Example:

• page url → testbooru
  • https://www.pixiv.net/en/artworks/unlisted/{alphanumeric-id}
• file url → testbooru
  • https://i.pximg.net/img-original/img/{timestamp}/{digits}-{32-hex-digits}_p{page}.png

I was able to upload using the direct file URL without issues, so it should be feasible to implement support for the page URLs.
There's a number of angles to consider though — such as whether there's a way to map from file url to page url (normalisation) like with ordinary pixiv works, and whether the pixiv: metatag needs to be extended to cater for the alphanumeric IDs.
Presumably they've used these non-integer IDs to prevent scrapers discovering them by simply iterating through artwork IDs sequentially.