Sizes
Not every collection has three distinct tiers (the Codex Arundel’s reading
files are its full-resolution files; Windsor has no reading tier). Asking
for a tier a collection lacks returns the nearest larger one.
Archives
0006_ff-3v-12r.jpg), a manifest.csv mapping every file to its folio label,
source URL and page on the site, the collection’s transcriptions.json where
e-Leo transcribed it, and a README.txt carrying the institutional credits.
The all archive holds one folder per collection.
Archives are assembled while you download — nothing is cached ahead of time
unless a pre-built copy exists — so a full-resolution codex takes a few
minutes and the whole library at full resolution is many gigabytes. For those,
use the URL list.
URL lists
A plain-text list of every source file, forwget, curl or aria2, which
can resume and retry:
&format=csv for index,label,file,source_url,page_url rows instead.
One page
Content-Disposition so browsers save it under a
self-describing name (veni-vidi-da-vinci_codex-arundel_0006_ff-3v-12r_full.jpg).
The same image URLs are in every /v1/folios/{slug}/{index}
response under image.