pirl-unc / hitlist / 24921801715
80%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 23
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: 59.318% (+0.2%) from 59.093%
24921801715

push

github

web-flow 
v1.19.0: Preserve sample provenance + thread cell_name to anchor resolver (#149) (#156)

Issue #149: generate_ms_samples_table dropped most curated provenance
(286 entries with 'source', 51 with 'reference_proteomes', etc.).
generate_sample_expression_table only emitted anchor columns and lost
the rest of the sample row.  The peptide-origin attach didn't pass
row-level cell_name / source_tissue into the anchor resolver, so labels
that only matched via cell_name fell to tier 6.

- hitlist/export.py:
  - generate_ms_samples_table now exposes condition (raw, alongside
    perturbation), source, profiled, peptides, reference_proteomes
    (serialized as 'UPID:label;...').  profiled is empty when
    not curated, 'true'/'false' when explicit.
  - New _serialize_reference_proteomes helper handles list-of-dict,
    list-of-str, partial entries (UPID-only / label-only), already-
    serialized strings, and None.
  - _SAMPLE_PROVENANCE_COLUMNS canonicalizes the column set
    generate_sample_expression_table now carries through alongside
    its anchor columns — so one export captures sample identity +
    acquisition + expression anchor.
  - _attach_peptide_origin now passes cell_name and source_tissue
    (when present) into resolve_sample_expression_anchor as
    cell_name / lineage_tissue, so observations whose sample_label
    is bare (e.g. donor IDs, tumor biopsy notes) resolve through
    the registry by cell_name and through HPA tier-5 by tissue.
  - peptides count is cast to str so the export round-trips through
    parquet (mixing int + empty-string broke pyarrow).
  - n_samples=0 placeholder rows are still skipped at the export
    boundary; #149 leaves that follow-up open.

- tests/test_export.py:
  - 7 new tests covering: original condition preserved, source field
    populated, reference_proteomes serialized in real overrides,
    _serialize_reference_proteomes unit tests for dict-list / partial
    / passthrough / None / empty cases, generate_sa... (continued)

2451 of 4132 relevant lines covered (59.32%)

0.59 hits per line

Coverage Regressions

Lines Coverage ∆ File
98
80.78
 0.54% export.py
Jobs
ID Job ID Ran Files Coverage
1 24921801715.1 23
59.32
 GitHub Action Run
Source Files on build 24921801715