Conference Agenda

Session Overview
Developer track 1: Integrations and APIs
Wednesday, 28/Jun/2017:
11:00am - 12:30pm

Session Chair: Sam Searle, Griffith University
Location: Redlands/Lockyer/Moreton
Hilton Brisbane


(Not) Lost in translation

Elizabeth Krznarich

ORCID, United States of America

Systems that serve a global audience face the unique challenge of ensuring access to users who speak a wide range of languages. Translating application interfaces into many languages (and maintaining them over time) is not only technically challenging; it can also be time-consuming, expensive and downright maddening. For those in the open-source community, supporters and stakeholders willing to contribute their local language expertise can ease the burden, however, distributing work across many translators with varying levels of technical skill brings challenges of its own. To tame the chaos, ORCID recently implemented Transifex, a translation management tool that allows users with limited technical skills to easily create and edit translations, as well as a customized Ruby application, TXGH, which allows automated syncing of translations between Transifex and Github. This presentation will provide an overview of both tools and a demo of the Transifex/Github syncing process.

Collaborative ingenuity @ UNE: Engineering an in-house solution with ORCiD and APIs

Andrew Devenish-Meares, Melissa Abbott

University of New England, Australia

The University of New England recently resolved a number of repository issues with an in-house solution developed in collaboration between the Information Technology Directorate, the University Library and Research Services. The Chute application uses RESTful APIs to integrate external databases, internal systems and ORCiD to the repository. UNE would like to share how the new system integration has allowed the Library to provide new services to researchers, has enabled an accurate and timely ingest of data, and has empowered us to analyse our institution’s publication data in new ways. Our experience with Chute provides potentially useful lessons to other libraries.

End to end reproducibility using CSIRO’s Data Access Portal API and the Workspace scientific workflow software

Dominic Laurence Hogan1, Dale Matthews2, Christopher Franks1, Matthew Birks1, Matt Bolger1, David Benn1

1CSIRO, Australia; 2Australian National University

An increasingly frequent request from researchers in the Commonwealth Scientific and Industrial Research Organisation (CSIRO – Australia's national science agency) has been for the ability to programmatically create collections in the Data Access Portal (DAP). Since 2012 the DAP has supported the publication and archiving of research data and software through a self-service web interface. One use of a programmatic method to create DAP collections would be to easily deposit data about samples as they are collected. In addition to providing flexibility for specific tasks, such infrastructure is a useful component when attempting to create fully reproducible research outputs. This can be of particular importance when the purpose of the research is to inform government policy, where decisions may need to be defended from legal challenges. Allowing external parties to access all data and software that was used to produce the results allows all aspects of the work to be scrutinised. We created a plugin for CSIRO's Workspace software that would simplify use of the DAP API for researchers. This allows projects to have source data, workflow software, workflows, and data outputs all made accessible via one repository, requiring minimal setup by the end user to re-run all processes.

Snek: a lightweight Fedora frontend

Mike Lynch, Peter Sefton

University of Technology Sydney

We demonstrate Snek, a lightweight Django frontend to Fedora Commons 4, and pcdm-lite, a command-line tool for quickly building a structured repository from a spreadsheet.

Snek provides a user-friendly web interface to a FC4 repo with a minimum of code, templates, and http requests per page.

Its features include:

* Where possible, all data retrieval, and transactions are done with the Solr index, avoiding calls to Fedora as much as possible.

* Simple mapping between RDF types and Python classes representing resources

* Standard RDF links (like PCDM) are mapped to web navigation elements in a sensible way by default, but it's easy to override this with templates

pcdm-lite is a simple and flexible tool for populating a repository with linked data based on a spreadsheet. Both pcdm-lite and Snek use fcrepo4py, a simple Python interface to the FC4 REST API.