[Question] The images for MS234 on Cantus are in PDF format and aren't numbered by folio; what do we do?

[JoyfulGen]

The link to the external images for MS234 on CantusDB brings the user to a PDF scan of the full manuscript. Since we're planning on running the manuscript through Mothra, there are two issues that I can perceive with this:

1. We need image formats of individual folios, so I guess we'll need to convert every page?
2. Since the images are in a PDF doc, the metadata numbers them by page, not by folio. But CantusDB numbers them by folio, which means that in Mothra, if I lock in the CantusID and want to select a folio or range of folios, the dropdown menu shows me folio numbers. Some of the folio images have folio numbers written in Roman numeral at the top, but some of them don't. So far I've dealt with this by figuring out the folio number myself, one folio at a time, but that won't work for the whole manuscript!

@kyrieb-ekat thoughts?

[annamorphism]

For image formats: we may be able to get better images from the library (maybe)--I think they have .tiff files somewhere, but didn't put them online. That might not change the need to convert every page, though.

2 is a pretty similar problem to what happens in the mapping tool in Cantus Ultimus, except there of course we have a IIIF manifest to play with. But a similar solution might work. (There is in any case no guarantee that Cantus's numbering of pages matches some other metadata source, even if we weren't dealing with a PDF...)

[JoyfulGen]

@annamorphism what's the mapping solution for Cantus Ultimus?

[annamorphism]

there's a screenshot in DDMAL/cantus#616 (the interface could be improved, but it didn't really seem worthwhile if the folios were going to be coming from Neon anyway (see discussion in DDMAL/cantus#645). But the mapping has to happen somewhere...

[kyrieb-ekat]

I have all of MS234's images, which I'll pop onto hugging face shortly. For a short-term solution, I'll write a script that'll match my images to the csv from CDB, but long-term we definitely will want a process at hand to sync this, though again this only addresses those manuscripts which also happen to be CDB manuscripts...