{rgeowheels}

Downloads pre-compiled Windows ‘wheel’ files (.whl) for Python geospatial packages from https://github.com/cgohlke/geospatial-wheels. These are unofficial binary installers for some Python geospatial libraries on Windows prepared by Christoph Gohlke. Wheels for various packages, Python versions, and architectures are made available via GitHub releases, providing the ability to revert to prior versions when needed.

Key Features:

• Automatic Python Version Detection: Use pyversion = "auto" to automatically detect and match your Python version
• Virtual Environment Support: Works seamlessly with Python virtual environments and conda environments
• Environment Discovery: Discover available Python environments with detect_python_envs()
• CI/CD Friendly: Perfect for GitHub Actions and other CI workflows on Windows
• Simplified Installation: Install complex geospatial libraries in seconds

Installation

You can install the development version of {rgeowheels} like so:

# install.packages("remotes")
# remotes::install_github("brownag/rgeowheels")
install.packages("rgeowheels", repos = c('https://brownag.r-universe.dev', 'https://cloud.r-project.org'))

Quick Start

The simplest way to install a wheel is to use automatic Python version detection:

library(rgeowheels)

install_wheel("GDAL", pyversion = "auto")

For more details, see the vignette: vignette("installing-gdal-windows")

Examples

List available packages and versions

library(rgeowheels)
#> rgeowheels 0.1.0
#>  - Latest cached release: 2025-10-26
#>  <https://github.com/cgohlke/geospatial-wheels/releases/tag/v2025.10.25>

# See all available wheels
assets <- list_rgeowheels_assets()

# Filter for specific package
gdal_wheels <- subset(assets, package == "gdal")
head(gdal_wheels[, c("package", "version", "pyversion", "architecture")])
#>    package version pyversion architecture
#> 61    gdal  3.11.4      3.11        win32
#> 62    gdal  3.11.4      3.11    win_amd64
#> 63    gdal  3.11.4      3.11    win_arm64
#> 64    gdal  3.11.4      3.12        win32
#> 65    gdal  3.11.4      3.12    win_amd64
#> 66    gdal  3.11.4      3.12    win_arm64

Get URL to Wheel File

library(rgeowheels)

# Auto-detect your Python version and return URL
install_wheel("gdal", pyversion = "auto", url_only = TRUE)
#> Auto-selected Python 3.12 for gdal
#> [1] "https://github.com/cgohlke/geospatial-wheels/releases/download/v2025.10.25/gdal-3.11.4-cp312-cp312-win_amd64.whl"

Install with automatic Python version detection

# Auto-detect your Python version and install
install_wheel("gdal", pyversion = "auto")

# Suppress the info message
Sys.setenv(R_RGEOWHEELS_QUIET_AUTO = "TRUE")
install_wheel("rasterio", pyversion = "auto")

# Or specify Python explicitly (suppresses message)
install_wheel("fiona", pyversion = "3.11")

Discover available Python environments

library(rgeowheels)

# Find available Python environments
detect_python_envs()

Install with specific versions

# Latest version of GDAL, latest cpython, available for win_amd64 architecture
install_wheel("GDAL", url_only = TRUE)

# Specific GDAL version for Python 3.10
install_wheel("rasterio", version = "1.3.4", pyversion = "3.10", url_only = TRUE)

# Most recent version of rasterio for Python 3.8
install_wheel("rasterio", pyversion = "3.8", url_only = TRUE)

Use with reticulate

library(reticulate)
library(rgeowheels)

# Create and activate a virtual environment
venv_path <- file.path(getwd(), ".venv")
virtualenv_create(venv_path, python = Sys.which("python"))
use_virtualenv(venv_path)

# Install geospatial packages
install_wheel("GDAL", pyversion = "auto")
install_wheel("rasterio", pyversion = "auto")

# Verify installation
py_run_string("from osgeo import gdal; print(f'GDAL {gdal.__version__}')")