0.3

Full Changelog: 0.2...0.3

Release Notes: ellalgo

📅 Version: 0.2 → 0.3

Date: September 12, 2025
Contributor: luk036

📦 Overview

This release includes significant enhancements to the ellalgo project, a Python implementation of the Ellipsoid Method for convex optimization. The updates focus on improved testing, documentation, code quality, and new features.

✨ New Features & Improvements

1. Added Comprehensive Documentation

• A new markdown file GEMINI.md was introduced, providing:
  • Project overview and build instructions
  • Dependency management using tox and pytest
  • Code style enforcement via black, isort, and flake8
  • Documentation build process with Sphinx

2. Enhanced Testing Suite

• Added multiple unit tests using Gemini CLI
• Expanded test coverage for:
  • LMI0Oracle
  • LowpassOracle
  • ProfitOracle
  • ConjugateGradient
  • CuttingPlane algorithms
  • Ellipsoid update methods

3. Code Quality & Refactoring

• Applied ruff format and ruff check --fix across the codebase
• Improved type annotations and docstrings
• Fixed zero-gradient handling in ellipsoid updates
• Enhanced error messages and validation

4. New Oracle Implementations

• Added LMI0Oracle for linear matrix inequality feasibility
• Enhanced LowpassOracle with better frequency response handling
• Improved ProfitOracle for Cobb-Douglas production optimization

5. Algorithm Optimizations

• Updated stopping criteria and tolerance handling
• Improved binary search logic in bsearch
• Better handling of parallel cuts in ellipsoid updates

🐛 Bug Fixes

• Fixed assertion in lowpass benchmark test (result >= 12300)
• Corrected environment build commands in orgx-build-commands.txt
• Fixed gradient validation in ellipsoid update methods
• Improved handling of edge cases in LDLT factorization

📊 Statistics

• 27 files changed
• 993 additions and 451 deletions
• 9 commits included in this release

🔧 Development Conventions

• Pre-commit hooks now enforce code style (black, isort, flake8)
• All code must pass pre-commit run --all-files before committing
• Documentation is built automatically via tox -e docs

🚀 Getting Started

To use this version:

# Install dependencies
conda env create --file environment.yml --prefix ./venv

# Run tests
tox

# Build documentation
tox -e docs

# Format code
pre-commit run --all-files

📄 Note

This release maintains backward compatibility while significantly improving test coverage, documentation, and code quality. The ellipsoid method implementation remains stable with enhanced numerical robustness.

Fixed Bugs in Parallel-Central Cut

Full Changelog: 0.1...0.2

0.1

Full Changelog: https://github.com/luk036/ellalgo/commits/0.1