Skip to content

Latest commit

 

History

History
35 lines (25 loc) · 2.61 KB

File metadata and controls

35 lines (25 loc) · 2.61 KB

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[1.0.1] - 2026-06-14

Fixed

  • MultiVAE(losses=None) no longer raises TypeError: argument of type 'NoneType' is not iterable. losses=None (the documented default, meaning MSE for all modalities) now works correctly in both the model wrapper (_multivae.py) and the torch module (_multivae_torch.py).
  • MultiVAE.load_query_data no longer raises AttributeError when the reference model was trained with integrate_on but without an nb/zinb loss (i.e. theta is None).
  • StratifiedSampler (used when integrate_on is set) no longer silently produces zero training batches on small datasets. Training will now always produce at least one batch per group, and raises a clear ValueError if the batch size is too large for the data.
  • organize_multimodal_anndatas now preserves .var metadata from each modality. When the same column name appears in multiple modalities it is suffixed with the modality index (e.g. mean_0, mean_1). Fixes #4.
  • Defensive .get("columns", []) in the continuous-covariate registry lookup prevents a KeyError with certain scvi-tools versions. Fixes #45.
  • Removed deprecated dtype= argument from anndata.AnnData() call (anndata 0.12+).

Changed

  • Requires Python ≥ 3.12 (aligned with scvi-tools 1.4.3 minimum).
  • Bumped core dependencies: scvi-tools>=1.4.3, anndata>=0.12; removed numpy<2 and jax<0.6 upper-bound pins.
  • Migrated developer tooling from hatch to native uv. Dependency groups (dev, test, doc) are now declared via PEP 735 [dependency-groups] in pyproject.toml.
  • CI rewritten to use uv sync --group test + uv run pytest against Python 3.12 and 3.13.
  • ReadTheDocs build updated to use uv sync --group doc + uv run sphinx-build.
  • How the masks for missing modalities are calculated: now assume that the modality is missing when all features are zero instead of the sum over all features being <= 0 previously (only makes a difference when the input can be negative, which doesn't happen with standard single-cell modalities)

Added

  • support for scvi-tools >= 1.0 (#37)
  • support for multiple NB lossed (#34)
  • tests