Wherever you have input data that needs to be processed to obtain some output, Kerblam! can help you out by dealing with the more tedious and repetitive parts of working with data for you, letting you concentrate on getting things done.
Kerblam! lets you work reproducibly, rigorously and quickly for big and small projects alike: from routine data analysis to large, multi-workflow projects.
You can watch a demo of what Kerblam! can do for you, or you can jump right in by reading the quickstart guide. If you are here often, you might need the installation command.
These are some of the features of Kerblam!:
- 🪄 You can quickly start a new project with
kerblam new. Kerblam generates the folder structure of your project using sensible defaults, similar to whatcookiecutterdoes. - 🚀 Kerblam! can manage your workflows, written for different
workflow management systems. It chiefly prefers GNU
make, but can manage anything that can be executed via the command line. - ♻️ Kerblam! can hotswap input data just before your workflows start, letting you work with different sets of input data on the fly without having to touch your configuration. For example, this is very useful to temporarily work with test data.
- 🎁 Kerblam! lets you run arbitrary workflows into containers, dealing with volume mounting and other details for you. This lets you package workflows that are not natively containerized into Docker for greater reproducibility. For instance, you might just need to run a very tiny data processing pipeline, and don't want to deal with the verbosity of robust workflow management systems like Snakemake, CWL or Nextflow.
- 📦
kerblam packagepackages your pipelines and exports a container image for execution later. This leaves a tarball with all the data a reproducer needs to run the analysis again, so they may do so quickly and easily. The reproducer can use the reply tarball manually, or also let Kerblam! un-package the project for them. - 📦 Kerblam! can fetch remote data and save it locally, manage local data and clean it up, preserving only files that must be preserved. It also shows you how much local data is on the disk, how much data is remote and how much disk space you can free without losing anything important. It can also export the important data to share it with colleagues quickly.
Kerblam! is not and does not want to be:
- 🚱 A workflow runner like
snakemake,nextfloworcwltool;- It supports and helps you execute pipelines written for other WMS, but it does not interfere from then on;
- ♻️ A replacement for any of the tools it leverages (e.g.
git,dockerorpodman,pre-commit, etc...); - 🔎 Something that insulates you from the nuances of writing good, correct
pipelines and container recipies.
Specifically, Kerblam! will never:
- Parse your
.gitignore,.dockerignore, pipes or container recipies to check for errors or potential issues; - Edit code for you (with the exception of a tiny bit of wrapping to allow
kerblam packageto work); - Handle any errors produced by the pipelines or containers.
- Parse your
Kerblam! wants to streamline and standardize data analysis projects as much as possible. For this reason, projects are opinionated:
- The folder structure of your project adheres to the Kerblam! standard,
although you may configure it in
kerblam.toml. Read more about it here. - You use some kind of workflow management system to run your analysis.
- Kerblam! natively uses
make, but nothing stops you writing shell files that execute other tools, likesnakemake.
- Kerblam! natively uses
- You use
dockerorpodmanas your virtualisation service. - You use
gitas your version control system. - You generally execute your pipelines in a container, and not in your development environment.
- Most of your input data is remotely downloadable, especially for large and bulky files.
If you don't like this setup, Kerblam! is probably not for you.
Tip
If you wish to learn more on why these design choices were made, please take a look at the kerblam! philosophy.
The full Kerblam! documentation is online at kerblam.dev 🚀. Please take a look there for more information on what Kerblam! can do.
You might find the quickstart interesting. It provides an overview of what Kerblam! is and does with an hands-on project.
Warning
Release candidates (e.g. 1.0.0-rc.x) are not available on crates.io! Use the pre-compiled
binaries in the releases tab, the command below or compile directly from git with
cargo install --git https://github.com/MrHedmad/kerblam.
In short, use a unix-compatible OS and either:
# Install a prebuilt binary
curl --proto '=https' --tlsv1.2 -LsSf https://github.com/MrHedmad/kerblam/releases/latest/download/kerblam-installer.sh | shor, alternatively
# Install from source with rustc and cargo
cargo install kerblamYou will need git, make and docker or podman installed for Kerblam! to work.
To contribute, please take a look at the contributing guide.
Code is not the only thing that you can contribute. Written a guide? Considered a new feature? Wrote some docstrings? Found a bug? All of these are meaningful and important contributions. For this reason, all contributors are listed in the contributing guide.
If you use Kerblam! or want to add your opinion to the direction it is taking, take a look at the issues labelled with RFC. They are requests for comments where you can say your opinion on new features.
Thank you for taking an interest in Kerblam! Any help is greatly appreciated.
Kerblam! is licensed under the MIT License. If you wish to cite Kerblam!, please provide a link to this repository or use the Zenodo DOI 10.5281/zenodo.10664806
This project is named after the fictitious online shop/delivery company in S11E07 of Doctor Who. Kerblam! might be referred to as Kerblam!, Kerblam or Kerb!am, interchangeably, although Kerblam! is preferred. The Kerblam! logo is written in the Kwark Font by tup wanders.
And remember! If you want it...
