Skip to content

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]

Fixed

  • CI "Static analysis (advisory)" job failed: mypy aborted with Source file found twice under different module names because synthetic_data/ had no __init__.py. Added the package marker, fixed the genuine type errors it uncovered (distribution parameters annotated as object, callable used as a type, unguarded n=None in utils/power_utils.py), relaxed the mypy strictness flags that do not fit the partially annotated codebase, and made the advisory steps continue-on-error so they report without failing the job.

[1.1.0] - 2026-09-23

Added

  • scipymasterpro command line (info, app, generate-data) and a scipymasterpro package holding the version; the previous scipy-app / scipy-generate-data entry points now work.
  • Streamlit smoke tests: every page is executed headlessly with streamlit.testing.v1.AppTest.
  • Plot smoke tests covering every helper in utils/viz_utils.py.
  • Hypothesis property tests (ECDF monotonicity, bootstrap bounds, effect-size antisymmetry, confidence-interval containment, interpolation at knots, SVD reconstruction).
  • Tests that execute every Python code block in the tutorials and API overview.
  • Notebook execution job in CI (nbmake) for all 12 notebooks.
  • Documentation site content: user guide, tutorials, developer guide, per-module API reference generated from docstrings; deployed to GitHub Pages by docs.yml.
  • CodeQL workflow, tag-driven release workflow (GitHub Release + optional PyPI trusted publishing).
  • Multi-stage, non-root Docker image with health check and an entrypoint that runs app, jupyter or both; multi-arch (amd64/arm64) publishing to GHCR.
  • requirements_dev.txt is now a fully resolved lock (make freeze-deps).
  • ruff linting; Makefile targets test-app, test-notebooks, test-all, lint-advisory, docker-test.

Changed

  • Minimum Python is 3.11; CI tests 3.11-3.13 with 3.14 as an experimental job.
  • Dependency floors raised (numpy>=1.26, pandas>=2.0, scipy>=1.11, streamlit>=1.50) and the Dependabot updates for pandas, matplotlib, plotly, jupyter-core, jupyterlab-server, pyarrow, cffi, executing, narwhals, ipykernel and wcwidth applied via the new lock.
  • GitHub Actions updated to current majors (checkout v7, setup-python v7, upload-artifact v7, codecov v7, build-push-action v7, metadata-action v6, CodeQL v4); Docker workflow uses docker compose v2.
  • CI lint job is now blocking (black, isort, ruff, bandit); mypy and pylint report as advisory.
  • Streamlit pages use a single, cross-platform PROJECT_ROOT bootstrap and package-qualified imports; use_container_width replaced by width="stretch".
  • Synthetic data generator writes next to its own file instead of relative to the working directory, and exposes main().
  • Coverage is no longer forced through pytest addopts; use make test-cov.
  • pre-commit configuration modernised (ruff, current hook versions, no types-all).

Fixed

  • perform_ks_test, fit_multiple_distributions* and the ECDF goodness-of-fit helpers failed on SciPy >= 1.18 (ndtr() takes from 1 to 2 positional arguments); they now test against the frozen distribution.
  • stratified_sample dropped the stratification column on pandas >= 3; it now uses GroupBy.sample.
  • Three Streamlit pages imported streamlit_utils unqualified and failed outside streamlit run.
  • Duplicate fit_distribution definition (the name-based variant is now fit_named_distribution).
  • Mid-module imports, unused imports and an unused walrus assignment in utils.
  • Docker workflow failed on current runners (docker-compose: command not found).

[1.0.0] - 2026-03-09

Added

  • Initial release of ScipyMasterPro
  • 10 comprehensive Jupyter notebooks covering all SciPy domains:
  • 01: Descriptive Statistics (moments, ECDF, skewness, kurtosis)
  • 02: Hypothesis Tests (t-tests, Mann-Whitney U, normality tests)
  • 03: Distribution Fitting (MLE, KS tests, multiple distributions)
  • 04: Sampling & Resampling (stratified, weighted, multinomial)
  • 05: Bootstrap & Simulation (confidence intervals, resampling)
  • 06: Multivariate Analysis (PCA, Mahalanobis distance, chi²)
  • 07: Optimization & Minimization (constraints, cost functions)
  • 08: Linear Algebra & Statistics (SVD, eigendecomposition, least squares)
  • 09: Interpolation & Curve Fitting (splines, curve_fit)
  • 10: Inference from Raw Data (SEM, confidence intervals)
  • 2 shared notebooks for comparative analysis:
  • PDF/ECDF comparison with statsmodels
  • Statistical power analysis
  • Interactive Streamlit application with 12 pages
  • Clean UI with custom CSS styling
  • Dataset selector and filtering
  • Interactive visualizations with Plotly
  • Export functionality for plots and tables
  • 10 modular utility modules:
  • stats_tests_utils.py - Statistical hypothesis testing
  • distribution_utils.py - Distribution fitting and evaluation
  • viz_utils.py - Statistical visualization functions
  • optimization_utils.py - Optimization algorithms and constraints
  • linear_algebra_utils.py - Matrix operations and decomposition
  • interpolation_utils.py - Interpolation and curve fitting
  • pdf_ecdf_utils.py - PDF and ECDF visualization tools
  • inference_utils.py - Statistical inference calculations
  • sim_utils.py - Simulation and bootstrap utilities
  • power_utils.py - Statistical power analysis
  • Synthetic data generator with 6+ reproducible datasets:
  • Normal and skewed distributions
  • Mixed distributions (beta, gamma, exponential, lognormal, Poisson)
  • Multivariate Gaussian with covariance
  • Optimization test data (sinusoidal + noise)
  • Curve fitting data (exponential decay + noise)
  • Categorical counts for chi-square tests
  • Poisson discrete data
  • Docker support for containerized deployment
  • Python 3.11-slim base image
  • Both JupyterLab and Streamlit in one container
  • Optimized system and Python dependencies
  • Comprehensive documentation:
  • Detailed README with setup instructions
  • CONTRIBUTING.md with contributor guidelines
  • CODE_OF_CONDUCT.md for community standards
  • SciPy cheatsheet for quick reference
  • Proper .gitignore covering Python, Jupyter, and OS files
  • GPL-3.0 license for open-source distribution
  • Consistent export structure for all plots and tables

Project Structure

ScipyMasterPro/
├── notebooks/               # 10 core concept notebooks
├── shared_notebooks/        # 2 comparison notebooks
├── streamlit_app/           # Interactive web application
│   └── pages/              # 12 Streamlit pages
├── synthetic_data/          # Data generation scripts and outputs
├── utils/                   # 10 reusable utility modules
├── cheatsheets/            # SciPy syntax reference
├── exports/                # Generated plots and tables
│   ├── plots/
│   └── tables/
├── requirements.txt        # Production dependencies
├── requirements_dev.txt    # Development dependencies
├── Dockerfile             # Container configuration
├── README.md              # Project documentation
├── CONTRIBUTING.md        # Contribution guidelines
├── CODE_OF_CONDUCT.md     # Community standards
└── LICENSE               # GPL-3.0 license

Features

  • ✅ Pure SciPy implementation (minimal statsmodels dependency)
  • ✅ Reproducible synthetic datasets (seed=42)
  • ✅ Modular, reusable utility functions
  • ✅ Consistent notebook structure across all modules
  • ✅ Professional Streamlit UI with custom styling
  • ✅ Automated plot and table exports
  • ✅ Docker-ready for easy deployment
  • ✅ Portfolio and interview prep ready

Dependencies

  • Python ≥ 3.10
  • NumPy ≥ 1.23
  • Pandas ≥ 1.5
  • SciPy ≥ 1.10
  • Matplotlib ≥ 3.6
  • Seaborn ≥ 0.12
  • Plotly ≥ 5.18
  • Statsmodels ≥ 0.14
  • Streamlit ≥ 1.33
  • JupyterLab ≥ 3.6
  • scikit-learn ≥ 1.3

Release Notes

Version 1.0.0 - Initial Release

This is the first official release of ScipyMasterPro, a comprehensive toolkit for mastering the SciPy library through hands-on practice with synthetic datasets, interactive notebooks, and a professional web application.

Target Audience: - Data scientists learning SciPy - Students studying statistical methods - Professionals preparing for technical interviews - Researchers needing quick SciPy reference materials

Key Highlights: - Complete coverage of scipy.stats, scipy.optimize, scipy.interpolate, and scipy.linalg - Clean, synthetic datasets for focused learning without domain noise - Interactive exploration via both Jupyter and Streamlit - Production-ready utilities for real-world statistical analysis - Fully containerized with Docker for reproducible environments

What's Next: - See CONTRIBUTING.md to get involved - Check out the notebooks in sequential order for structured learning - Experiment with the Streamlit app for interactive exploration - Use utilities in your own projects by importing from the utils/ folder


Full Changelog: https://github.com/SatvikPraveen/ScipyMasterPro/commits/v1.0.0