Re-map score sets with mapped variants that have neither a mapping nor an error #820

Problem

Three of the 1,170 mapped score sets in the public dump have mapped variants with both post_mapped and error_message null. They came from dcd-mapping runs that were killed during BLAT alignment, most likely out of memory, before the mapper could report a reason, and the worker stored null for the missing error.

Both causes are gone on the code shipping in 2026.3.0:

• dcd_mapping2 runs BLAT with shell=False, a timeout and an output file (align.py:185), so BLAT's output is no longer held in memory.
• RT (the allele-centric mapping work on feature/bencap/allele-centric-mapping-and-storage)'s mapping job requires a typed outcome on every mapped record and fails the job when one is missing (worker/jobs/variant_processing/mapping.py), so a null/null row can't be written.

Scope

• After 2026.3.0 deploys, find the score sets whose current mappings have neither a mapping nor an error.
• Re-run mapping for those score sets.

Acceptance criteria

• The query for current mapping records with no post-mapped allele and no error returns no rows in prod.
• The affected score sets are listed in this issue, with their re-mapping job ids.