This was the last working week before Christmas, and it was a four-day week due to Friday being given in lieu of Christmas Eve, which is on Sunday this year. I’ll be off for the next two weeks, which I’m really looking forward to.
The Books and Borrowing project officially comes to an end on the 31st of December, although we’re not going to launch the front-end until sometime in January. I still had rather a lot of things to do for the project and therefore spent the entirely of my four days this week working for this project. Of course, the other team members were also frantically trying to get things finished off, which often led to them spotting something they needed me to sort out, so I found myself even busier than I was expecting. However, by Thursday I had managed to complete all of the tasks I’d hoped to finish, plus many more that were sent my way as the week progressed.
At the end of last week I’d begun updating the site text that the project PI Katie had supplied me with, and I completed this task, finally banishing all of the placeholder text. This also involved much discussion about the genre visualisations and what they actually represent, which we thankfully reached agreement on. I also added in some further images of a register at Leighton library and wrote a script to batch update changes to the publication places, dates of publication and formats of many book edition records. One of the researchers also spotted that the ‘next’ and ‘previous’ links for the two Selkirk registers were not working, due to an earlier amalgamation of page records into ‘double spread’ records. I therefore wrote another script to sort these out.
I then added new ‘Top ten book work’ lists to the site-wide ‘Facts’ page (overall and by the gender of borrowers). This required me to update the script that generated the cache that I developed last week, to rerun the script to generate fresh data, to update the API to ensure that works were incorporated into the output and to update the front-end to add in the data. Hopefully the information will be of interest to people.
I then overhauled the highlighting of search terms in the search results. This was previously only working with the quick search, and only when no wildcards were used in the search. Instead I used a nice JavaScript library called mark.js (https://markjs.io/) that I’d previously used for the DSL website to add in highlighting on the client-side. Now any the values in any search fields that are searched for will be highlighted in the record, including when wildcards are used. I also updated the highlight style to make it a bit less harsh.
It should be noted that highlighting is still a bit of a blunt tool – any search terms will be highlighted throughout the entire record where the term is found. So if you search for the occupation ‘farmer’ then wherever ‘farmer’ is found in the record it will be highlighted, not just in the normalised occupation list. Similarly, if you search for ‘born’ then the ‘born’ text in the author information will be highlighted. It’s not feasible to make the highlighting more nuanced in the time we have left, but despite this I think that on the whole the highlighting is useful.
I reckoned that the highlighting could end up being a bit distracting so I added in an option to turn results highlighting on or off. I added a button to process this to the search results page, as part of the buttons that include the ‘Cite’ and ‘Download’ options. The user’s choice is remembered by the site, so if you turn highlighting off and then navigate through the pages of results or perform a filter the highlights stay off. They will stay off until you turn them on again, even if you return to the site after closing your browser.
One of the researchers noticed that an unnecessary near-duplicate genre had somehow been introduced into the system (‘Fine Art’ instead of ‘Fine Arts’) so I removed the and reassigned any records that were assigned to the erroneous version. The PI Katie also spotted some odd behaviour with the search form boxes. When using the browser’s ‘back’ button search data was being added to the wrong search boxes. This took quite some time to investigate and I couldn’t replicate the issue in Firefox (the browser I use by default), but when using a Chrome-based browser (MS Edge) I experience the issue. It turns out it’s nothing to do with my code but a bug in Chrome (see https://github.com/vuejs/vue/issues/11165). The fix mentioned on this page was to add ‘autocomplete=”off”’ to the form and this seems to have sorted the problem. It’s crazy that this issue with Chrome hasn’t been fixed as the posts on the page identifying the issue started in 2020.
Katie also spotted another issue when using Chrome. Applying multiple filters to the search results wasn’t working in Chrome, even though it worked fine in Firefox. This time it was caused by Chrome encoding the bar character to %7C while Firefox keeps it as ‘|’. My filter script was splitting up filters on the actual bar character and as this wasn’t present in Chrome multiple filters were not working (even though they were working fine in Firefox). Thankfully once identified this was relatively easy to fix.
I also managed to implement a ‘compact’ view of borrowing records this week, something that had been on my ‘to do’ list for a while. Borrowing records can be extremely verbose and rather overwhelming so we decided to give the option to view compact versions of the records that contain a narrower set of fields. I added a compact / full record view switcher to the bar of options in the top right of the search results and library register page, beside the ‘Cite’ option. As with the highlighting feature I previously discussed, the choice is remembered in your browser, even if you return to the site in a later session (so long as you’re using the same device and browser, of course).
For the compact view I decided to retain the links to the library, register and page as I figured it would be useful to be able to see these. Also included are the borrowed and returned dates, the borrowers (names only), the title of the Book Work (or Works) if the record has such an association and the title of the Holding if not, any associated authors and genres, plus a list of the volumes borrowed (if applicable). The following screenshot shows what the compact looks like:
My final tasks of the week were to add in a cookie banner for the site and install Google Analytics. In the New Year I’ll need to regenerate the Solr index and then integrate the development site with the live site. This will include making updates to paths throughout the code, ensuring the existing Chambers Maps continues to function, adding links to the pages of the development site to the site menu and adding the quick search option to the site header. It will be great once the site is fully accessible.
Also this week I created a new Google Analytics property for a site the DSL launched a month or two ago and spoke to Geert, the editor of the AND about an issue he’d spotted with entry dates (that I’ll investigate after Christmas). I finished off my work for the year by removing the Twitter widget from every site I’m responsible for. Twitter blocked access to the widget that allows a feed to be embedded in a website a few months ago and it looks like this is a permanent change. It meant instead of a nice Twitter feed an empty feed and a ‘nothing to see here’ message was displayed on all of my sites, which was obviously no good. It feels quite liberating to drop Twitter (or X as it is currently called).
That’s all from me for this year. If anyone is reading this I wish you all the best for Christmas and 2024!