Installation¶
This section describes how to set up the development environment and install the package from source.
Prerequisites¶
This project uses Mamba and Conda-Lock for dependency management. You’ll need to install these tools first:
1. Installing Mamba¶
Install Mamba globally using the MiniForge distribution:
curl -L -O "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-$(uname)-$(uname -m).sh"
bash Miniforge3-$(uname)-$(uname -m).sh -b
For detailed instructions, see the MiniForge Repository.
2. Installing Conda-Lock¶
Install Conda-Lock globally:
pip install conda-lock
Alternative Setup (Optional)¶
If you prefer not to install these tools globally, you can create a dedicated conda environment:
conda create -n ccc-gpu-dev -c conda-forge mamba conda-lock
conda activate ccc-gpu-dev
Setting Up the Development Environment¶
Create the Conda Environment¶
Create the environment using the pre-generated lock file:
conda-lock install --name ccc-gpu conda-lock.yml --conda mamba
Installing the Package¶
Activate the conda environment and install the package from source:
conda activate ccc-gpu
# (Make sure you are at the root of the repository)
pip install -e .
To uninstall:
pip uninstall ccc-gpu
Updating Dependencies¶
To update the environment when dependencies change:
# Regenerate the lock file
conda-lock --file environment/environment-gpu.yml --conda mamba
# Update the environment
conda-lock install --name ccc-gpu conda-lock.yml --conda mamba
Note
Future versions may transition to using Pixi for dependency management.