csv not pandas

Instead of using pandas/numpy to load CSV into XLSX files, use the csv stdlib

This will allow us to remove pandas from the ironbank-modules and subsequently pipeline-runner, reducing the uncompressed size by nearly 200MiB. Also, pandas and its dependencies have not been updated since 2022 in ironbank-modules.

If for some reason pandas was to be kept, it needs to be updated/tested as the old versions are not packaged for python3.12 nor musl.

trigger: https://code-ib-zelda.staging.dso.mil/ironbank-tools/tools/trigger/-/jobs/215255

fb-all-before.xlsx fb-all-after.xlsx fb-just-before.xlsx fb-just-after.xlsx