Changelog

All notable changes to Real Simple Stats are documented here.

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

[0.2.0] - 2025-07-26

This is a major quality improvement release focusing on code standards, testing, and developer experience.

Added

  • Comprehensive Type Hints: Added detailed type annotations to all functions

  • Enhanced Documentation: Google-style docstrings with examples and mathematical explanations

  • Testing Infrastructure: 35 comprehensive unit tests with pytest framework

  • Command Line Interface: rss-calc CLI tool for quick statistical calculations

  • Development Tools:

    • Black code formatting

    • Flake8 linting

    • MyPy type checking

    • Pre-commit hooks

    • Makefile for development tasks

  • Quality Assurance:

    • Automated code formatting

    • Comprehensive error handling

    • Input validation with meaningful error messages

    • Coverage reporting

  • Documentation:

    • Professional Sphinx documentation with RTD theme

    • Installation guide

    • Quick start guide

    • CLI reference

    • Contributing guidelines

    • Code quality standards

    • Jupyter notebook tutorial

  • Package Improvements:

    • Enhanced pyproject.toml with tool configurations

    • Development dependencies and optional extras

    • Project metadata and classifiers

    • .gitignore for Python projects

Changed

  • Version: Bumped from 0.1.1 to 0.2.0

  • Code Quality: All code now follows Black formatting standards

  • Error Handling: Improved error messages and input validation

  • Documentation: Completely rewritten README with badges and comprehensive information

  • Package Structure: Better organization and modern Python packaging practices

Fixed

  • Coefficient of Variation: Fixed calculation bug for edge cases

  • Empty Input Handling: Proper error handling for empty datasets

  • Division by Zero: Added checks for zero denominators

  • Import Issues: Cleaned up unused imports and circular dependencies

  • Type Safety: Fixed type inconsistencies and added proper annotations

Removed

  • Unused Imports: Cleaned up all unused import statements

  • Dead Code: Removed commented-out and unreachable code

  • Redundant Functions: Consolidated duplicate functionality

Security

  • Input Validation: Added comprehensive input validation to prevent errors

  • Type Safety: Static type checking helps prevent runtime errors

[0.1.1] - 2024-XX-XX

Initial release with basic statistical functionality.

Added

  • Basic descriptive statistics (mean, median, mode, variance, standard deviation)

  • Probability utilities (simple, joint, conditional probability)

  • Hypothesis testing functions (t-tests, F-tests, critical values)

  • Probability distributions (normal, binomial, Poisson)

  • Linear regression utilities

  • Chi-square test functions

  • Confidence interval calculations

  • Basic plotting capabilities

  • Statistical glossary

  • Sphinx documentation setup

  • PyPI package distribution

[0.3.1] - 2025-01-XX

Fixed

Fixed several critical bugs that were causing crashes and incorrect results:

  • Resolved segmentation fault when running tests. The issue was caused by incompatible system-wide NumPy/SciPy installations. Package initialization now prefers the bundled virtualenv dependencies.

  • five_number_summary() was crashing on small datasets. It now handles n=1, n=2, and n=3 cases properly instead of trying to calculate quartiles from empty halves.

  • One-sided power analysis was ignoring the sign of effect sizes. Negative effects with “less-than” alternatives were returning the same power as positive effects. The calculations now correctly handle directionality.

Added

  • Normal distribution functions are now fully implemented. The CLI previously lacked normal distribution support; normal_pdf() and normal_cdf() are now available both in Python and via the command line.

  • CLI argument validation across all subcommands. Missing or invalid arguments now show helpful error messages instead of cryptic Python exceptions.

  • Input validation for probability helper functions. Functions like joint_probability(), bayes_theorem(), and expected_value() now check that probabilities are in valid ranges and raise clear errors.

  • stratified_split() now ensures minority classes get at least one test sample when possible, preventing rare classes from being completely excluded from test sets.

Changed

  • Package initialization was tweaked to avoid dependency conflicts. The __init__.py now sets up the environment before any imports happen.

  • Error messages are more helpful. Instead of cryptic Python exceptions, you’ll see clear messages like “Error: –n (number of trials) is required for binomial distribution”.

  • Added 10 new tests covering edge cases and validation (511 tests total, all passing).

[0.3.2] - 2026-03-21

Fixed

  • CLI tests were failing in CI because commands use logger.info() but logging was only configured in main(). Tests that call commands directly now configure logging with the captured stdout.

Changed

  • Removed emojis project-wide from docs, examples, and scripts

  • Interactive examples documentation now recommends Chart.js and Observable for web visualizations instead of Streamlit

[Unreleased]

(No unreleased changes)

Migration Guide

Upgrading from 0.1.1 to 0.2.0

Breaking Changes: None - this is a backward-compatible release.

Recommended Actions:

  1. Update your installation:

    pip install --upgrade real-simple-stats

  2. Try the new CLI:

    rss-calc --help

  3. Check the new documentation for enhanced examples and tutorials

  4. Consider using type hints in your code for better IDE support

Deprecated Features: None in this release.

New Opportunities:

  • Use the CLI for quick calculations

  • Leverage improved error messages for debugging

  • Benefit from comprehensive type hints in your IDE

  • Contribute to the project using our development tools

Version Support

Supported Versions:

  • 0.2.x: Active development, bug fixes, and new features

  • 0.1.x: Security fixes only (until 0.3.0 release)

Python Version Support:

  • Python 3.7+: Fully supported

  • Python 3.6: No longer supported (use version 0.1.x)

Dependency Updates:

  • NumPy: 1.19.0+ (was 1.18.0+)

  • SciPy: 1.5.0+ (was 1.4.0+)

  • Matplotlib: 3.3.0+ (was 3.1.0+)

Contributing to Changelog

When contributing to the project:

  1. Add entries to the [Unreleased] section

  2. Use the standard format: Added/Changed/Deprecated/Removed/Fixed/Security

  3. Be descriptive: Explain what changed and why

  4. Link to issues: Reference GitHub issues when applicable

  5. Credit contributors: Acknowledge community contributions

For more details, see our Contributing to Real Simple Stats guide.