Reads archive identifiers out of free-text citations, and proposes a structured source record for each one.