The borrowed page

I see that the Archive.org now has a ‘no-Borrow preview page’ on search-landing pages for its ‘Borrow’ books. They’ve also announced that their extended “Borrow” feature is to come to an end, as was always to be the case after the emergency period had passed.

One wonders if this new ‘page preview’, similar in nature to Google Books, is about bringing Google in on the new lawsuit from publishers? ‘If this goes down, so does Google Books?’ Just my guess.

On auto-downloading open access books

Martin Paul Eve has a new post on Zotero and auto-downloading open access books…

all I really wanted was to be able to embed an ISBN and a citation_pdf_url and have Zotero do the lookup and save the file. However, out of the box there is no easy way to do this.

His test book is quite interesting, his own new Close Reading with Computers: Textual Scholarship, Computational Formalism, and David Mitchell’s Cloud Atlas (April 2020), which applies textual computing to the science-fiction-philosophy novel Cloud Atlas.

I don’t know about or use the current version of Zotero, so I’m unsure what advantages it confers. I assume Eve intended to find a way to automatically harvest all CC-SA books in PDF, and build a local collection for automated analysis.

But I see his book is already on the OA book aggregator catalogue OAPEN. Theoretically then, since OAPEN is comprehensive and timely, one could have a harvester look at all the pages hanging off library.oapen.org/handle/ and save out only those pages with the required permissive CC “Rights” label on them. These pages each have a uniform PDF link URL in their HTML, in the form of library.oapen.org/bitstream/ and these could be easily extracted to a list. One would end up with a set of PDF links for a linkbot, ready to download to a local folder for computational analysis. I presume that’s what Eve intended to have Zotero do.

One would need to reference the OAPEN record page first, in the way I’ve suggested, since the PDF itself can have different or non-uniform or contradictory licence information. For instance in its interior Eve’s book is labelled as both “©” … “No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying and recording, or in any information storage or retrieval system without the prior written permission of Stanford University Press.” and also “Creative Commons Attribution-ShareAlike 4.0”.

How many items on OAPEN have a creativecommons.org/licenses/by-sa/ “Rights” label at present, as Martin’s book does? A Google site: search suggests around 650 titles. Half an hour of my filtering the OAPEN CSV suggests it’s actually just over 3,000 under some form of permissive CC that permits commercial use. That’s still a manageable harvest at present. But as the supply of OA books and monographs grows rapidly, the likely result of various OA mandates in the near-future, it might be a useful time-saver for text-miners and digital humanists if OAPEN were to maintain a single torrent of all the PDFs. Inside which a half dozen folders would neatly organise the books by CC licence type. Such a one-click solution might save a lot of faffing around with digging into and filtering their XML and CSV feeds, wrangling with harvester scripts and timeouts, or trying to wrestle with third-party services such as Zotero. A torrent could also save OAPEN’s bandwidth.

Dead Carrot

The Carrot2 search-engine appears to be kaput. It’s been down for days now, unable to show results. Multiple different browers, on different systems, even different VPNs, give no results. A query can be input, but… no results. Even a TOR browser can’t get it to respond. Such a pity, it was so promising.

The worthless Peter Lowe ad-blocking list

The Peter Lowe ad-blocking list is obviously now worthless, due to its over-reach and scattergun blocking of all sorts of legitimate things. Back in June I found it blocking Harvard. I’ve since found all sorts of similar blocks on things that should not be blocked. I’ve unsubscribed my browser’s ad-blocker addon from the Peter Lowe list, and I suggest that you also consider doing so.

Carrot2 search – a new script for multi-column results

The Carrot2 search-engine has changed both URL and layout. It was at search.carrot2.org/stable/search and it’s now just at the search.carrot2.org URL. I guess the public-facing search may have come out of beta? This is what their new layout looks like…

Not good, on a widescreen desktop monitor.

Which means I’ve made another custom CSS for it. This slices the Carrot into a multi-column layout suited to a widescreen 1920px monitor. It works in the Stylus browser addon, and you need to tell the script to target the search.carrot2.org site

Yum, crunchy Carrot! All the z-depths are set up nicely, so you can still click on/in the filters and search box.

I like to read the URLs in search-results and so I’ve turned them dark green and wrapped them to make this possible. The results look good with URLs that line-wrap by up to three lines. But it’s unavoidable that some very long URLs will wrap over four lines, and will thus spill over the element below.

In most cases there should be no need for any scrolling.

It only works in Day mode, with the Carrot’s new toggle-able Night and Day mode…

If you want a night mode, you’ll have to manually change the colour chips to a charcoal black on…

body, body.light {

and

div.ResultList

The other drawback is that whereas the old multi-column fix showed about 25-30 results, now we’re down to about 16. If you want it up to 20+ add the following code…


.ResultClusters {
display: none;
}


There’s also the lack of a “more…” button, but this usefully forces the user to use the Carrot’s innovative faceting systems over in the left-hand pane.

---

To install my fix, simply go to Carrot2, then left-click on your icon for the Stylus browser addon and click “Write style for…”.

Then paste in this…


/* ==== CARROT2 - Multi-Columns v.02 Oct 2019 ==== */
/* run this Stylus script on search.carrot2.org */
div.document div.url {
overflow: hidden;
color: #3a7730;
font-size: 110%;
}
body, body.light {
background-color: #ece5db;
}
.ResultList > div > a > span.url {
color: #3f7126;
font-size: 80%;
font-weight: bold;
white-space: pre; /* CSS 2.0 */
white-space: pre-wrap; /* CSS 2.1 */
white-space: pre-line; /* CSS 3.0 */
white-space: -pre-wrap; /* Opera 4-6 */
white-space: -o-pre-wrap; /* Opera 7 */
white-space: -moz-pre-wrap; /* Mozilla */
white-space: -hp-pre-wrap; /* HP Printers */
word-wrap: break-word; /* IE 5+ */
}
a {
font-size: 100%;
}
.SearchForm {
z-index:20;
}
div.ResultList {
z-index:4;
position:absolute;
background-color: #ece5db;
padding-top: 180px;
top:0;
bottom:0;
right:0;
column-count: 5;
width: 70%
}
.ResultList > div > a >div {
font-size: 80%;
}
div.sources {
padding-top: 25px;
}
div.ClusterList {
background-color: #ece5db;
padding-top: 10px;
column-count: 2;
}
div.clusters-tabs {
width: 40%
}
div.clusters {
z-index:4;
width: 50%
}


Tested in Opera, which is a browser that runs on Chrome. It will probably work with other CSS style injectors.

Be warned that Carrot2 will perma-block an entire IP address (in the case of BT in the UK, or VPNs, that can mean hundreds of thousands of users) if it detects “excessive traffic”.

WordPress to ebook – without plugins

Needed: an easy three-click “WordPress to ebook” service that works on any free WordPress.com blog account. Such a thing doesn’t currently seem to exist. I can only find ebook plugins for install on third-party hosted installs of WordPress, or which need WordPress.com’s £240 per year Business account upgrade (which allows use of plugins).

Zinepal used to be a service that would do it from an RSS feed, but has been withdrawn. The similar ebookGlue is dead. I would have thought this would be an in-demand service for the millions of free WordPress blogs. But no, it seems not. Ideally such a service would allow you to mark only your wanted posts by previewing a snippet as well as the title, then let you re-shuffle them into themed chapters and sections, before outputting to an editable ebook-friendly format. I’m assuming a non-fiction blog here, with perhaps 3,000 posts.

What it doesn’t need to be is something just pushes the whole blog into an ePub and calls it done.

Update: Solved. WordPress2Doc – a free ebook converter for free WordPress.com blogs.

Button-holed

The Google CSE team is obviously updating with some tweaked code, on the front-end of the Custom Search Engines (CSEs).

As a result, JURN’s custom “click to search button” has become a small generic-looking one done in .SVG code. This is smaller and tighter than I’d like, and has replaced JURN’s nice custom .JPG button.

This “click to search” button still works fine, so it’s only an aesthetic change…

I’ll leave it a few days before jiggling the CSS to try to get the custom button back, just in case Google decides to revert the changes.

GRAFT updates

GRAFT has updated with a new batch of repository URLs added. Search across full-text and records alike, in 4,723 repositories.

JURN pagination links fixed

In the last week or so Google has made some slight changes to the default styling templates for CSEs, resulting in the numbered pagination links at the foot of the search results becoming very small and grey. This has now been fixed on JURN, and your per-page links to more search results should now look like this. They should be far more easily selectable now, and especially for touch-screen users…

My thanks to Amit Agarwal of India, for the elegant snippet of commented CSS for the .gsc-cursor-page element. If you have the same problem with your own CSE, this snippet goes in the style header of your page. Colours are controlled elsewhere, in the ‘Look & Feel’ | Customise | Refinement section of your CSE admin dashboard.

Changes may not show up until you and your users refresh your main page a few times, due to Web browser caching.

GRAFT has also had the same fix applied.

---

Update:

Also add padding for the pagination row, by adding the following to your CSS style (I have mine in the page itself)…

Search serendipity

An example of the strange ways in which search serendipity can (and often does) work…

Ahead of the release of Zbrush 2019, I was searching for “npr” 3d shader in: Google Search | ‘Last Month’.

On the second page of results, I discover the ‘moved and lost from JURN a year ago’ archive of the RAF’s Air Power journals.

A search for “npr” 3d shader has nothing to do with the British Royal Air Force. ‘NPR’ being non-photorealistic rendering with ‘3D’ computer models that have been fitted with ‘shader’ materials, to make them look like hand-drawn cartoons when they’re rendered into graphical form. The results were arising because the same keywords were shared.

Air Power et al will be back in JURN soon.