Add a word-by-word diff option

[james-perretta]

Likely using wdiff and writing a custom find and replace for the [-xxx-] and {+xxx+} markers. Needs to be able to handle those markers potentially being in the original output.

We can probably convert the parsed output into the list format that we currently use so that we can reuse the current html code on the frontend.

[james-perretta]

Parsing wdiff output is challenging. It's unclear what the best way to handle occurrences of the start/end markers that happen to be in the original text.