pirl-unc / hitlist / 25163184658
80%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 25
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: 69.937% (+0.2%) from 69.733%
25163184658

push

github

web-flow 
v1.30.7: End-to-end test coverage for supplementary build + parquet export (#29) (#191)

Closes #29.

The two gaps the issue called out:

1. ``scan_supplementary()`` had a parquet round-trip test, but only
   against the shipped manifest. CI that doesn't carry the
   supplementary CSVs (or that's running on a stripped-down install)
   got nothing. The new fixture-based test writes a tiny synthetic
   manifest + CSV under ``tmp_path``, monkeypatches
   ``_MANIFEST_PATH`` / ``_SUPP_DIR``, and exercises the whole scan +
   normalize + classify + parquet round-trip path unconditionally.

2. ``generate_observations_table()`` had session-scoped tests via
   ``full_observations_df``, which silently skip when
   ``observations.parquet`` isn't built (most CI environments). The
   new fixture-based test builds a 3-row synthetic obs.parquet (one
   mono-allelic row, one multi-allelic, one class-only sentinel) plus
   a fixture ``load_pmid_overrides`` dict, and asserts each row gets
   the right ``sample_match_type`` (allele_match for the first two,
   pmid_class_pool for the sentinel — which is why the multi PMID's
   sample list has TWO entries: single_sample_fallback only fires for
   PMIDs with exactly one sample).

3. The CLI parquet output path of ``hitlist export observations -o
   foo.parquet`` was never exercised in tests. Two new tests invoke
   ``main()`` against the synthetic fixture with a ``.parquet`` and a
   ``.csv`` output target and round-trip the result.

Six tests in total (~1.3s on the test runner). They run on every CI
without needing a built corpus, so they catch regressions in the
build-flow and export-flow contracts that the existing skip-if-built
tests would miss.

3429 of 4903 relevant lines covered (69.94%)

0.7 hits per line

Jobs
ID Job ID Ran Files Coverage
1 25163184658.1 25
69.94
 GitHub Action Run
Source Files on build 25163184658