ReproZip Examples

This repository holds some examples and use cases from different domains that showcase the ReproZip software packing tool.

You'll find more information about each particular example in its associated README.

Examples

Creating a provenance graph

• bash-count: A simple bash script that counts the number of pages of an input file.

Reproducing published results

• bechdel-test: A data analysis experiment that tries to reproduce the claims of an article from FiveThirtyEight.
• data-polygamy: A paper published at SIGMOD 2016 whose plots were made reproducible using ReproZip.
• ising-model: A typical simulation in statistical physics, consisting of large scale Monte Carlo simulations followed by an involved statistical analysis of the results.
• irish-schools: An example from digital humanities, where publication ready graphs and materials are generated from tabular data using R.
• mongodb-vls: A paper published at ICDE 2016 whose plots were made reproducible using ReproZip.

Extending the original experiment in VisTrails

• digits-sklearn: A supervised learning and classification experiment written in Python, using scikit-learn.

Packing interactive and GUI applications

• bus-vis: A Java application for visualizing the Konstanz bus transportation system.
• object-tracking: A Python application to do multiple object tracking (MOT) in an input video.

Reproducing the experiment with a different input

• brain-segmentation: A Python script that extracts brain information and mask from an input b0 image using dipy.
• digits-sklearn-opencv: A similar experiment to digits-sklearn, but using the OpenCV library to extract the digits from a photograph before predicting their values.

Archiving and porting web applications

• django-blog: A simple website built with Django that displays a list of blog posts and allows the admin to edit or post. Despite its simplicity, it represents a realistic small web application using a database (SQLite3).
• stacked-up: A website called Stacked Up, also built with Django, to explore the textbook inventory of Philadelphia public schools. All the data is stored in a PostgreSQL database.
• repromatch: A website designed to help you find the tool that best matches your reproducibility needs. Built with Django and Apache Solr for search.

Instructions for reproducing the examples use the vagrant and the docker unpackers. However, any of the available unpackers can be used.

Vagrant Machine

This repository contains a Vagrantfile that automatically configures a machine with the following examples:

• bechdel-test
• brain-segmentation
• bus-vis
• digits-sklearn
• digits-sklearn-opencv
• irish-schools
• stacked-up

To startup the machine, make sure you have Vagrant and VirtualBox installed, and then run the following:

$ vagrant up --provider virtualbox

Note that this will take some time for the first time, since all the dependencies will need to be properly installed. By default, the machine starts in headless mode (no UI visible on the host machine). To boot it with a GUI, uncomment the v.gui = true line in the Vagrantfile.

Instructions on how to run each of the examples in this machine are available in their respective pages.

Other Useful Links

• Main Website
• Documentation
• GitHub Repository
• Packing and Unpacking with ReproZip: A YouTube video showing how to pack and unpack an experiment using ReproZip.

Team

ReproZip is currently being developed at NYU. The team includes:

• Fernando Chirigati
• Juliana Freire
• Rémi Rampin
• Dennis Shasha
• Vicky Steeves