Has anyone run into problems with PDFs in Islandora not rendering on Apple products?

[Annika Peterson]

Hello all,

I have a very specific question: has anyone encountered trouble with PDFs rendering incorrectly on Apple products? My institution has been using Islandora for more than a year now, and we've uploaded many collections with PDFs without a hitch. However, we have one collection where the PDFs refuse to render properly on iPhones/Macs. Only the OCR layer of the PDF loads. The image layer will not appear. On PCs, the OCR layer renders first, but the image layer appears after a second or two. (Here is an example of an item in the collection: https://uplink.nmu.edu/islandora/object/nmu%3A87163)

This set of PDFs came to us already digitized, so I have no idea how the original items were scanned or what software was used. Our site uses the PDF.js viewer, and we're still using Islandora 7. Our site is hosted by Lyrasis. I've talked to our contact there, but he didn't know what to do. I'm hoping someone in the larger community might have run into this problem before.

Thanks!

Annika Peterson

UPLINK Digital Project Manager

Central Upper Peninsula and Northern Michigan University Archives

[Rosie Le Faive]

Hi Annika. I hope I can help or at least rule out some things! 

Issues of things loading/not loading usually points to a problem in the browser. Does the mac have any browsers that do render it correctly? (Firefox, safari, chrome, etc). I tried those three, but can't repeat your issue on my mac (Ventura 13.1). The pdf appears in the viewer as a black & white scanned newspaper.

If it's still not loading for you, can the mac/iphone load the PDF at https://uplink.nmu.edu/islandora/object/nmu%3A87163/datastream/OBJ/view ? 

If it's just one browser that's acting weird, open up the developer tools (easier on a computer) and see if the Console or Network tabs indicate what's going on.

When I was testing just now, I was sure that Chrome "just wouldn't load the PDF at all for me". But it turns out I keep that browser with Javascript off by default, and when I enabled javascript it worked fine. 

I'm curious to hear what you figure out!

[Willow Gillingham]

Hi Annika,

I would propose a couple troubleshoothing steps to help figure out what might be happening.

 - are the images actually external sources -- and that source is not reachable from the current location?

 - when it fails to display in Chrome, are there any console messages related to it not showing?

  - It is possible that the PDF has all of the content on the same page (and even though it looks like ONLY the OCR is being shown, the images could all be on the same page but behind the OCR layer). Could the PDF just be very old and needs to be edited with a current version of PDF?

As for the performance of big PDF files on any browser, there is a fairly large payload - both binary and HTML to support the spans in the viewer. I am sure that there is a little bit of CPU overhead when the page loads but I doubt that it is the entire source of the slow loads on a PC. This can be tested if the page is cached and compare that time to the uncached time it takes to render.

I hope this helps or something that Rosie mentioned above.

If you do figure this out, please let us know -- we'd look forward to hearing what causes this too.

Willow Gillingham

[Annika Peterson]

Rosie and Willow,

Thank you for your advice (and sorry for my delay in replying)! I couldn't figure out exactly what was causing the problem (if you're interested--it was happening in all browsers on Apple products, but viewing items via /datastream/OBJ/view worked). However, I did find a work-around by completing re-doing the PDFs (exporting the pages as images, creating new PDFs, and re-doing the OCR). I suspect the issue probably was age--the original files were very glitchy whenever I tried to open them in Acrobat.

Thanks again!

-Annika