vidrl_upload: fix serum_strain used for human sera measurements

[joverlee521]

While working on #161, we learned that VIDRL uses proxies for human sera references that do not exactly match the vaccine strains listed in VACCINE_MAPPING.

We should update the mapping to the correct proxy strains that are used in the flat files. Because the serum_strain is one of the index fields, we will need to delete and re-upload the previously ingested human sera data to correct them.

TODOs

• vidrl_upload: Update VACCINE_MAPPING to match proxies #167
• delete previously ingested human sera data
• re-upload with the updated VACCINE_MAPPING.

[joverlee521]

Deleted previously ingested human sera records with

$ envdir ../env.d/seasonal-flu/ python3 scripts/delete_records.py -db vidrl_tdb -v flu --filter serum_host:human
Connected to the "vidrl_tdb" database
Delete filters: {'serum_host': 'human'}
Delete intervals: {}
Deleted 7283 records

[joverlee521]

Used this Snakefile to re-upload all human sera data from 2024. This uploaded 7859 measurements (576 new measurements from https://github.com/fludata/VIDRL-Melbourne-WHO-CC/commit/7f71cbca8e6f3d964b2d036386513a148d04521a that were missed in my previous backfill)