25170780743
80%

Ran 30 Apr 2026 02:24PM UTC

Jobs 1

Files 26

Run time 1min

Badge

Embed ▾

Committed 30 Apr 2026 02:22PM UTC coverage: 70.223% (+0.03%) from 70.19%

Build # 25170780743

Build Type

push

github

Committed by

web-flow

Commit Message

v1.30.11: exclude_class_label_suspect filter on load_observations / export (#182) (#196)

Adds a one-line opt-out for IEDB class-label drift. v1.30.0's
mhc_class_label_suspect flag tagged rows where the bimodal length
distribution disagrees with the curated class (class II ≤10aa or
class I ≥18aa) but consumers still had to roll their own filter.

API surface:
- load_observations(..., exclude_class_label_suspect=True)
- load_binding(..., exclude_class_label_suspect=True)
- load_ms_observations(..., exclude_class_label_suspect=True)
- load_all_evidence(..., exclude_class_label_suspect=True)
- generate_observations_table(..., exclude_class_label_suspect=True)
- generate_ms_observations_table(..., exclude_class_label_suspect=True)
- hitlist export observations --exclude-class-label-suspect

The filter pulls mhc_class + peptide into the read columns whenever
it's enabled — even if the caller projects with columns=["pmid"] —
so the suspect flag can be computed before trimming back to the
caller's projection. No behavior change unless explicitly opted in.

Coverage Stats

3521 of 5014 relevant lines covered (70.22%)

0.7 hits per line

Coverage Regressions

Lines	Coverage	∆	File
80	55.01	0.07%	cli.py
73	81.46	-0.11%	export.py
8	93.94	0.29%	observations.py

Jobs

ID	Job ID	Ran	Files	Coverage
1	25170780743.1	30 Apr 2026 02:24PM UTC	26	70.22	GitHub Action Run

pirl-unc / hitlist / 25170780743
80%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Coverage Regressions

Jobs

Source Files on build 25170780743

pirl-unc / hitlist / 25170780743 80%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Coverage Regressions

Jobs

Source Files on build 25170780743

pirl-unc / hitlist / 25170780743
80%

README BADGES
x