Add filter and setup Markdown normalization for sentence-per-line

[alerque]

The filter situation for sentence-per-line seems to have improved. I might need to adapt the existing options some for Turkish, but it seems like we might be able to crack this nut.

jgm/pandoc#9780

Ideally we'd have a filter for all imports/exports such that we can keep normalized Markdown as one line per sentence locally for diff purposes while still round tripping to other formats.

[alerque]

This (sentence wrapping in Markdown, import/export that round trips to/from that wrap model) is done and in the master branch.

I'll probably make it optional somehow before the next release, for now it's just the default.