.. _getting_started-installation: ============ Installation ============ Paks can be installed from pypi, or from source. We recommend a container interaction for the quickest way to get up and running via our `container bases `_ Note that `spack `_ is required to be on your path. If you use the container, this will already be provided. .. code:: console $ git clone --depth 1 https://github.com/spack/spack $ source spack/share/spack/setup-env.sh Container ========= It's recommended to interact with paks via `one of the container bases `_ mentioned previously. That might look like the following: .. code:: console $ docker run -it ghcr.io/syspack/paks-ubuntu-18.04 # which paks /usr/local/bin/paks # which spack /opt/spack/bin/spack # which oras /usr/local/bin/oras You'll notice that the container, along with paks, also provides spack and `oras `_ which is used to interact with the GitHub packages registry. Pypi ==== If you want to use Paks locally on your machine or via some custom install, the module is available in pypi as `the paks project `_. .. code:: console $ pip install paks This will provide the latest release. If you want a branch or development version, you can install from GitHub, shown next. Virtual Environment =================== Here is how to clone the repository and do a local install. .. code:: console $ git clone https://github.com/syspack/paks $ cd paks Create a virtual environment (recommended) .. code:: console $ python -m venv env $ source env/bin/activate And then install (this is development mode, remove the -e to not use it) .. code:: console $ pip install -e . Installation of paks adds an executable, ``paks`` to your path. .. code:: console $ which paks /opt/conda/bin/paks Once it's installed, you should be able to inspect the client! .. code-block:: console $ paks --help You'll next want to install or build packages, discussed in :ref:`getting-started`.