For your first question, I am not aware of any changes to the PUMA definitions since 2010 so I don’t see why using '12-'19 PUMA data would cause any issues.
Sadly there is no straighforward way to accomplish your second ask around a 1:1 crosswalk of PUMAs to counties. Some counties are in multiple PUMAs and vice-versa. You’d need to do some kind of probabilistic match, which may be difficult if it’s not been done already.
See this helpful answer from an IPUMS Staff member: Crosswalk County FIPS to PUMA/MIGPUMA from 2005-2019 - #2 by JonathanSchroeder
“Some counties correspond to multiple PUMAs, and vice versa, so it’s not possible to allocate data directly from one to the other.”