merge Adam's word segmentation improvements into main branch

[emylonas]

Adam split out the word segmentation script into components and cleaned up some bits. This is in a branch on Adam's fork of the iip-texts repo.
Some information is in an issue on iip-production repo.

Should test and merge. This includes a separate script to add IDs to elements. Could be very useful for problem files.

adding @zeichman as an fyi