Experimental request analysis tool

This is a purely-experimental set of scripts to automatically evaluate open requests to flag specific checks for human review. This is vaguely progress for enwikipedia-acc/waca#425 , but not integrated into the UI yet, and reliant on database queries and API calls in a single batch. It is my hope that this can eventually be refactored into a set of functions which take a request ID and return a result, and that this can be cached in the main schema and retrieved on-demand.