On Friday 6th we held the first Anthologizr Hackathon for the project team. Patrick McSweeney joined us at Senate House to get to discuss the requirements for Antholgizr and best approaches to achieving them. We aimed to make progress on the first critical milestone, namely to get a barebones EPrints export plugin working, into which we can start piling the gory innards of the EPub creation code.
We decided to focus on building the Anthologizr tool on two existing EPrints Bazaar plugins, Shelves and Collections.
Each of these plugins provides a framework for creating arbitrary aggregations of items in a repository, suitable for us to start anthologising. Shelves works by creating a new EPrints dataset which users can add items too; Collections works slightly differently, by creating a new EPrints item, of type ‘collection’, which aggregates other items. We will assess over the course of development if either approach offers significant advantages over the other (particularly with regard to the end-user experience).
Rory and Patrick focused on getting the new export plugin working with Shelves, and, at length, managed to get some XML output. José investigated the Perl EPUB library, with promising results. The end results were encouraging and we briefly had an EPUB exporter running on our test repository. We are very close now to the point where will have a working plugin that will provide an option like the one illustrated below: