pirl-unc / hitlist

77%

LAST BUILD ON BRANCH main

branch: SELECT

CHANGE BRANCH

x

Sync Branches

• No branch selected
• add-load-all-evidence
• add-parquet-export-test
• add-refs-aggregator-20260420
• align-version-with-pypi
• allele-resolution
• attribution-per-donor-rows
• audit-pmid-overrides
• binding-index-split
• builder-memory-reduction
• bulk-proteomics-abundance-and-metadata
• bulk-proteomics-non-tryptic-bj
• category2-per-donor-curation
• cell-name-parser
• chore/audit-cleanup-stale-docs
• chore/test-speed
• ci/corpus-cache
• codex/apc-lineage-curation
• codex/curate-pmid-24366607
• codex/curate-pmid-29093164
• codex/export-ms-peptide-summary
• codex/training-export-unified
• curate-batch-2
• curate-batch-3
• curate-top-studies
• expand-proteome-registry-bacteria
• faridi-per-transfectant
• feat/261-cell-type-column
• feat/allele-bag-expansion-137
• feat/apm-perturbation-columns
• feat/assay-iri-evidence-row-id
• feat/binding-response-measured-stacked
• feat/build-top-level
• feat/bulk-proteomics-n-replicates-possible
• feat/class-label-severity-tiers
• feat/condition-categories
• feat/curate-shapiro-2025
• feat/discrepancies-per-sample
• feat/engineered-mhc-flag
• feat/exclude-class-label-suspect
• feat/export-cleanup
• feat/export-provenance
• feat/gomez-raji-plasma
• feat/instrument-category-and-gomez-zepeda
• feat/line-expression-anchors
• feat/line-expression-cache
• feat/list-args-space-separated
• feat/maptac-dp-dq-match
• feat/normalize-and-class-suspect
• feat/observations-export
• feat/pmhc-binder-classification
• feat/pmhc-flat-and-optional
• feat/pmhc-query-and-toplevel-reshuffle
• feat/pmhc-sample-paired
• feat/pmhc-serotype-expansion
• feat/ptm-aware-peptides
• feat/qc-plan-roadmap
• feat/qc-proteome-coverage
• feat/qc-top-level
• feat/quantitative-binding-fields
• feat/register-aav-proteomes
• feat/register-mtb-plasmodium-proteomes
• feat/rename-map-source-proteins
• feat/report-from-index
• feat/severity-tiers-in-curation-plan
• feat/severity-tiers-in-qc-and-cli
• feat/training-export-136
• feat/transcript-aware-mappings
• feat/vectorize-and-version
• fix-allele-resolution-categorical-fillna
• fix-alpizar-2017-split-b-alleles
• fix-chen-2020-hela-abc-ko
• fix-classify-allele-pair-gene-gene
• fix-ebv-lcl-direct-ex-vivo
• fix-hla-only-filter
• fix-illing-2018-split-b57-transfectants
• fix-pandas-categorical-warnings
• fix-pmid-arrow-conversion
• fix-sarango-2022-precision
• fix-trolle-2016-split-721-221-transfectants
• fix-weingarten-gabbay-2021-no-hbec
• fix/255-download-timeout
• fix/audit-19-remaining
• fix/c1r-allele-curation
• fix/class-ii-implausible-45
• fix/class-ii-implausible-threshold
• fix/cross-reference-class-only-pmids
• fix/derived-column-passthrough
• fix/multi-allele-genotype-tokenization
• fix/per-sample-mhc-curation
• fix/pmhc-progress-wording
• fix/pmid-curation-audits-128-132
• fix/proteome-cache-bounded
• fix/report-keyerror-and-cli-help
• fix/run-all-include-proteome-coverage
• fix/sample-resolver-tiebreak
• fix/scanner-close-csv-files
• fix/severity-tier-ptm-aware
• ingest-abelin-2019-maptac
• ingest-bekker-jensen-peptides
• ingest-ccle-bulk-proteomics
• ingest-strazar-2023-hla2
• is-non-peptide-ligand-228
• issue-104-108-digest-and-bounds
• issue-105-atomic-rebuild
• issue-106-export-bulk-cli
• issue-110-proteome-index-memory
• issue-117-summary-from-indices
• issue-119-export-progress
• issue-12-human-51-100
• issue-121-118-normalize-and-lengths
• issue-122-short-mhc2-filter
• issue-8-nonclassical-hla
• issue-85-scanner-perf
• issue-86-proteome-index-cache
• issue-93-remove-hla-filter
• issue-98-fractionation-ph
• issue-99-proteome-kmer-set
• main
• mono-allelic-detection
• mono-allelic-method
• obliterate-legacy-index-cache
• parallelize-mappings
• per-transfectant-splits
• perf/262-arrow-xdist-cache
• perf/263-tighten-dtypes
• perf/batch-drops-and-fillna
• perf/memory-aware-test-workers
• perf/merge-and-copy
• perf/proteome-build-inline
• perf/proteome-cache-build-order
• perf/test-speed
• perf/test-suite-integration-marker
• perf/unique-map-string-ops
• pin-mhcgnomes
• pmhc-filter-flags
• pmhc-group-by-species
• predictor-multi-allele-narrowing
• profile-build-pipeline
• proteome-cache-followups
• qc/discrepancies-report
• refactor/258-rename-pmhc-helpers
• rename-proteome-label
• retire-human-only
• retro-allele-predictor
• sarkizova-monoallelic-recovery-45
• scanner-v2
• serotype-mapping
• split-gene-mapping-from-observations
• test/supplementary-build-e2e
• unified-observations
• update-readme-1.10
• upgrade-mhc-unknown
• vectorize-export-discriminator-loop

Build # 26977481678

Build Type

push

github

Committed by web-flow

Commit Message

v1.30.61: run integration suite in CI via prebuilt-corpus cache (#272)

CI couldn't run the ~20 integration tests (or exercise the #262 Arrow
mmap fixture) because the built observations corpus isn't present —
full_observations_df self-skips when ~/.hitlist/observations.parquet is
absent, and CI can't build it (the ~15 GB IEDB/CEDAR sources aren't
there). The OUTPUT is only ~237 MB, so we ship that.

- scripts/publish_ci_corpus.sh: operator-run, uploads the 5 built
  parquets to a PUBLIC ci-corpus-<version> GitHub release. (Licensing:
  the IEDB/CEDAR-derived corpus becomes publicly downloadable — a
  deliberate choice for full-fidelity CI coverage.)
- tests.yml (3.11 job only, to keep 3.9/3.10/3.12 fast): restore
  ~/.hitlist from Actions cache -> on miss, gh release download the
  corpus -> run the full suite incl. integration. Cache is only SAVED
  when the download actually succeeded, so a pre-seed run can't poison
  the key with an empty dir. A missing release logs a warning and lets
  integration tests skip, so CI stays green whether or not the corpus
  has been seeded yet.

Bump CORPUS_VERSION in tests.yml in lockstep with the release tag when
republishing.

Coverage Stats

4729 of 6155 relevant lines covered (76.83%)

0.77 hits per line