioconductor mirror set to: https://mirrors.tuna.tsinghua.edu.cn/bioconductor

# CRAN only
set_mirror("cran", "ustc")
#> v CRAN mirror set to: https://mirrors.ustc.edu.cn/CRAN

# Bioconductor only
set_mirror("bioc", "nju")
#> v Bioconductor mirror set to: https://mirrors.nju.edu.cn/bioconductor

Available mirrors:

Mirror CRAN Bioconductor
official v v
tuna v v
ustc v v
westlake v v
nju v v
rstudio v -
aliyun v -
sjtu v -
pku v -
hku v -
sustech v -

CRAN-only mirrors cannot be used with repo = "all":

set_mirror("all", "aliyun")
#> Error in `set_mirror()`:
#> ! Mirror "aliyun" is CRAN-only and cannot be used with `repo = "all"`.
#> i Use `set_mirror("cran", "aliyun")` instead.
#> i Shared mirrors: "official", "tuna", "ustc", "westlake", "nju"

The previous settings are returned invisibly for easy restoration:

prev <- set_mirror("cran", "tuna")
# ... do work ...
options(prev)  # restore original mirror

2 Package Exploration

pkg_functions() — List exported functions

Returns an alphabetically sorted character vector of exported symbols from any installed package. Use key to filter by a case-insensitive keyword.

pkg_functions("evanverse")
#>  [1] "%is%"          "%map%"         "%match%"       "%nin%"
#>  [5] "%p%"           "comb"          "df2list"       "df2vect"
#>  ...
# Filter by keyword (case-insensitive)
pkg_functions("evanverse", key = "plot")
#> [1] "plot_bar"     "plot_density" "plot_forest"  "plot_pie"
#> [5] "plot_venn"

pkg_functions("stats", key = "test")
#> [1] "ansari.test"   "bartlett.test" "binom.test"
#> [4] "chisq.test"    "cor.test"      "fisher.test"
#> ...

An unmatched keyword returns an empty vector rather than an error:

pkg_functions("stats", key = "zzzzzz")
#> character(0)

Querying an uninstalled package raises an informative error:

pkg_functions("somefakepkg")
#> Error in `pkg_functions()`:
#> ! Package `somefakepkg` is not installed.

We found that pak provides a more elegant and performant package management workflow. It supports parallel installation, automatic dependency resolution, and a unified interface for all package sources — which is why inst_pkg(), check_pkg(), update_pkg(), and pkg_version() were removed from evanverse.

# Install pak
install.packages("pak")

# Set a fast mirror first (evanverse)
set_mirror("all", "tuna")

# Install from any source with a unified interface
pak::pkg_install("ggplot2")                              # CRAN
pak::pkg_install("tidyverse/ggplot2")                    # GitHub
pak::pkg_install("bioc::DESeq2")                         # Bioconductor
pak::pkg_install(c("dplyr", "bioc::limma", "r-lib/pak")) # mixed

# Check status
pak::pkg_status(c("ggplot2", "DESeq2"))

# Update all or specific packages
pak::pkg_update()
pak::pkg_update("ggplot2")

Getting Help