Support for juicing entire directories

Workaround is something like this `find coverage/srcbackup -name '*.html' | xargs -P 16 -I@ yarn juice @ @`
but it would be great if juice had native support for doing this as I suspect it'd be faster

Reproduce the current workaround with `find coverage/srcbackup -name '*.html' | xargs -P 16 -I@ yarn juice @ @`. Then trace the `yarn juice` command entry point and existing input handling; the work is done when native directory input processes the relevant HTML files without requiring the external `find` and `xargs` pipeline.

Written by the indexing model from the issue text.