Digital Collections revamped
This week we have rolled out an update to the St Andrews Digital Collections site, making it easier to discover, navigate to and explore the content of the varied range of items available.
The project to build the Digital Collections began in 2013 when the Digital Humanities Research Librarian liaised with Research Computing staff to create an Islandora repository with two document viewers for displaying digitised books and manuscripts. The first item to be digitised was poet Douglas Dunn’s Red Notebook, from the Special Collections archives, and we were pleased to find that this was much used by students and staff in the School of English. We next worked with the Rare Books librarians to identify books in the collections which were unique or important to St Andrews, or had a particularly interesting provenance, and arranged for these to be digitised. The result is now the 18 items which can be reached from the St Andrews Rare Books tab. Other materials were added over time, again selected on the basis of significance to St Andrews, and include 45 beautifully illuminated congratulatory addresses received by the University on its 500th anniversary in 1911, a selection of letters from historian and anthropologist Andrew Lang, and over 100 18th-century music manuscripts from the library of Gerald Finzi. A project initiated by a School of English PhD student researching the history of photopoetry resulted in an opportunity to digitise over 40 out-of-copyright, attractively illustrated books from the Library’s Photopoetry Books collection, which has added an interesting dimension to the range.
Today the Digital Collections consist of nearly 300 individual items, and are attracting an encouraging number of page views. We look forward to seeing the collections expand further – we are currently in the process of digitising some of our St Andrews Craftsmen’s Guild Books, for example – and are pleased to be able to showcase an increasing number of Special Collections treasures via this platform.
The Digital Collections site is (and always has been) built using Islandora, an open-source framework which combines several robust and well-supported open-source applications to provide a solution for the management and discovery of digital assets.
A Fedora repository is used to store and manage the digital objects (e.g. a book, or a page of a book), their relationships (e.g. page X belongs to book Y) and their datastreams (e.g. a high resolution TIFF image of a page, a lower resolution JPEG version of that image for display on the web, a small thumbnail of the cover of a book or a detailed MODS XML description of a book). Fedora is very robust, well supported and widely used, but it’s not especially user-friendly.
The Drupal content management system is used to create and manage countless dynamic websites, and its extensible architecture allows users to apply custom themes to alter its appearance and use modules to add or change the functionality of their site. The community around Drupal is huge and the platform is very well supported. The community provides and maintains a wide range of modules and themes and developers of Drupal sites can create their own.
The Islandora community has created and maintains a suite of Drupal modules which interact with a Fedora repository to ingest and retrieve/display digital objects. This includes automatically creating derivatives on ingest, such as creating a thumbnail for a book from an image of the first page, and displaying content appropriately according to its content type, e.g. using the Internet Archive BookReader to allow users to “turn the pages” of a book.
Apache Solr (another widely used and well supported open-source component) is used to index the metadata describing the objects in the repository and handles queries of that index. In the current implementation, this facilitates the display of metadata (e.g. shelf location) alongside objects; it can also be used to provide search and browse (e.g. by author or year) functionality – watch out for that in the coming months.
Towards the end of 2015, we updated from an older version of Islandora which worked with Drupal 6 to the current version which works with Drupal 7. While doing so, we overhauled the look of the site so that it is in keeping with the University’s new visual identity, but the structure and functionality of the site remained unchanged. Indeed, we worked hard to maintain the previous level of functionality – changes to the way books are described required the time-consuming updating of most of the content and changes to the implementation of the BookReader meant that additional development was required to make it possible to embed it in other webpages as it had been previously. This was particularly important because what users saw were custom, hand-made Drupal pages describing a book/manuscript into which the BookReader displaying the content of that item would be embedded.
This model didn’t allow us to take full advantage of the capabilities of the Islandora stack and increased the overhead involved in maintaining and, particularly, expanding the collections. We couldn’t use Solr for search, as it would result in users being directed directly to the standard, Islandora-generated pages for the objects rather than the custom ones which had been created. Also, some objects contained in the repository weren’t actually accessible via the Digital Collections site – the Finzi scores were only available via the Special Collections web pages, and the Photopoetry Books were only available via a separate site. Further, users could only explore books using the BookReader, and couldn’t use the functionality offered by Islandora to examine individual pages in detail.
We needed to expose the Islandora-generated pages directly and this involved a number of steps, carried out by the Research Computing team and the Digital Humanities Research Librarian:
- Reorganising the collections into a hierarchy that could be easily navigated.
- Customising the template used to display collections to use the tiles in use elsewhere on the University’s new-look pages
- Creating new thumbnail images for all the books and manuscripts (the automatically-generated ones are too small for the tiles used for navigation).
- Extending the metadata profile used to describe books to include elements for the catalogue URL and the shelf location. Previously, the catalogue URL had only been included in the text in the custom pages, while the shelf location had been included in the title.
- Adjusting the Solr configuration so that the appropriate metadata was indexed for display.
- Updating the metadata records for all the items to make use of the new elements.
- Customising the template used to display books to include appropriate metadata.
- Various other tweaks to the appearance of the site.
- Redirecting the URLs for all the old custom pages to the appropriate Islandora pages.
This work was completed earlier this week and we are very pleased with the results. More content than ever before is accessible via the Digital Collections site, it will be easier to maintain and the workflow for making new content available has been significantly improved. We are now on a sound footing to expand both the content and functionality of the site – watch out for future updates.
Thanks to Alice Crawford, our Digital Humanities Research Librarian, for the paragraphs on the history and content of the Digital Collections.