New Framework for Search Results Page

Screenshot displaying an example of the first two search results for the search phrase Ruby on Rails

We recently updated the unified All search results page linked from the Duke Libraries homepage. Users may notice the following updates to the interface:

  • Catalog results: Items from the catalog displayed in either the Books & Media or the Archival Materials sections include the following updates:
    • Item availability is displayed using either a green check mark or a red x
    • The call number is displayed for titles with only one item
    • A view online link is displayed for titles that are online
    • Cover images have been moved to the right
  • Chat with a Librarian: Patrons will now click to expand the chat box to use it, and the new modal is always accessible by patrons in contrast to the old version that disappeared completely once dismissed.
  • Removing two unused sections: The Images and Other Resources sections of the results page have been removed. Data for the 2020 calendar year show that links within these two sections each received less than 0.3% of all clicks on the page — these sections were simply not being used.
  • Related searches: Terms related to the current search are displayed near the bottom of the page, allowing patrons to quickly perform a new search (this feature only appears for some searches).
  • More search options: Options to search beyond the initial results using tools such as Research Databases, Online Journal Titles, etc. (displayed at the bottom of the page) have been reformatted to conserve space and to include an icon indicating that each link leads to a different website.

Changes to underlying technology

Most of the work to implement this new version is behind the scenes and focuses on the following changes.

New framework

Our old version of the unified results page was built using the Drupal 7 content management system that supports our main website; however, Drupal 7 will soon be replaced with a newer version of Drupal. Rather than migrate our unified search results page to the newer version of Drupal, we opted to migrate to an application called Quicksearch that was developed by our colleagues at North Carolina State University (NC State) and is built with the Ruby on Rails framework.

Since many of our discovery tools are Rails-based, this is a framework that is familiar to our developers, and using NC State’s Quicksearch as our starting point also saved time.

Because the new unified search results page is now a separate application from Drupal, it has a new URL, quicksearch.library.duke.edu, but the primary starting point for accessing it will continue to be the All tab on the library homepage.

Website search

The website search section of our unified search results page previously used a deprecated version of Google Custom Search Engine that was not accessible from Duke Kunshan University. We have switched to a website search based on two open source tools, Nutch (a web crawler) and Solr (a search platform). Using Nutch and Solr for our Website search will allow us to continue displaying ad-free results, will cost less to maintain over time, will be usable by patrons at Duke Kunshan University, and will help us maintain patron privacy.

Purpose

The unified search results page provides users with quick access to content across several of the discovery platforms provided by Duke University Libraries, allowing users to see wide ranging results when starting from the All tab on the library homepage. Users with more granular research needs are invited to explore more focused research paths listed on our Search & Find portal page.

Search box from library homepage with the All tab highlighted

Teamwork

Moving our unified search results page to a new framework is the result of a collaborative project undertaken by IT staff within the library. The Digital Strategies and Technologies Scrum Team completed the implementation; special thanks goes to Cory Lown, Derrek Croney, Michael Daul, Sean Aery, and Zeke Graves.