IIIF and Axiell EMu

[Brian McKenna]

Hi,

I'm new to IIIF but I'm looking to develop an IIIF server solution at my institution, the University of Glasgow, probably in Java.

Our collection images are stored in EMu. Here are my initial thoughts on solution design:

-Deploy Cantaloupe to our Java server container

-Export hi res images from Emu to Cantaloupe

-Create bespoke Java web app to implement IIIF presentation API based on EMu data, pointing images to Cantaloupe image server deployment

-Create a web page that uses Mirador, pointing to the bespoke web app

Has anyone done something similar with EMu? Does this sound resonable?

Thanks,

Brian

[Tom Cramer]

At the risk of compounding a question with another question: I know that Axiell Collections has native support for IIIF, but can’t find any evidence that Axiell EMu does. Does anyone know the reason why it’s in one but not the other? And/or if Axiell might introduce it as a core feature to EMu? 

- Tom

--
-- You received this message because you are subscribed to the IIIF-Discuss Google group. To post to this group, send email to iiif-d...@googlegroups.com. To unsubscribe from this group, send email to iiif-discuss...@googlegroups.com. For more options, visit this group at https://groups.google.com/d/forum/iiif-discuss?hl=en
---
You received this message because you are subscribed to the Google Groups "IIIF Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to iiif-discuss...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/iiif-discuss/89db9060-6760-4f66-8a33-d5ac905ab899n%40googlegroups.com.

[Andrew Wilson]

Hi Brian and Tom,

(Brian - I think your colleague Sarah may have been talking with one of my colleagues about this..my direct work email is aw...@st-andrews.ac.uk if you want to contact direct :-) ).

Here at St Andrews we've got a stack set up for Emu and IIIF : good news it's possible, bad news..it's a bit of a PITA!   Your plan sounds like a good 'un though and is relatively similar to what we did.

There's no native IIIF interface in Emu and when questioned Axiell have been very vague about it ever being implemented.  I might be wrong, but Axiell Collections only uses IIIF for the staff back-end interface, with no api/output available for image/presentation api either.

We initially used Cantaloupe to access the images via http and the IMU api (usually as jpg, converted on the fly using imu : the usually original tiff was just too slow), but performance wasn't great just dumping a jpg into cantaloupe via http.  So we now have a nightly job that exports a list of all modified multimedia files within the last 24hrs.  We then have a bash script running on the Emu server that converts the listed files into JP2 via kakadu(if an image) or just moves (other media such as 3D files etc) to a fast filestore that's also attached to the cantaloupe server.   The filenames are all named after their Emu multimedia IRN (i.e 283848.jp2), which then can be configured with cantaloupe for filestore access.   This is massively quicker than getting the images directly from Emu.    In addition to the above, I spent a looooonnnnnng month running the initial conversion scripts to actually get 400K images into JP2...that was not a fun month (as I'm sure others have had to do here too!!).

You could go down the route of attaching your emu/multimedia directory to the cantaloupe server then write a script to tell Cantaloupe to grab the correct file but you can't guarantee performance if the original image isn't in a decent format (JP2 or pTIFF).

That's for the Image API side.    For the presentation side, we export all Emu "publish on internet" catalogue records nightly as XML, then run an XSLT on them and copy to our funnelback (search service) server for ingest and index.  Funnelback is then used for producing all the front-end, including iiif manifests.

Hope that helps! Get in touch if you want a deep-dive :-)

Andrew

To view this discussion on the web visit https://groups.google.com/d/msgid/iiif-discuss/3BDE987A-D864-4149-AD22-D8EA3A2C7697%40stanford.edu.

[Brian McKenna]

Thanks for the valuable advice. Andrew, I may be in touch early next year.

[Brian McKenna]

Hi Andrew,

I finally got back to this a couple of weeks ago. I decided against a full dump of master images from EMu to Cantaloupe. The total size is .5TB and growing and that's not including the other 4 resolutions that EMu has for each image. Instead, the first person to request the image will take a hit (about 1 second) of caching the image in Cantaloupe. After that, the cached version will be used unless the image has been updated more recently in EMu (the update check is just a few ms).

How do you handle images that have been deleted in EMu? I have a couple of ideas of how to handle this but I'm curious how you dealt with this.

Thanks,

Brian