Add an endpoint to api/extractors that allows rerunning of specific extractors. Specifics can be discussed (e.g. re-run on all? re-run only on objects that are missing? option for both? etc.) but the goal is to make it easy to deploy new extractors on older files or re-run updated extractors on existing files.

This (and elsewhere that fetches file lists) should use Mongo's file iterator wrapped, so we aren't pulling a huge list of files at once.