datadevtools should be able to produce the final data by reading in a datadevtools repo

[njtierney]

So something like:

read_datadevtools(<url>)

• Finds the appropriately tagged "final" data
• Uses the metadata files to give columns the appropriate class
• checks an MD5 hash to sure the data is the most recent