Overview

MyTardis is an open source Data Repository product developed by Monash University for managing xray diffraction data. It has subsequently been used for electron microscopy data.  MyTardis ingests data files and metadata, and allows users to view files, download them, share them with colleagues and "publish" them.  The default MyTardis UI has a modern look with thumbnails and carousels, and is (for the most part) intuitive.

The repository codebase is implemented using Python, Django and South.  It can use either MySQL or PostgreSQL as its backend database.  The Django framework allow you to add standard or custom "apps" to the base MyTardis.  These can be used to enhance or skin the user interface, add new ingestion mechanisms and metadata extractors, and so on.  The latest version of MyTardis includes a RESTful API.

The MyTardis primary data model has 3 levels:

  • Experiment is the top level entity
  • DataSets belong to one or more Experiments; i.e. not strict containment.
  • DataFiles are strictly contained by a single DataSet.

Metadata is represented as name/value pairs defined by schemas.  Metadata can be attached at each level of the model.

MyTardis has a flexible approach to the storage of the data associated with a DataFile.  It can store the data directly, or simply manage references to the location in which the data is stored.  In addition, it can be configured to store replicas of data in multiple Locations.  (There was also a project to support offline storage of Experiments as archives, but this was not completed.)

Installation

There is a Chef cookbook for installing MyTardis in the NeCTAR Cookbooks collection.

Health report

There have been a number of MyTardis sites in production but not all are still alive.  (For example, the UQ Mirage site is "off".) My best guess is that there are at least 8 "live" installations at the moment, at Monash, Univ of Sydney, Swinburne, RMIT, Australian Synchrotron and ANSTO.

There are 2 core developers at Monash, and the core team has always been small, which has tended to slow progress.  Over the life of the project, there have been contributions from contributions from 20+ developers.

The primary code repositories and issue trackers are fully public, and there is a Google Group for developer discussion.  The code has a BSD license, and copyright is not assigned.