Kliko - The Scientific Compute Container Format
Gijs Molenaar, Spheshile Makhathini, Julien N. Girard, Oleg Smirnov

TL;DR
Kliko introduces a formalized container specification for scientific computing that enhances scriptability and pipeline composition by defining standardized input, output, and task parameters within Docker-based containers.
Contribution
The paper presents Kliko, a new container format with formalized parameters, and demonstrates its implementation through a library and two reference applications in scientific workflows.
Findings
Kliko enables standardized parameter handling in containers.
The Kliko library facilitates workflow chaining and parameter parsing.
Two real-world applications demonstrate Kliko's utility in astronomy.
Abstract
Kliko is a Docker-based container specification for running one or multiple related compute jobs. The key concepts of Kliko are the encapsulation of data processing software into a container and the formalization of the input, output and task parameters. By formalizing the parameters, the software is represented as abstract building blocks with a uniform and consistent interface. The main advantage is enhanced scriptability and empowering pipeline composition. Formalization is realized by bundling a container with a Kliko file, which describes the IO and task parameters. This Kliko container can then be opened and run by a Kliko runner. The Kliko runner will parse the Kliko definition and gather the values for these parameters, for example by requesting user input or retrieving pre-defined values from disk. Parameters can be various primitive types, for example: float, int or the path…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
