Document tokenization limits

[jacobwegner]

Refs https://github.com/scaife-viewer/backend/blob/35f792914d04152cecce7426a061a9824ae5c45c/core/scaife_viewer/core/indexer.py#L140

New URNs means these will fail:

• https://scaife-dev.perseus.org/reader/urn:cts:greekLit:tlg4013.tlg001.1st1K-grc1:4
• https://scaife.perseus.org/search/?q=%22%CE%A4%E1%BD%B8%CE%BD%20%E1%BD%85%CE%BB%CE%BF%CE%BD%20%CF%83%CE%BA%CE%BF%CF%80%E1%BD%B8%CE%BD%20%CE%BA%CE%B1%CF%84%E1%BD%B0%20%CF%84%E1%BD%B4%CE%BD%20%CF%80%CE%B5%CF%81%E1%BD%B6%20%CE%BF%E1%BD%90%CF%81%CE%B1%CE%BD%CE%BF%E1%BF%A6%22&kind=form&format=instances&p=1

refs https://github.com/scaife-viewer/scaife-viewer/releases/tag/v2023-06-14-001

[jacobwegner]

I'm working on this for Brill too and will likely have a scaife-viewer-core update to address.

The limit appears to be around ~99,380 tokens.