Best Practice - MVZ Media Upload Process

[Jegelewicz]

See https://github.com/ArctosDB/documentation-wiki/issues/119#issuecomment-887078338

I uploaded a photo and managed to exit before attaching it to the specimen record. Now I cannot locate the media to make the connection and I also managed to attach an incorrect photo to the specimen.

The specimen is UTEP:Mamm:7465 The media is supposed to be at http://arctos.database.museum/mediaUploads/jegelewicz/UTEPMamm7465.jpg - but when I try to pick it, Arctos finds no matches. I tried to remove the relationship from the incorrect photo, but Arctos won't let me do it. Help.

[dustymc]

I uploaded a photo

How did you do that? (eg, does the documentation need help? There's a drag-n-drop specimen uploader, I haven't received much feedback - never sure if that means it's perfect or nobody's found it!)

You couldn't edit http://arctos.database.museum/media/7465 because the date label was not ISO8601.

I did...

search/Media find your image by URL - not there, I don't think there's a Media record (which is not the same thing as a binary object) create media, including the relationship to http://arctos.database.museum/guid/UTEP:Mamm:7465

The media record is http://arctos.database.museum/media/10566358

Should be all cleared up, but the media may take a while to clear the cache.

[Jegelewicz]

I think the documentation needs help. Here's what happened. I used the "click to upload" functionality and once my photo was there I started editing the documentation (license, etc.) and I accidentally closed the window in a moment of distraction before finishing up. So my media was created, but not attached to the specimen record. Then no matter what I tried, I couldn't make the "attach media" function find the media I had created, even with the url that Arctos was giving me (http://arctos.database.museum/mediaUploads/jegelewicz/UTEPMamm7465.jpg). As an experiment, I tried using the specimen # (7465) as the media ID, and ended up with a plant photo attached to my mammal specimen. I don't know the difference between a media record and a binary object, so I may not be the person to say what exactly we need. I did set up an account with TACC and notified Carla so that I could (hopefully) bulk upload media, but I will confess that that process seems opaque to me when reading the documentation, however I was going to try it out.

[dustymc]

Thanks.

I couldn't make the "attach media" function find the media I had created

ArctosDB/arctos#1070, perhaps - I'll did deeper when I have a chance to.

media ID

Media ID is an arbitrary (mostly-persistent) integer created when a media record is created.

difference between a media record and a binary object

A binary object (that's probably not quite the right name for it) is something with a URI, generally a URL (eg, you CAN link to stuff on an FTP-or-whatever server, but you probably shouldn't). Images (on Arctos or elsewhere), web pages, youtube videos are all "binary objects" (from the viewpoint of Arctos Media).

Media is the Arctos metadata concerning those, including a primary key (the media ID) which can be used to form relationships to other stuff (http://arctos.database.museum/info/ctDocumentation.cfm?table=CTMEDIA_RELATIONSHIP).

bulk upload media

For clarity in documentation, it should be noted that there are two completely independent components of this:

1) Getting the binary data (ie images) to an exposed server (eg, creating a URL). TACC can host, but you can also upload to your local server, flickr, or whatever.

2) Creating Media records (potentially including relationships to things like specimens) for those. If you use TACC (or something else with an XML directory list) Arctos can semi-automate this, or you can manually fill in the media bulkloader (which bulkloads metadata, NOT binaries), or you can create media records one at a time.

I'm going to create a "needs documentation" label for this. I'm happy to write docs, but they'll probably be more useful to users if they come from users.

[Jegelewicz]

Cool. I'll try to write something if I can find any time and especially once (if) I can get the TACC option working, although I may just go with Flickr... In the mean time, I'll be uploading one off photos as I have students and volunteers taking them.

[dustymc]

Flickr

Brave!

I don't think I'd trust them with my only source of data, but maybe ArctosDB/arctos#1071 ....

[ekrimmel]

@Jegelewicz I worked with Chris at TACC to bulk upload thousands of images to their server and I'd recommend going that route. He can automate the re-sizing and TACC is a safer, more dependable place to store media. We are now at a point that sounds similar to you, where we have interns and volunteers creating new images just a few at a time. My plan is to still process these in batches with TACC/Arctos media records. I think doing so provides better consistency and is more efficient than uploading and creating records singularly or in small groups... I don't think MVZ's TACC tutorial is on the handbook. @mkoo or @ccicero could link?

[ekrimmel]

nvmd I found the MVZ tutorial. It won't be exactly the same for you but if you haven't already seen this it's a good overview of how the bulk media process works.

[Jegelewicz]

@ekrimmel Thank you!!

[Jegelewicz]

@mkoo could your MVZ workflow be made into a best practice?

[mkoo]

let me clean up the spilled coffee from snorting so much...

I think that's a good idea once we create a new workflow with TACC's help since we've been talking about changing up some of the key parts of the workflow (eg., management of uploaded material via SFTP or SSH etc). Also, if Arctos starts getting IIIF savvy then hopefully some of our workflow will become obsolete! But let's put that on the to-do list here...

On Mon, Jul 26, 2021 at 3:49 PM Teresa Mayfield-Meyer < @.***> wrote:

@mkoo https://github.com/mkoo could your MVZ workflow be made into a best practice?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/ArctosDB/documentation-wiki/issues/119#issuecomment-887078338, or unsubscribe https://github.com/notifications/unsubscribe-auth/AATH7UKCSVMVBJYX35OUN6DTZXQ7PANCNFSM4I4YIU3A .

[Jegelewicz]

I figured as much - we'll just leave this here for now.