Hypothesis tests for VCF

sgkit-dev / bio2zarr

Convert bioinformatics file formats to Zarr

Apache License 2.0

23 stars 5 forks source link

Hypothesis tests for VCF #264

Closed tomwhite closed 3 weeks ago

tomwhite commented 3 weeks ago

Fixes #249

This should be working now, but does it need a hypothesis-vcf release before merging (or not)?

coveralls commented 3 weeks ago

coverage: 98.848% (-0.04%) from 98.887% when pulling dc9d398f95021a71c98ed1bee8d728032a000091 on tomwhite:hypothesis-vcf-tests into d1920541bf36101d5e3aa7656094901182c0e98b on sgkit-dev:main.

tomwhite commented 3 weeks ago

The failure here was because hypothesis created a VCF with a GT INFO field of type Character, and bio2zarr assumes all GT fields have type Integer. The VCF is valid (but almost certainly never occurs in practice), so I think the fix is to change bio2zarr to only interpret GT FORMAT fields as Integer.

coveralls commented 3 weeks ago

coverage: 98.848% (-0.04%) from 98.887% when pulling 48df4cc312eb64ae3bb5a5f0ad6860f062fa3c35 on tomwhite:hypothesis-vcf-tests into d1920541bf36101d5e3aa7656094901182c0e98b on sgkit-dev:main.

coveralls commented 3 weeks ago

coverage: 98.89% (+0.003%) from 98.887% when pulling 48df4cc312eb64ae3bb5a5f0ad6860f062fa3c35 on tomwhite:hypothesis-vcf-tests into d1920541bf36101d5e3aa7656094901182c0e98b on sgkit-dev:main.

jeromekelleher commented 3 weeks ago

I'm happy to merge if you are?