Installation¶
This guide walks you through setting up GIANT on your local machine.
Prerequisites¶
- Python 3.11+ (tested with 3.11, 3.12)
- OpenSlide library (for reading whole-slide images)
- uv package manager (recommended) or pip
System Dependencies¶
macOS¶
# Install OpenSlide
brew install openslide
# Install uv (recommended package manager)
curl -LsSf https://astral.sh/uv/install.sh | sh
Ubuntu/Debian¶
# Install OpenSlide
sudo apt-get install openslide-tools python3-openslide
# Install uv
curl -LsSf https://astral.sh/uv/install.sh | sh
Windows¶
# OpenSlide Windows binaries: https://openslide.org/download/
# Add OpenSlide bin directory to PATH
# Install uv
powershell -c "irm https://astral.sh/uv/install.ps1 | iex"
Clone and Install¶
# Clone the repository
git clone https://github.com/The-Obstacle-Is-The-Way/gigapixel-goblin.git
cd gigapixel-goblin
# Install dependencies with uv (recommended)
uv sync
# Activate the virtual environment
source .venv/bin/activate
Verify Installation¶
# Check GIANT version
giant version
# Verify OpenSlide is working
python -c "import openslide; print(f'OpenSlide version: {openslide.__version__}')"
# Run unit tests
uv run pytest tests/unit -x
API Keys¶
GIANT requires API keys for LLM providers. Create a .env file in the project root:
See Configuring Providers for detailed setup instructions.
Next Steps¶
- Quickstart - Run your first inference
- Data Acquisition - Download WSI files for benchmarks
- Configuring Providers - Set up OpenAI/Anthropic