0.19.1

0.19.1 (2025-12-20)

Fix

• perf: upgrade genoray for (much) faster writes. test: idempotent fixutres in test_ds_haps

[main e41f7f7] bump: version 0.19.0 → 0.19.1
2 files changed, 7 insertions(+), 1 deletion(-)

0.19.0

0.19.0 (2025-12-03)

Feat

• allow subsetting by region name. fix: convert eligible ak.Array to sp.rag.Ragged for gvl.RaggedVariants whenever possible

[main babe777] bump: version 0.18.3 → 0.19.0
2 files changed, 7 insertions(+), 1 deletion(-)

0.18.3

0.18.3 (2025-11-09)

Fix

• perf: faster reverse complementing

[main 0690647] bump: version 0.18.2 → 0.18.3
2 files changed, 7 insertions(+), 1 deletion(-)

0.18.2

0.18.2 (2025-11-03)

Fix

• correctly parse and load variant fields (skip duplicates)
• use ak.str.length instead of ak.num to get ref and alt lengths. docs: shape docstring

[main c1900e9] bump: version 0.18.1 → 0.18.2
2 files changed, 8 insertions(+), 1 deletion(-)

0.18.1

0.18.1 (2025-10-23)

Fix

• track file format version

[main 877b6d4] bump: version 0.18.0 → 0.18.1
2 files changed, 7 insertions(+), 1 deletion(-)

0.18.0

0.18.0 (2025-10-22)

Feat

• make RaggedVariants an Awkward Array subclass supporting arbitrary additional fields.

[main 27420c0] bump: version 0.17.0 → 0.18.0
2 files changed, 7 insertions(+), 1 deletion(-)

0.17.0

0.17.0 (2025-08-22)

❗Breaking Changes❗

• The layout of Ragged offsets has changed to match Awkward for better performance. This ONLY affects SVAR datasets which must have their offsets rewritten to have shape (2, n_regions, n_samples, ploidy) instead of (n_regions, n_samples, ploidy, 2). The SVAR genotype metadata shape info must be updated to reflect this as well. In addition, any code that uses Ragged.offsets directly may need to be adjusted as 2-D offsets have had their axes transposed.
• The transform and return_indices arguments have been moved to Dataset.to_torch_dataset and Dataset.to_dataloader

📦Additions📦

• Dataset.n_intervals() and Dataset.n_variants() methods to get arrays of how many intervals and variants are present in each region/sample/ploid/track
• Dataset.with_seqs('variants') and Dataset.with_tracks(kind='intervals') to return RaggedVariants and RaggedIntervals, respectively. These are simple containers of Ragged arrays. Methods to convert RaggedVariants and RaggedIntervals to dictionaries of PyTorch Nested tensors.
• Ragged arrays are now a subclass of Awkward Arrays. As a result, Ragged arrays now support all element-wise operators, all awkward functions, and NumPy ufuncs. Their indexing behavior also directly inherits from awkward arrays. As a result, the Ragged.to_awkward() method is deprecated. If you want to remove the Ragged behavior from a Ragged array and get a plain awkward array, you can use Ragged.to_ak(). Awkward arrays can be converted to Ragged ones by simply passing an awkward array to the Ragged constructor directly: Ragged(ak_array).
• gvl.data_registry.fetch() to obtain pre-processed datasets. Currently includes Geuvadis with RNA-seq tracks and the full 1000 Genomes Project WGS data as VCF/PLINK and BigWigs that can be passed to gvl.write().

Commits

Feat

• use new seqpro.Ragged interface
• method to obtain the number of variants per region,sample,haplotype.
• support for reverse complementing alt alleles of RaggedVariants
• ragvariant methods for pytorch, fix squeeze for scalar indexing, upstream genoray fix for extending genotypes from PGEN, (perf) batched contig normalization
• subsetting for RefDataset

Fix

• right shape for double slice indexing
• shape/broadcasting bug for reverse complementing variants
• perf: breaking changes to ragged data layout from seqpro to eliminate copies when converting non-contiguous ragged data to/from awkward arrays.
• when ref not passed to Dataset.open, emit a warning instead of error and let ds return raggedvariants
• refdataset transforms with indices

[main d38797c] bump: version 0.16.0 → 0.17.0
2 files changed, 19 insertions(+), 1 deletion(-)

0.16.0

0.16.0 (2025-06-05)

Feat

• make DatasetWithSites return both wild-type and mutant haplotypes
• data_registry submodule

Fix

• transform not applied when dataset returns single item. docs: add basenji2 evaluation
• use genoray>=0.12. docs: basenji2 eval
• ensure samples are re-ordered by subset_to if necessary
• Let GVL recognize bgz-compressed VCFs
• pad ref_coords with max value for dtype to ensure ref coords are sorted
• remove transform arg for dummy dataset
• finish deprecating the transform setting on Dataset, which was moved to dataloading functionality
• PR #101, ensure variable length output corresponds to ArrayDataset
• update genoray pixi version
• permit ragged output for dataloading, emitting a warning instead of raising an error
• constrain genoray for breaking changes
• torch dataset issues
• bump seqpro to 0.4.2

[main 30ef648] bump: version 0.15.0 → 0.16.0
2 files changed, 23 insertions(+), 1 deletion(-)

0.15.0

0.15.0 (2025-05-23)

⚠️ Breaking Changes ⚠️

• The .with_indices and .with_transform methods have been moved to be arguments to the .to_dataloader and/or .to_torch_dataset methods. Motivation: returning indices and transforms are not necessary outside of a dataloading context. Best practice to facilitate subsetting with GVL is to obtain a PyTorch Datset via gvl.Dataset.to_torch_dataset(), pass this to torch.utils.data.Subset, and then provide to this gvl._torch.get_dataloder. This is a very thin wrapper around the the native PyTorch DataLoader with better defaults for GVL to take advantage of multithreading (better than multiprocessing). Documentation updates are pending.

📦 Notable Features 📦

• The RefDataset class, an interface for lazily accessing and organizing subsequences from a reference genome
• The DatasetWithSites class, which combines a Dataset with a table of site-only variants e.g. from ClinVar or gnomAD and inserts the site-only variants into haplotypes. Currently only 1 variant at a time and only SNPs.
• Annotation tracks for Datasets, which take a BED file and broadcast its values to be treated as a track for all samples.
• Now includes SeqPro >= 0.4.2, which adds functionality for scanning GTFs as polars.LazyFrames and extracting GTF attributes via the scan and attr functions under the seqpro.gtf namespace.

Feat

• add reference property to Dataset and add path attribute to Reference
• add RefDataset to work with one or more reference genomes. Also change internal indexing to never materialize a full dataset index, dramatically reducing memory usage to support datasets with 1M+ regions. BREAKING CHANGE: move returning indices and transforms to torch API, since these features are generally unnecessary for non-dataloading contexts.
• sites-only changes for QoL. fix: consider output length < region length for sites overlap
• allow tracks to pass-through dataset with sites since SNPs have no affect on them
• wip: pad ragged annotated ref coords with max dtype value. pass sanity checks.
• wip: change ref_coord annotation so that right-pad values have position MAX_I32
• type-safe Dataset, passes all tests.
• refactor Dataset implementation to be (almost) fully type-safe.
• wip: sites-only variants
• wip: use sp.bed functions.
• wip: small updates
• apply sites-only SNPs, filtering non-SNPs out from VCFs.
• sites-only classes, intersecting them with Datasets, and obtaining information necessary to apply variants.
• add annot track to dummy dataset
• wip: initial implementation for read/write annotation tracks, incorporating them along the track dimension.
• deprecate unphased variants
• wip: dosages/CCFs on ragged variants
• prototype of returning ragged variants from Dataset

Fix

• shape of single item from RefDataset
• update init for seqpro bump
• bump seqpro to 0.4.0 which includes basic gtf ops
• update dummy dataset for changes to Reference. docs: add more docstrings
• jittering by folding it into data (re)construction
• contig offset mapping for in-memory reference and incrementing offset when writing cache
• bump genoray version to handle unsorted PVAR contigs
• bump genoray version for filtered PGEN fix
• ensure annot tracks match on-disk ordering
• update for internal breaking changes
• check for SNPs
• bump genoray so bioconda pgenlib is valid
• pass all tests
• internal breaking changes
• treat POS as 1-based to match VCF spec
• type annotations
• type annotation
• contig naming for reference fasta.
• contig normalization
• pass all tests.
• pass tests.
• pass tests.
• Dataset.open returns highest complexity ds by default (haps + all tracks, sorted).
• use pandera polars not pandas
• correct manipulation of active tracks
• dummy dataset
• wrong germline ccfs for 3rd germline variant and beyond
• wrong geno path
• parsing SVAR metadata, bump genoray

Refactor

• use genoray

[main c0fd8be] bump: version 0.14.4 → 0.15.0
2 files changed, 60 insertions(+), 1 deletion(-)

0.14.4

0.14.4 (2025-05-12)

Fix

• data corruption when rc_helper is parallelized

[main 39e4ac3] bump: version 0.14.3 → 0.14.4
2 files changed, 7 insertions(+), 1 deletion(-)