Add script for creating drafts with alternate postprocessing configurations

[isaac091]

Towards #746

---

This change is 

[benjaminking]

This looks good. At some point, we'll have to discuss what adding quotation denormalization to this will look like. NLLB sometimes doesn't translate all of the quotation marks from the source, so we can't go back to the source to figure out what they originally were. We could re-apply quotation normalization, but there are combinations of source and target quote conventions where normalization->denormalization->normalization will be different from the original. Anyway, not super important right now.

Reviewed 2 of 2 files at r1, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @ddaspit)

[isaac091]

The intention is that this script is only used with unedited drafts (i.e. straight from the translate pipeline) and their original sources. This is because the ScriptureRefs of the source and draft have to match up exactly (with the exception of \rem elements) so that alignment can be done for marker placement. So at least in the way that I'm thinking about this script right now, it would have access to the quotation convention of the source and assume that the quotations in the draft are either normalized or have the same convention as the source.

But yes, we will need to have that conversation.