Changelog

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.

Unreleased#

0.2.0#

Added#

  • ParTI-style quantile-based distance-ranking enrichment methods are now part of partipy, see for example pt.compute_quantile_based_gene_enrichment.

0.1.0#

Added#

  • Added optimizer aliases ("PCHA" for "projected_gradients" and "FW" for "frank_wolfe") across the public API, including caching filters and documentation, plus regression tests to ensure both names yield identical results.

  • Added quantile-based continuous enrichment for gene expression and numeric adata.obs columns with ParTI-style binning, max-in-bin0 criteria, and optional NaN ignoring for obs columns.

  • Added quantile-based categorical enrichment for adata.obs labels with ParTI-style binning, max-in-bin0 filtering, hypergeometric over-representation testing, configurable background contrast, and minimum category count filtering.

0.0.6#

Added#

  • partipy.write_h5ad and partipy.read_h5ad helpers that make the AnnData.uns caches HDF5-compatible by serializing and restoring ArchetypeConfig keys.

  • Automatic restoration of cached dictionaries keyed by ArchetypeConfig when using the public accessors, enabling use of .h5ad files saved with the helper utilities.

0.0.5#

Added#

  • Public accessor layer for cached archetypal analysis artifacts (get_aa_result, get_aa_cell_weights, get_aa_metrics, get_aa_bootstrap, summarize_aa_metrics) with consistent filtering semantics.

  • Comprehensive documentation on caching and retrieval flows, including the new docs/notebooks/data_access.ipynb tutorial and updates to other notebooks.

  • New bootstrap and selection-metric plotting enhancements that rely on the unified accessors.

Changed#

  • Reworked AA caching to remove the eagerly stored adata.uns['AA_metrics_df'], generating summaries on demand instead using summarize_aa_metrics

  • Refactored t-ratio significance testing and AA result handling to better reuse cached runs and ensure typing/mypy compliance.

  • Updated plotting APIs (plot_var_explained, plot_IC, plot_bootstrap_*, plot_archetypes_*) to require precomputed caches and use the new result filters.

  • Streamlined schema defaults and test fixtures after the accessor refactor.

  • Multiple unit-test adjustments to align with the new caching workflow.

0.0.4#

First release alpha version of partipy

Contents