pirl-unc / hitlist / 25174346523

Build # 25174346523

Build Type

push

github

Committed by web-flow

Commit Message

v1.30.13: qc plan — combined per-PMID curation roadmap (#198)

Joins discrepancies + cross_reference + normalization_drift into
one priority-ranked table. Answers "which study should I curate
next?" without bouncing between three reports.

Each PMID gets one row with:
- suspect_class_label_n / _rate (#182)
- monoallelic_class_only_n (#45 candidates)
- class_pool_n (#37 candidates)
- nonstandard_aa_n
- yaml_only_alleles_n (curation typo / unmeasured)
- data_only_alleles_n (gap in YAML)
- normalization_drifts_n (YAML allele needs editing)
- priority_score (weighted sum: drift x10000 > yaml-only x1000 >
  data-only x100 > row-level signals x1)
- severity (worst across all signals)

Sorted descending by priority_score so the top of the list is the
highest-leverage curation work — drift fixes the YAML once and
benefits every downstream consumer; row-level cleanups only matter
for that one study.

API: qc.curation_plan(mhc_class=None, min_rows=50)
CLI: hitlist qc plan [--class I|II] [--min-rows N] [--top N]
                     [--severity info|warn] [--output PATH]

Coverage Stats

3578 of 5079 relevant lines covered (70.45%)

0.7 hits per line

Source File

Press 'n' to go to next uncovered line, 'b' for previous

Source Not Available