Skip to content

Major feature release

Latest
Latest

Choose a tag to compare

View all tags
@jeromekelleher jeromekelleher released this 08 Jun 11:30
0.2.0
7d4f86b

Major feature release adding new output formats, cloud store support, a Python API, performance improvements and bug fixes.

New output formats:

  • Add view-plink command for PLINK 1 binary output (.bed/.bim/.fam), following the semantics of plink2 --vcf X --make-bed. Use --no-bim/--no-fam to suppress sidecars.
  • Add view-bgen command for Oxford BGEN output, streamed to stdout or written with .bgen.bgi/.sample sidecars (-o STEM). Supports haploid and mixed-ploidy data.

Cloud and remote storage:

  • Read VCZ from cloud and remote stores via the optional [obstore] and [icechunk] extras; fsspec and HTTP stores are also supported.
  • Add support for .vcz.zip files (#280).
  • Support "proportional chunking" across arrays (#356).

Filtering and CLI:

  • Support -R/--regions-file and -T/--targets-file (#268).
  • Add -v/--types, -V/--exclude-types, -m/--min-alleles, and -M/--max-alleles to view and view-plink, matching bcftools view,
    along with a new N_ALT filter identifier.
  • Add N_MISSING and F_MISSING filter variables
  • Add --fill-tags to emulate bcftools +fill-tags
  • Add --log-level and --log-file options; report throughput in MiB/s.

Data types:

  • Add support for float16 and float64 data stored in Zarr (#413, #414).
  • Add support for 64-bit integer POS and INFO/FORMAT fields.

Python API:

  • Add a public Python API, documented under {ref}sec-python-api:
    • VczReader for variant data access, with sample selection by ID via set_samples(sample_ids, complement=, ignore_missing_samples=) (or by raw index via set_sample_indexes).
    • One-shot writers write_vcf / write_plink / write_bgen and streaming BedEncoder / BgenEncoder byte encoders.
    • Sidecar writers write_bim / write_fam / write_sample / write_bgi.
    • Click option bundles (ViewBgenOptions, ViewPlinkOptions, SelectionOptions, ZarrStoreOptions, ReaderOptions, LogOptions) and GroupedCommand for downstream CLI reuse.
    • Array-sentinel helpers is_missing / is_fill / trim_fill for interpreting missing and end-of-vector (fill) values in iterated arrays.

Platform and packaging:

  • Add Windows support.
  • Provide prebuilt wheels for Linux, macOS, and Windows (CPython 3.11-3.13).
  • New required dependencies: pandas and humanfriendly; pyparsing>=3.1.0. Genomic-range operations now use ruranges_py on Python 3.12+ and pyranges on earlier versions.

Documentation:

  • Add a documentation website covering installation, storage backends, the CLI, PLINK/BGEN output, and the Python API.

Deprecations:

  • --zarr-backend-storage is deprecated in favour of --backend-storage; the old name still works and emits a warning.

Bug fixes:

  • vcztools query silently truncated output on multi-chunk stores (#283)
  • vcztools view crash on FILTER expressions on multi-chunk stores (#282)
  • Incorrect output on vcztools query on FORMAT fields (#286, #287)
  • Per-allele INFO field + -s crashes (#295)
  • VCF header emitted even when filter expression is invalid (#221)
  • Incorrect behaviour for bcftools query with FORMAT scoped filters and sample subsetting (#297)
  • Incorrect output for filtering with missing Number=A INFO field + numeric (#299)
  • Noise on stderr when performing arithmetic with missing values (#301)
  • Null samples included in plink output (#310)
  • Fill values handled incorrectly by query (#415)
  • Crash evaluating non-scalar INFO field expressions (#435, #436)
Assets 2
Loading