As discussed in Yoshi Python Chat andgoogleapis/python-bigquery#590 (comment)

It'd be good to have some utilities for constructing prefixes for resources based on the repo name and directory name, relative to repo root. This would include a timestamp (so that resources without a created time property could be aged) as well as a random identifier.

Likewise, we could have helpers to parse resource identifiers and determine (1) they match the repo name and directory name and (2) they are older than some time-to-live (24 hours seems a good default)