has someone already written the javascript to scrape the hi-res tiles from the minnesota institute of art? they are rudely claiming they provide hi-res downloads which, um, no, not *close* to what they have in the viewer.
(I Need a print of this)
https://collections.artsmia.org/art/139987/blood-collage-john-bingley-garland
scraping procedure, long
@maya Not in the brainspace to write code right now, but the size is encoded in the image's metadata at https://iiif.dx.artsmia.org/139987.jpg/info.json (number matches the item number in the original URL), in the width/height properties.
Each tile is then at https://iiif.dx.artsmia.org/139987.jpg/0,0,512,512/512,/0/default.jpg, where first number is the item ID again, the URL segment after that is startX,startY,endX,endY and the URl segment after *that* is outputWidth,outputHeight (latter can be omitted like it is here, will default to the former).
If outputWidth is equal to endX-startX (and same for height/y), then you get the tile at the original resolution, if not then it gets scaled to the specified output size.
512 pixels seems to be maximum output dimensions per tile, but a loop to fetch all the tiles based on the width/height metadata + something like imagemagick/graphicsmagick should be able to stitch it back together
re: scraping procedure, long
@joepie91 @maya here is the spec for this; it’s an open standard https://iiif.io/api/image/3.0/#21-image-request-uri-syntax
but keywords like `full` and `max` don’t seem to be working
re: scraping procedure, long
@maya @Lady cc @huertanix museum-archival-digital tech
re: scraping procedure, long
@maya yeah if you just email like “hey could you hook me up with the original for X public domain resource” i would be very surprised if they were like “no”
re: scraping procedure, long
@maya i’m wondering if some of the resources in their collections have licensing restrictions which allow them to display them publicly but forbid duplication and that’s why their IIIF throws an error above a certain size
re: scraping procedure, long
@Lady I have written this email! Thank you for giving me the courage to do so lol.
They have a policy where they don't provide hi res versions of material not in the public domain so maybe it's something there getting twisted
re: scraping procedure, long
@maya yeah. the reason why i was expecting it to work is because this is what the “save detail” thingy in the bottom right (also broken) uses. i feel like they would have just hidden it if it being unavailable was intentional?? (but i also know how understaffed GLAM tech teams can be so who knows)
re: scraping procedure, long
@Lady they seem to use a different API for their internal use, and the README suggests they *mean* this only to be internal https://github.com/artsmia/collection-tools/blob/master/miamg
still, I'll give 'em an email to see if they can hook me up. the blood collages are public domain (in letter and spirit) so I'm hopeful :)