You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Size greater than 100% should only work with the ^ notation
Elapsed time (ms): 1842
url: https://lbiiif.riksarkivet.se/v3/arkis!B0000998_00014/full/1881,1881/0/default.jpg
got: 200
expected: !200
type: size-upscalling
message: Retrieving upscailed image succeeded but should have failed as 3.0 requires the ^ for upscalling. Size: 1881,1881
warning: false
Of course, download of an image with full region and a size with values like the ones above (1881,1881) retrieves a squared image and therefore a distorted image as the aspect ratio is changed. But we do NOT understand why the validator issues a fail for "Size greater than 100%", as this size is not exceeded.
I have the same issue. It looks like the test in question generates a random number between 1100 and 2000, which assumes that the full-sized image is smaller than 1100 pixels.
s=random.randint(1100,2000)
It seems like it would make more sense for the test to read the dimensions from the image's info.json and attempt to upscale to dimensions that are known to be larger than the image dimensions.
I would like to inquire whether the size_noup test maybe is unfairly failing?
Example for our own IIIF Image API 3.0:
https://lbiiif.riksarkivet.se/v3/arkis!B0000998_00014/info.json (version 3, level1)
https://iiif.io/api/image/validator/results/?server=https%3A%2F%2Flbiiif.riksarkivet.se&prefix=%2Fv3%2F&identifier=arkis%21B0000998_00014&version=3.0&level=-1&id_basic=on&size_noup=on
Test result:
Size greater than 100% should only work with the ^ notation
Elapsed time (ms): 1842
url: https://lbiiif.riksarkivet.se/v3/arkis!B0000998_00014/full/1881,1881/0/default.jpg
got: 200
expected: !200
type: size-upscalling
message: Retrieving upscailed image succeeded but should have failed as 3.0 requires the ^ for upscalling. Size: 1881,1881
warning: false
HOWEVER,
max size for this image is 2577 x 4339, see even: https://lbiiif.riksarkivet.se/v3/arkis!B0000998_00014/full/max/0/default.jpg
So, validation against "Size greater than 100%" does not seem to be correct.
Of course, download of an image with full region and a size with values like the ones above (1881,1881) retrieves a squared image and therefore a distorted image as the aspect ratio is changed. But we do NOT understand why the validator issues a fail for "Size greater than 100%", as this size is not exceeded.
Support for feature "sizeUpscaling" (Syntax: ^size) is optional according to https://iiif.io/api/image/3.0/compliance/#32-size
We have NOT added any implementation for it, but throw errors. For instance, according to https://iiif.io/api/image/3.0/#42-size - a request of https://lbiiif.riksarkivet.se/v3/arkis!B0000998_00014/full/%5Emax/0/default.jpg results in a 501 - Not Implemented error:
"Size parameter upscaling with '^' is not implemented".
We have tried to compare with other organizations and have found that the same issue arises for their respective IIIF Image API implementations.
For instance,
https://digital.bodleian.ox.ac.uk/developer/iiif/#tag/Image
https://digital.bodleian.ox.ac.uk/developer/iiif/#tag/Image/paths/~1iiif~1image~1{image_id}~1{region}~1{size}~1{rotation}~1{quality}.{format}/get
with the image URI that is given as an example in the documentation, results in a FAIL for size_noup test:
Size greater than 100% should only work with the ^ notation
Elapsed time (ms): 1076
url: https://iiif.bodleian.ox.ac.uk/iiif/image/1363b336-260d-4f22-a6cf-4e1320dbb689/full/1177,1177/0/default.jpg
got: 200
expected: !200
type: size-upscalling
message: Retrieving upscailed image succeeded but should have failed as 3.0 requires the ^ for upscalling. Size: 1177,1177
warning: false
However, the max size image is larger (3734 x 4000): https://iiif.bodleian.ox.ac.uk/iiif/image/1363b336-260d-4f22-a6cf-4e1320dbb689/full/max/0/default.jpg
We would be very grateful for any comments and/or hints regarding the question of how to proceed from here.
Thank you so much.
The text was updated successfully, but these errors were encountered: