Skip to content

SETUP.md

Latest commit

 

History

History
88 lines (61 loc) · 2.28 KB

SETUP.md

File metadata and controls

88 lines (61 loc) · 2.28 KB
Raw

Setup

This repository supports two ways to create a Python environment for the training examples:

  • Pixi, the preferred setup for this repository.
  • Mamba, a compatibility setup based on the existing Conda environment file.

Both approaches assume that you run commands from a terminal.

Pixi

Pixi uses the checked-in pixi.toml and pixi.lock files. This is the recommended setup.

Install Pixi following the instructions at https://pixi.sh/.

Check that the pixi command is the Prefix.dev Pixi tool:

pixi --version

Create the default environment from the repository root:

pixi install

On systems with a small home directory quota, set PIXI_HOME to a location with enough space before installing environments:

export PIXI_HOME=/path/to/scratch/pixi

Pixi and its dependencies may also use the system temporary directory. On HPC systems where the default temporary directory is unsuitable, set TMPDIR to a site-appropriate location:

export TMPDIR=/path/to/scratch/tmp

See the root README.md and the source-code subdirectory READMEs for usage commands.

Mamba

The repository also contains a Conda-style environment.yml file that can be used with Mamba. This environment is broader than the default Pixi environment and includes several optional packages.

Install Mamba or a Mamba-compatible Conda distribution, then create the environment from the repository root:

mamba env create -f environment.yml

Activate the environment:

mamba activate scientific_python

On systems with a small home directory quota, configure Conda/Mamba to store environments and package caches on a filesystem with enough space. For example:

conda config --add envs_dirs /path/to/scratch/conda/envs
conda config --add pkgs_dirs /path/to/scratch/conda/pkgs

Mamba reads the same Conda configuration, so this also affects mamba env create and mamba env update.

If the environment already exists and environment.yml changes, update it with:

mamba env update -f environment.yml --prune

Topic-specific Conda environment files are still present in a few subdirectories for compatibility with older versions of the training material. See the root README.md and the source-code subdirectory READMEs for usage commands.