Installation¶

Basic Installation¶

Install from PyPI:

pip install cng-datasets

Or install from source:

git clone https://github.com/boettiger-lab/datasets.git
cd datasets
pip install -e .

Development Installation¶

To install with development tools:

pip install -e ".[dev]"

This includes:

pytest for testing
black for code formatting
ruff for linting
mypy for type checking

Raster Processing Support¶

For raster processing, GDAL requires system libraries. Install GDAL first:

Ubuntu/Debian¶

sudo apt-get install gdal-bin libgdal-dev python3-gdal
pip install -e ".[raster]"

macOS¶

brew install gdal
pip install -e ".[raster]"

Using Docker (Recommended)¶

The easiest way to use this package with full GDAL support is via Docker:

# Pull the pre-built image
docker pull ghcr.io/boettiger-lab/datasets:latest

# Run interactively
docker run -it --rm -v $(pwd):/data ghcr.io/boettiger-lab/datasets:latest bash

# Or run a specific command
docker run --rm -v $(pwd):/data ghcr.io/boettiger-lab/datasets:latest \
  cng-datasets raster --input /data/input.tif --output-cog /data/output.tif

The Docker image includes:

GDAL with full NumPy array support
All Python dependencies
AWS CLI and rclone for cloud storage
Pre-installed cng-datasets package

Verifying Installation¶

Check that the package is installed correctly:

cng-datasets --help

You should see the command-line interface help message.