pirl-unc / hitlist / 24700010067
80%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 21
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 48.105%. Remained the same
24700010067

push

github

web-flow 
v1.13.1: remove silent HLA-only filter from pMHC aggregators (#93) (#100)

`aggregate_per_pmhc` and `aggregate_per_pmhc_with_refs` both ended
with a trailing filter:

    result[result["mhc_restriction"].str.startswith("HLA-", na=False)]

Silently dropped any mouse / rat / other non-HLA rows with no warning
or opt-out. Species filtering already happens upstream at the scan /
load_observations layer via `mhc_species=`; having the aggregator also
enforce HLA-only is a second independent filter that doesn't belong
there and re-introduces the split-API pattern that 1.9.0 consolidated
around.

Surfaced by tsarina PR #17 where `tsarina hits --species "Mus musculus"
--format refs` started silently returning empty after switching to
aggregate_per_pmhc_with_refs (v0.9.0 used tsarina's local aggregator
which did not filter by locus prefix).

Breaking change: callers relying on the silent HLA filter will see
non-HLA rows reappear. Low likelihood — mhcflurry / NetMHCpan
downstream consumers typically species-filter upstream anyway — but
worth a minor bump. The fix aligns with the 1.9.0-era scanner API
consolidation.

- Delete trailing HLA startswith filter from both aggregators.
- Update `aggregate_per_pmhc_with_refs` docstring to document that
  species filtering is the caller's responsibility upstream.
- Rename `test_aggregate_per_pmhc_with_refs_filters_non_hla` to
  `test_aggregate_per_pmhc_with_refs_preserves_non_hla`; asserts
  non-HLA rows now pass through unchanged.

Closes #93.

1472 of 3060 relevant lines covered (48.1%)

0.48 hits per line

Jobs
ID Job ID Ran Files Coverage
1 24700010067.1 21
48.1
 GitHub Action Run
Source Files on build 24700010067