openvax / varcode / 25521488744
88%

DEFAULT BRANCH: main
Ran
Jobs 0
Files 0
Run time –
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
25521488744

push

github

web-flow 
Drop runtime PyVCF3 dep: own VCF header/INFO/FORMAT parsing (#302) (#356)

* Drop runtime PyVCF3 dependency: own VCF header/INFO/FORMAT parsing (#302)

PyVCF3's vcf/filters.py imports rpy2 unconditionally at module top, so
even after the lazy-import shim from #346 the moment any caller did
`from varcode import load_vcf` the rpy2/R init kicked in. varcode only
ever used PyVCF3 for header metadata + INFO/FORMAT cell decoding (via
two private methods, _parse_info / _parse_samples). Pandas already
parses the row structure.

Replace that surface with varcode/vcf_parsing.py:VCFHeader, ~250 lines.
Behaviour is pinned to PyVCF3's via tests/test_vcf_parsing.py — for each
fixture VCF we parse with both, compare every header field, every row's
INFO, every row's per-sample data. Plus targeted unit tests for header
quoting, multi-value fields, reserved-but-undeclared keys, missing cells.

PyVCF3 stays as a test-only dep (oracle); requirements.txt loses it,
pyproject.toml gains a [test] extra. The lazy __getattr__ shim and its
regression test are gone — `varcode/vcf.py` no longer imports the `vcf`
package, so import varcode is rpy2-free unconditionally.

* Add canonical VCF spec example fixtures + dedicated tests + usage script

Drop in two small VCFs reproduced from the official samtools/hts-specs
documents (MIT licensed):

- tests/data/spec_examples/vcf42_spec_trio.vcf  — VCFv4.2 §1.1 three-sample
  trio. Exercises Number=A/2, Flag fields with embedded commas in
  Description, multiallelic ALTs, microsatellite indel, ALT="." rows,
  phased+unphased GT, missing HQ values.
- tests/data/spec_examples/vcf43_spec_sv.vcf    — VCFv4.3 §5.4 structural
  variant example. Exercises symbolic ALTs (<DEL>/<INV>), END=, CIPOS
  pairs, IMPRECISE flag, SVLEN with Number=. variable count.

Both are picked up automatically by the existing PyVCF3-oracle parametrized
tests (test_vcf_parsing.py now globs recursively under tests/data/), and
also pinned to hand-verified expected values... (continued)
Source Files on build 25521488744
Detailed source file information is not available for this build.