pirl-unc / tsarina / 24684145316
73%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 30
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: 42.345% (+0.8%) from 41.534%
24684145316

push

github

web-flow 
v0.9.0: tsarina hits --format refs — per-pMHC ref/tissue breakout (#13)

New aggregation between the existing peptides (per-peptide roll-up) and
pmhc (per-(peptide, allele), lean) formats. Groups by (peptide, allele)
and surfaces the provenance columns reviewers want for therapy
prioritization: ref_count, pmids (sorted numerically), tissues, diseases,
cell_lines, in_cancer, in_healthy_tissue, mono_allelic_hit_count.

Implementation notes

- tsarina.cli_hits._aggregate_refs is a thin groupby over raw scan
  columns. Optional columns absent from the input are silently skipped
  so both the cached-observations fast path and the raw-CSV slow path
  produce valid output.
- PMIDs sort numerically (not lexicographically) — prevents a dormant
  bug when PMIDs roll to 9 digits.
- Docstring explicitly documents which output columns are guaranteed
  vs conditional on input.
- CLI plumbed through _SUPPORTED_FORMATS + --format help text.

Test coverage: tests/test_cli_hits_refs.py — 6 tests (per-pMHC
aggregation with PMIDs/tissues/cancer flags, canonical empty-frame
shape, skinny-input robustness, CLI choice registration, 9-digit PMID
numeric sort regression, positive in_healthy_tissue aggregation).
Live-tested with `tsarina hits --gene PRAME --serotype A24 --format refs`.

Follow-ups: CLI smoke test + merge-path coverage (#14), single-pass
pmid aggregation (#15), upstream to hitlist.aggregate (hitlist#88).

791 of 1868 relevant lines covered (42.34%)

0.42 hits per line

Coverage Regressions

Lines Coverage ∆ File
82
45.68
 6.68% cli_hits.py
Jobs
ID Job ID Ran Files Coverage
1 24684145316.1 30
42.34
 GitHub Action Run
Source Files on build 24684145316