CMS Part D back from the dead

[jrlegrand]

Resolves #239

Explanation

Changed to data.cms.gov API to get file name.

For some reason, zipfile works fine without needing zipfile_deflate64 now.

Found all sorts of other odd things that I will have to deal with for automating. lightning round:

• random double space in the insulin beneficiary cost file
• various files lost a column since the last time i ran
• 2 files no longer exist since last time i ran
• zip file says 2025 but each interior zip file named 2024 Q4 - which broke my DAG and i had to hard code some stuff for now

Rationale

CMS Part D prescription plan data.

Tests

Ran DAG.

extract (download) takes 1 min
unzip takes under 5 min
Each of the load tasks run in parallel with the Postgres COPY command

• almost all take under 10 seconds
• pricing takes 3 mins
• pharmacy networks took 9 mins (i cram all 6 files into one SQL load which might be dumb) and errors out b/c my old laptop is out of space