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.
- Compliance with scikit-learn Transformer API.
- Usage of scikit-learn's
set_output
API to output pandas.DataFrames (keep gene and observation names).
- Coarse-graining of Zero-Inflated genes is now performed as zero-or-not instead of quantile.
- All R dependencies: scBoolSeq is now implemented in pure Python, built on top of pandas and scikit-learn.
- Temporarily, the CLI has been removed as it needs major refactoring to accomodate the API's changes.
First version of scBoolSeq, buit as a wrapper around a series of R scripts.