Skip to content

Releases: xRiskLab/pearsonify

v1.0.1

08 Jun 18:54
Compare
Choose a tag to compare

Release Notes for v1.0.1

✨ Major packaging modernization:

  • Migrated from Poetry to uv packaging system - following the same pattern as woeboost, fisher-scoring, and fastwoe
  • Restructured package layout - moved from src/pearsonify to root-level pearsonify/
  • Updated pyproject.toml to modern project specification format with SPDX license
  • Added comprehensive development dependencies - black, isort, ruff, pylint, pytest, pre-commit
  • Added uv.lock and .python-version files for reproducible environments

🔧 Technical improvements:

  • Cleaner build process with no deprecation warnings
  • Better tooling configuration (black, ruff, pytest)
  • TestPyPI publishing support configured

v1.0.0

17 Feb 17:26
Compare
Choose a tag to compare

Pearsonify 🐍 is a lightweight Python package that provides model-agnostic classification intervals for binary classification tasks. It uses Pearson residuals and principles from conformal prediction to quantify uncertainty without making strong distributional assumptions.