2016 · SIGMOD

ReproZip: Computational Reproducibility With Ease

ReproZip: Computational Reproducibility With Ease

Cite this work

Video Presentation

2016SIGMODreproducibility

Authors

Fernando Chirigati, Remi Rampin, Dennis Shasha, Juliana Freire

Abstract

We present ReproZip, a tool that automatically captures provenance of experiments and packs all the necessary files, libraries, and variables to reproduce the results. ReproZip uses system call tracing to identify dependencies and creates a self-contained package that can be run on different machines without requiring users to install any dependencies. The tool supports multiple unpackers including Docker, Vagrant, and native execution.

ReproZip enables computational reproducibility with ease by automatically capturing all dependencies needed to reproduce an experiment.

## Features

- **Automatic Dependency Tracking**: Uses system call tracing to capture all files and libraries
- **Cross-Platform Support**: Packages can be reproduced on different operating systems
- **Multiple Unpackers**: Support for Docker, Vagrant, and native execution
- **Easy Sharing**: Single file packages that can be easily shared with collaborators