We’ve seen that Snippy-NG will be by the people i.e. it is written in the lingua franca of the programming world: Python. In this post, we’ll explore how Snippy-NG will be for the people i.e. easy to install.

Installing Snippy

The original Snippy is a Perl script that required a number of system-level dependencies. While Conda helps to some extent, installation issues remain one of the most common problems reported on GitHub. This has led to Snippy being criticized for its degraded installation experience.

We plan to address these issues in Snippy-NG by adopting modern software development practices. Most notably, we will provide a pre-built environment that installs Snippy-NG and all its dependencies with a single command. This environment uses pinned versions of all dependencies to ensure Snippy-NG behaves consistently across systems. In addition, we use continuous integration and GitHub best practices to keep the environment up to date and Snippy-NG always functional.

You can install snippy-ng via the install script using the following command:

curl -sSL https://github.com/centre-pathogen-genomics/snippy-ng/releases/latest/download/install.sh | bash

This version of Snippy-NG includes all dependencies pinned via the pixi.lock file. Users installing Snippy-NG through this script will get the same versions of all dependencies, ensuring reproducibility.

Pre-built Environment

We are using Pixi and pixi-pack to create cross-platform, self-extracting binaries of the complete Snippy-NG environment. We’ve developed GitHub Actions that automatically pack Pixi environments and generate an install script. This approach not only solves installation issues but also enhances reproducibility.

GitHub Best Practices

By adopting automated semantic releases, we ensure Snippy-NG is always up to date and that issues are resolved quickly. We also use GitHub Actions to run tests and benchmarks on every pull request and release to guarantee performance and correctness.

We’re committed to achieving 100% code coverage and have checks in place to prevent any regression. To release a new version of Snippy-NG, code coverage must not decrease. Additionally, we have blocked direct commits to the main branch—ensuring that all changes go through pull requests, where they are reviewed and tested before being merged.

Conclusion

Snippy-NG will be for the people by offering a streamlined installation experience and following best practices in software development. We’re excited to see how this will improve the user experience and make Snippy-NG a more reliable tool for microbial genomics research. We also look forward to community contributions that will help us make Snippy-NG even better.