Changelog for ts-stat-tests🔗

v1.0.1

v1.0.1 - Refresh documentation structure, add new pages, update package infrastructure and enhance repo security🔗

v1.0.1
2026-02-01
data-science-extensions/ts-stat-tests/releases/v1.0.1

Release Notes

This release marks a significant milestone in the ts-stat-tests library, transitioning the project to Production/Stable status. The focus of this update is a comprehensive overhaul of the documentation architecture, the integration of mathematical rendering capabilities, and the hardening of the repository's security and CI/CD workflows.

🚀 Key Highlights🔗

Core Infrastructure & Security🔗

Production Status: Elevated the development_status to "5 - Production/Stable" in pyproject.toml, reflecting the library's maturity and stability.
CodeQL Analysis: Introduced a new CodeQL workflow to perform advanced static analysis for both Python and GitHub Actions, ensuring high code quality and security standards.
Automated Dependency Management: Configured Dependabot to monitor and update pip and GitHub Actions dependencies weekly.
Metadata Refinement: Updated project classifiers to better represent the library's focus on Mathematical Engineering and Scientific Research.

Documentation Overhaul🔗

Restructured Navigation: Organised the documentation into a clearer "Basic" vs "Advanced" hierarchy in mkdocs.yml to improve the onboarding experience for new users while maintaining deep references for experts.
New "Getting Started" Guide: Added a Getting Started page that explains the library's tiered API structure (Dispatchers, Boolean Checks, and Algorithms) and provides usage examples for data loaders.
Exhaustive Algorithm Catalog: Created the Algorithms Guide, documenting the rationale for library selections and providing a detailed catalog of all implemented statistical tests.
Comprehensive FAQ: Launched a FAQ section to address common conceptual questions, such as the handling of the Null Hypothesis ($H_0$) and the benefits of the library's unified interface.

Mathematical & Visual Enhancements🔗

KaTeX Integration: Integrated KaTeX for browser-side rendering of mathematical expressions, ensuring that statistical hypotheses and formulas are displayed precisely throughout the documentation.
Visual Iconography: Enhanced all primary documentation headers with relevant emoji icons (e.g., 💡, 📚, 🚀) to facilitate quicker visual scanning and brand consistency.

ADF Parameter Unpacking: Fixed an issue in the documentation for the adf() algorithm where return values were incorrectly unpacked when store=True.
Linearity Options: Standardised the naming of extra: parameters across the linearity module's tests and algorithms.
Spelling Standardisation: Synchronised all documentation and comments to adhere to Australian English standards (e.g., prioritising -ise over -ize).

📊 Change Metrics🔗

Metric	Value
Files Changed	13
Lines Added	+672
Lines Deleted	-201
Project Status	Production/Stable
Spelling Standard	Australian English

🛠️ Internal Changes🔗

Update .gitignore to exclude AI instruction sets and local output directories.
Bump uv-pre-commit version to 0.9.28.
Refine docstring formats to comply with strict DFC (Docstring Format Checker) rules.

💪 Pull Requests🔗

Refresh documentation structure, add new pages, update package infrastructure and enhance repo security by @chrimaho in https://github.com/data-science-extensions/ts-stat-tests/pull/34

Full Changelog: https://github.com/data-science-extensions/ts-stat-tests/compare/v1.0.0...v1.0.1

Updates

f796a00: Fix argument unpacking in ADF function call
- Remove the usedlag and nobs arguments from the unpacking of the adf() function.
- Ensure the call to adf() remains functional with the updated argument list. (by chrimaho)
b2fbf08: Properly indent steps: in codeql workflow
- Standardise indentation and formatting in the .github/workflows/codeql.yml file.
- Ensure consistent use of comments to describe each step in the workflow.
- Maintain clarity in the Checkout repository and Initialize CodeQL steps.
- Improve readability of the Run manual build steps section with clearer comments. (by chrimaho)
466984f: Fix yaml syntax error
The matrix include: list items are not indented under include: (the - language: ... entries are aligned with include:). This is invalid YAML and will prevent the workflow from loading. Indent the list so include: maps to a sequence. (by chrimaho)
59b2173: Update JavaScript assets in MkDocs configuration
- Replace the assets/js/katex.js entry with the CDN link for KaTeX.
- Add the https://unpkg.com/katex@0/dist/katex.min.js and https://unpkg.com/katex@0/dist/contrib/auto-render.min.js entries.
- Ensure proper loading of KaTeX for rendering mathematical expressions. (by chrimaho)
7611234: Fix typo
numpy is imported in the example but never used. Dropping the unused import will keep the quickstart snippet minimal and avoids lint warnings for readers who copy/paste it. (by chrimaho)
19c9583: Fix typos
get_sine_wave() is shown being called with n/periods arguments, but the actual function signature takes no parameters. As written, this example will raise TypeError. Either update the example to call get_sine_wave() as implemented, or extend the API to support the documented parameters. (by chrimaho)
1817f46: Fix typo
The snippets include syntax is inconsistent with the other pages (e.g., --8<-- "README.md") and may not be parsed correctly without quotes. Use the quoted form for the LICENSE include to match the existing working pattern. (by chrimaho)
7ad2618: Fix typo
Grammar: “libraries which implements these tests” should be “libraries which implement these tests”. (by chrimaho)
24c85c7: Correct import paths
In the “Python Import” table, the RUR and ZA rows label the import source as pmdarima, but the imports shown are from statsmodels. This is misleading when readers try to map tests to upstream libraries—please correct the library labels for those rows. (by chrimaho)
0d7e3d7: Refresh docs comment
The sample ValueError message here doesn't match what the library actually raises. generate_error_message() produces messages like Invalid 'algorithm': <value>. Options: <options>, not the multi-line Invalid value for parameter... Parsed... format shown. Updating the example to match the real output will prevent confusion when troubleshooting. (by chrimaho)
2c0a6d5: Reword docs
This section states that the project “strictly forbid[s] the use of Any”, but the current codebase does use Any (e.g., ts_stat_tests.stationarity.algorithms.adf assigns res: Any = _adfuller(...)). Either relax/reword this requirement (e.g., “avoid Any where possible”) or align the implementation to match the stated policy. (by chrimaho)
6b8ee18: Update getting started documentation for clarity and detail
- Enhance descriptions of data loading functions in the ts_stat_tests.utils.data module.
- Include links for load_airline(), load_macrodata(), get_sine_wave(), get_noise_data(), and get_random_generator().
- Revise the available test categories table to include links for each category and its corresponding algorithms.
- Improve formatting for better readability and navigation within the documentation. (by chrimaho)
0743f40: Add FAQ section to documentation
- Introduce a new FAQ section in the documentation for ts-stat-tests.
- Explain the differences between "Dispatcher", "Boolean", and "Algorithm" functions.
- Describe the advantages of using ts-stat-tests over direct calls to underlying packages.
- Clarify the handling of the Null Hypothesis ($H_0$) in statistical tests.
- Provide troubleshooting tips for common errors and usage questions.
- Include installation details and rationale for dependencies like numpy. (by chrimaho)
9de5928: Update documentation for algorithms and getting started pages
- Enhance section headings with relevant emojis for better visual appeal.
- Change heading for selection rationale to 💡 Selection Rationale.
- Change heading for exhaustive test catalog to 📚 Exhaustive Test Catalog.
- Change heading for advanced usage patterns to 🚀 Advanced Usage Patterns.
- Change heading for developer information to 🛠️ Developer Information.
- Change heading for available test categories to 📚 Available Test Categories.
- Ensure consistent use of Australian English spelling, changing prioritized to prioritised. (by chrimaho)
12a3976: Add KaTeX support for rendering mathematical expressions
- Introduce the katex.js file to handle rendering of mathematical expressions.
- Subscribe to document$ to render math in the document body.
- Define delimiters for inline and block math rendering.
- Update mkdocs.yml to include the new katex.js file in extra JavaScript.
- Add KaTeX CSS and JavaScript files for proper rendering. (by chrimaho)
ef0174b: Update documentation for algorithms and tests
- Add detailed analysis of algorithms in algorithms.md.
- Include rationale for library selection and coverage in the documentation.
- Introduce an exhaustive test catalog with implementation status and library references.
- Enhance the index.md file by consolidating content and improving clarity.
- Remove outdated sections and streamline the overall structure of the documentation. (by chrimaho)
5f40a50: Add algorithms page to docs
- Create a new algorithms.md file for algorithm usage documentation.
- Update mkdocs.yml to include the new Algorithms section in the navigation. (by chrimaho)
5515331: Fix typo in options for linearity modules
- Correct extras: to extra: in the options for ts_stat_tests.linearity.tests.
- Correct extras: to extra: in the options for ts_stat_tests.linearity.algorithms. (by chrimaho)
952a9ad: Update module links in documentation
- Change links for modules in the documentation to include .md file extensions.
- Ensure consistency in linking to module descriptions for better navigation. (by chrimaho)
bc32692: Add Getting Started documentation for ts-stat-tests
- Introduce the Getting Started section to provide an overview of the library.
- Outline installation instructions using uv and pip.
- Explain core concepts and structure of the library's modules.
- Provide a quickstart example for performing a stationarity test.
- List utility data loaders available in the ts_stat_tests.utils.data module.
- Detail available test categories and their corresponding dispatchers and checks. (by chrimaho)
cb50c1b: Add FAQ section to documentation
- Create a new faq.md file to address common questions.
- Update mkdocs.yml to include the FAQ section in the navigation. (by chrimaho)
42d5b60: Add license documentation to project
- Create a new file license.md to include licensing information.
- Update mkdocs.yml to add a navigation entry for the License section. (by chrimaho)
ae98c84: Refactor navigation structure in documentation
- Organise the navigation menu in mkdocs.yml for better clarity.
- Group related items under Basic and General categories.
- Ensure all links remain functional after restructuring. (by chrimaho)
ee6f791: Add Getting Started section to documentation
- Create a new getting_started.md file for user onboarding.
- Include the Getting Started entry in the navigation of mkdocs.yml. (by chrimaho)
eeaa0d0: Add CodeQL workflow for advanced code analysis
- Introduce the .github/workflows/codeql.yml file for CodeQL analysis.
- Configure triggers for push and pull_request events on the main branch.
- Schedule weekly analysis using a cron job.
- Define job matrix for multiple languages including actions and python.
- Set up permissions for security events and package access.
- Include steps for checking out the repository and initializing CodeQL tools.
- Add a manual build step for languages requiring custom build commands.
- Implement the Perform CodeQL Analysis step for code scanning. (by chrimaho)
58e78c9: Update dependency versions in pyproject.toml
- Adjust black version for compatibility with Python versions.
- Specify conditional dependencies for ipykernel, isort, and pylint based on Python version.
- Update docstring-inheritance and mkdocstrings versions for Python compatibility.
- Modify pytest version to align with Python version requirements.
- Change complexipy version to 5.*.
- Update uv_build version to 0.9.28 in the build system. (by chrimaho)
71cb5a1: Update project classifiers in pyproject.toml
- Change the development status to "5 - Production/Stable".
- Add new topics for utilities, scientific engineering, and information analysis.
- Update intended audience to "Science/Research".
- Include operating systems support for Unix, MacOS, and Windows. (by chrimaho)
9e1fe97: Update .gitignore to include additional files and directories
- Add git_output/ directory to be ignored.
- Include .github/copilot-instructions.md file in the ignore list.
- Add .github/instructions/* and .github/prompts/* directories to be ignored. (by chrimaho)
398acb6: Update uv-pre-commit dependency version
- Update the rev for the uv-pre-commit repository to 0.9.28 (by chrimaho)
aee5492: Update Dependabot configuration for Python dependencies
- Add configuration for updating Python dependencies using Dependabot.
- Set target branch to main and schedule updates weekly.
- Include labels, reviewers, and assignees for the updates.
- Specify commit message prefix as "🤖 deps(pip): ". (by chrimaho)
51ec99c: Update Dependabot configuration for GitHub Actions
- Add target branch for GitHub Actions updates.
- Include assignees and commit message prefix for clarity. (by chrimaho)
86d76b1: Update pre-commit configuration for improved linting and checks
- Upgrade black to version 26.1.0
- Upgrade uv-pre-commit to version 0.9.26
- Add docstring-format-checker with entry dfc --check
- Modify ty entry to run without specifying filenames (by chrimaho)

v1.0.0

v1.0.0 - The First Major Release: A Unified Foundation for Time Series Analysis🔗

v1.0.0
2026-01-18
data-science-extensions/ts-stat-tests/releases/v1.0.0

Release Notes

🚀 Overview🔗

We are proud to announce the first major release of ts-stat-tests. This milestone represents the culmination of an intensive development effort to create a unified, production-ready suite for time-series statistical testing in Python.

The primary goal of this release is to provide a single, consistent interface to the most important statistical tests from established libraries like statsmodels, scipy, arch, and antropy. By normalising inputs, outputs, and hypothesis testing logic, ts-stat-tests simplifies complex analysis workflows and ensures mathematical rigour across all modules.

🆕 Getting Started🔗

Install the library using your preferred package manager:

pip install ts-stat-tests

Or with uv:

uv add ts-stat-tests

Basic Usage🔗

Every module provides a consistent API. You can check for a property (Boolean) or get full results (Dispatcher).

from ts_stat_tests.stationarity import is_stationary, adf

# Check for stationarity at 5% significance
result = is_stationary(data, algorithm="adf", alpha=0.05)
if result["result"]:
    print(f"Series is stationary (p-value: {result['pvalue']:.4f})")

# Or get the raw statistics
stat, pvalue, lags, crit, ic = adf(data, autolag="AIC")

🧩 Module Synopses🔗

ts-stat-tests is organised into eight core namespaces, each dedicated to a specific domain of time-series analysis.

📊 Stationarity🔗

Tests to determine if the statistical properties of a series are constant over time. Includes Augmented Dickey-Fuller (adf()), KPSS (kpss()), and Zivot-Andrews (za()) for structural breaks.

📈 Normality🔗

Assess whether the residuals or data follow a normal distribution. Supports Jarque-Bera (jb()), Shapiro-Wilk (sw()), and D'Agostino's K-squared (dp()).

📏 Linearity🔗

Evaluate if a linear model is appropriate for the data. Includes the Ramsey RESET test (rr()), Harvey-Collier (hc()), and Rainbow test (rb()).

📉 Heteroscedasticity🔗

Detect whether the variance of errors is constant. Provides ARCH (arch()), Breusch-Pagan (bpl()), Goldfeld-Quandt (gq()), and White's test (wlm()).

♻️ Correlation🔗

Measure autocorrelation and cross-correlation. Features ACF (acf()), PACF (pacf()), and the Ljung-Box test (lb()) for residual independence.

🔄 Seasonality🔗

Identify seasonal patterns and measure their strength. Includes the QS test (qs()), OCSB (ocsb()), and STL-based features like seasonal_strength().

🏗️ Stability🔗

Analyse the stability and lumpiness of time-series data using STL decomposition features, helping identify periods of structural change or volatility shifts.

🕸️ Regularity🔗

Measure the complexity and entropy of a signal. Provides multiple entropy metrics including Approximate (approx_entropy()), Sample (sample_entropy()), and Permutation Entropy (permutation_entropy()).

🏛️ Architectural Refactor🔗

A core highlight of the v1.0.0 release is a complete re-architecting of the internal package structure (PR #32). We have migrated from a functional layout to a domain-centric layout.

Namespace Consolidation: Algorithms and tests are now co-located within their respective domain folders (e.g., src/ts_stat_tests/stationarity/).
Intuitive Imports: The new structure allows for cleaner, more descriptive import statements. Users can now import directly from the top-level namespace of each module.

Old Import Pattern:

from ts_stat_tests.algorithms.stationarity import kpss

New Import Pattern:

from ts_stat_tests.stationarity import kpss

💪 Quality Assurance🔗

Reliability is the cornerstone of this library. The v1.0.0 release maintains the highest standards of code quality:

100% Code Coverage: Every execution path in the library is verified by unit tests.
Strict Type Safety: Full Pyright strict-mode compliance and runtime @typechecked enforcement.
Optimal Linting: 10.00/10 Pylint score across all files.
Verified Documentation: All examples are tested via doctest to ensure they are runnable and accurate.

✅ Checklist🔗

Implement final Heteroscedasticity module.
Restructure package into domain-centric namespaces.
Standardise import patterns via __init__.py files.
Achieve 100% code coverage across the entire library.
Overhaul documentation and contribution guidelines.
Remove all skipped doctests and ensure complete reproducibility.

📊 Changes🔗

Metric	Value
Files Changed	46
Lines Added	+1,317
Lines Deleted	-490
Commit Count	16
Contributors	1

🔗 Pull Requests🔗

Prepare for First Major Release by @chrimaho in https://github.com/data-science-extensions/ts-stat-tests/pull/32

Full Changelog: https://github.com/data-science-extensions/ts-stat-tests/compare/v0.8.0...v1.0.0

Updates

38ae241: Fix mkdocs config structure
The mkdocs.yml configuration has a structural issue. The options: key should not be nested under extras:. Looking at the diff, the entire options: block (lines 132-153) should be at the same indentation level as paths:, not nested inside an extras: key. This is likely to cause the mkdocs build to fail or not apply the configuration correctly. (by chrimaho)
19eb899: Fix typo (by chrimaho)
9dac0b5: Fix typo (by chrimaho)
3ecd1af: Add headers and summaries to each namespace init file
- Add titles and purposes for the Heteroscedasticity, Linearity, Normality, Regularity, Seasonality, Stability, and Stationarity modules.
- Include detailed summaries for each module to describe their functionalities and the tools they provide for time series analysis. (by chrimaho)
4e40b4a: Fix typo (by chrimaho)
47f3dfb: Fix broken links in the Code overview page
- Update code style adherence to [black][] formatting.
- Ensure spell checking for documentation and code comments using [codespell][].
- Maintain type safety through type-checking with [ty][] and [pyright][].
- Manage import sorting with [isort][] and remove unused imports using [pycln][].
- Check code quality with [pylint][] maintaining a score of 10/10.
- Verify docstring quality using [docstring-format-checker][].
- Achieve 100% code coverage with unit testing using [pytest][].
- Build package with [uv][] and documentation with [mkdocs][]. (by chrimaho)
84f6591: Remove redundant checks (by chrimaho)
62a5908: Fix typos and formatting issues in testing documentation
- Correct spelling errors in the testing section.
- Ensure consistent formatting for library references.
- Improve clarity of the testing description. (by chrimaho)
24e728b: Revise README for clarity and structure
- Change section title from "Motivation" to "Overview"
- Introduce a new "Key Features" section with detailed descriptions
- Add installation instructions and quick start examples
- Include a "Quality Assurance" section outlining testing and code quality practices
- Update "Known limitations" to "Quality Assurance" for better organisation (by chrimaho)
53344e3: Update contribution guidelines for clarity and completeness
- Revise the contribution process to include detailed steps for raising issues, forking the repository, and creating branches.
- Add sections on testing changes and submitting pull requests.
- Ensure guidelines align with project goals and enhance contributor experience. (by chrimaho)
f640f36: Add title to changelog output
- Include a title in the changelog file for better clarity.
- Write the title as # Changelog for {repo.name}. (by chrimaho)
7b5ec65: Restructure docs pages for consistency
- Remove redundant section headers for tests in various documentation files.
- Introduce a unified "Modules" section across all relevant documentation files.
- Enhance clarity and organisation of the documentation structure. (by chrimaho)
7c9309a: Completely restructure package modules
- Old structure:
- ts_stat_tests/algorithms/<ALGO>
- ts_stat_tests/tests/<ALGO>
- New structure:
- ts_stat_tests/<ALGO>/algorithms
- ts_stat_tests/<ALGO>/tests
- This allows for more streamlined package structure, and a more intuitive import process by intelligent usage of __init__.py files per namespace (by chrimaho)
6a72832: Update docs to remove progress updates
- Remove implementation progress table from index.md
- Update formatting in README.md for better readability
- Ensure consistent header structure across documentation (by chrimaho)
216efc9: Remove any references for # doctest: +SKIP
- Replace numpy array generation with utility functions for consistency.
- Update p-values in examples to reflect new calculations.
- Ensure all examples use data_line and data_random for data generation. (by chrimaho)

v0.8.0

v0.8.0 - Add Heteroscedasticity Module🔗

v0.8.0
2026-01-18
data-science-extensions/ts-stat-tests/releases/v0.8.0

Release Notes

🚀 Overview🔗

Introduce the heteroscedasticity module, completing the seventh analytical pillar of the ts-stat-tests library. This release provides a unified interface for detecting Autoregressive Conditional Heteroscedasticity (ARCH) and classic heteroscedasticity using the Breusch-Pagan, Goldfeld-Quandt, and White tests. These tools are critical for identifying volatility clustering and non-constant variance in residuals, ensuring the validity of econometric inferences. This version also represents a significant technical milestone, transitioning the entire test suite to modern pytest assertions and standardising documentation across all core algorithms.

🛠️ Implementation details🔗

Heteroscedasticity module implementation🔗

Implement core heteroscedasticity algorithms in src/ts_stat_tests/algorithms/heteroscedasticity.py:
- arch(): Engle's test for Autoregressive Conditional Heteroscedasticity.
- bpl(): Breusch-Pagan Lagrange Multiplier test for heteroscedasticity.
- gq(): Goldfeld-Quandt test for homoskedasticity.
- wlm(): White's Lagrange Multiplier test for heteroscedasticity and specification error.
Provide standardised Boolean dispatchers in src/ts_stat_tests/tests/heteroscedasticity.py:
- heteroscedasticity(): Unified high-level interface for all heteroscedasticity tests.
- is_heteroscedastic(): Boolean checker returning a consistent "result" key, maintaining API parity across the library.
Migrate from deprecated statsmodels.sandbox imports to the stable statsmodels.stats.diagnostic module for long-term compatibility.

Infrastructure and documentation🔗

Create comprehensive theoretical documentation in docs/code/heteroscedasticity.md, including KaTeX mathematical definitions for variance testing.
Update README.md and docs/code/index.md to reflect 100% completion status for both the Linearity and Heteroscedasticity modules.
Extend the library's code overview with detailed feature highlights including unified interfaces, boolean checkers, and strict type safety.
Optimise navigation in mkdocs.yml to include the new diagnostic pages.

Quality assurance and refactoring🔗

Refactor the global test suite to transition from unittest-style methods to modern pytest assertions, improving diagnostic readability.
Modernise internal QA infrastructure by categorising checks in the check() function within src/utils/scripts.py.
Standardise type hints across the library, preferring native float over np.float64 for simplified user consumption.
Maintain 10-00/10 Pylint scores and 100% test coverage through rigorous docstring and type validation.

✅ Checklist🔗

Implement arch(), bpl(), gq(), and wlm() algorithms.
Implement heteroscedasticity() and is_heteroscedastic() dispatchers.
Refactor entire test suite to use standard pytest assertions.
Modernise statsmodels imports to remove deprecated sandbox references.
Author theoretical documentation for heteroscedasticity diagnostics.
Pass 100% path coverage and Pylint 10.00/10 quality gates.

📊 Changes🔗

Metric	Value
Files Changed	11
Lines Added	1047
Lines Deleted	22
Contributors	@chrimaho
Commit Count	22

🆕 Pull Requests🔗

Implement Heteroscedasticity Module by @chrimaho in https://github.com/data-science-extensions/ts-stat-tests/pull/31

*Full Changelog: https://github.com/data-science-extensions/ts-stat-tests/compare/v0.7.0...v0.8.0

Updates

782046a: Fix docstrings in the gq() function
The return type annotation in the docstring describes np.float64 values, but the actual code returns float values (line 389, 403). The docstring should be updated to reflect the actual return type of float instead of np.float64 to be consistent with the implementation. (by chrimaho)
ff8800e: Fix docstrings in the bpl() function
The return type annotation in the docstring describes np.float64 values, but the actual code returns float values (line 261). The docstring should be updated to reflect the actual return type of float instead of np.float64 to be consistent with the implementation. (by chrimaho)
3c436af: Fix docstrings in the arch() function
The return type annotation in the docstring describes np.float64 values, but the actual code returns float values (lines 177-180, 188). The docstring should be updated to reflect the actual return type of float instead of np.float64 to be consistent with the implementation. (by chrimaho)
3a4fd45: Fix another typo (by chrimaho)
4b4af91: Fix typo (by chrimaho)
6cfb3c6: Fix docstrings for the wlm() function
The return type annotation in the docstring describes np.float64 values, but the actual code returns float values (line 466). The docstring should be updated to reflect the actual return type of float instead of np.float64 to be consistent with the implementation. (by chrimaho)
260477e: Update README to reflect heteroscedasticity tests status
- Mark Engle's Test for Autoregressive Conditional Heteroscedasticity (ARCH) as ✅
- Mark Breusch-Pagan Lagrange Multiplier test for heteroscedasticity (BPL) as ✅
- Mark Goldfeld-Quandt test for homoskedasticity (GQ) as ✅
- Mark White's Lagrange Multiplier Test for Heteroscedasticity (WLM) as ✅ (by chrimaho)
50a0b90: Remove deprecated sandbox module
- Replace import from statsmodels.sandbox.stats.diagnostic with statsmodels.stats.diagnostic
- Ensure consistent usage of ResultsStore across modules (by chrimaho)
0f106e4: Extend Code overview page to add key info
- Introduce key features of the statistical tests package, including:
- Unified interface for consistent user experience.
- Boolean checkers for null hypothesis evaluation.
- Comprehensive documentation for user understanding.
- Robust testing suite achieving 100% code coverage.
- Standardised error handling for informative feedback.
- Strict type safety with @typechecked decorators.
- Modular design for extensibility of new tests.
- Integration with established libraries for validated methodologies.
- Quality assurance with stringent code quality checks.
- Document modules available for various statistical tests. (by chrimaho)
e8350b7: Refactor unit tests to use pytest assertions
- Replace self.assertEqual() with assert for brevity.
- Update type checks to use assert isinstance() for better readability.
- Standardise algorithm result checks across tests. (by chrimaho)
3f88884: Refactor params & returns type hints in heteroscedasticity module
- Standardise type hints to use float instead of np.float64
- Update return types to reflect consistent tuple structures
- Simplify return statements in arch(), bpl(), gq(), and wlm() functions
- Remove unnecessary imports and unused code (by chrimaho)
66ea028: Fix typo (by chrimaho)
d851214: Refactor check() function for improved structure and clarity
- Organise checks into categories: Formatting, Spelling, Type Safety, Imports, Quality, Docs, and Unit Tests
- Remove redundant calls to check_codespell() and check_pyright()
- Ensure all checks are clearly defined and grouped for better readability (by chrimaho)
4f586ae: Add documentation for heteroscedasticity modules
- Introduce an overview of heteroscedasticity and its significance in statistics.
- Detail the four heteroscedasticity tests available in the library, including their import scripts and links to documentation.
- Provide information on the source library statsmodels and its role in the implementation.
- List the source modules for the tests and algorithms. (by chrimaho)
5b75189: Fix typo (by chrimaho)
db10837: Add unit tests for heteroscedasticity module
- Implement setUp() method to prepare test data for homoscedastic and heteroscedastic cases
- Create tests for arch(), bpl(), gq(), and wlm() algorithms
- Add dispatcher tests for heteroscedasticity() function
- Include tests for is_heteroscedastic() function with various algorithms (by chrimaho)
7fc683b: Add extensive docstrings to the heteroscedasticity module
- Add comprehensive docstrings for arch(), bpl(), gq(), and wlm() functions.
- Include usage examples and parameter descriptions for better clarity.
- Implement a convenience wrapper function heteroscedasticity() for easier testing.
- Introduce is_heteroscedastic() function to evaluate residuals against significance levels. (by chrimaho)
0098430: Add module headers and descriptions for heteroscedasticity module
- Include a summary of the module's purpose and functionality.
- Document the various heteroscedasticity tests implemented:
- ARCH Test
- Breusch-Pagan Test
- Goldfeld-Quandt Test
- White's Test (by chrimaho)
e66b8b7: Add heteroscedasticity algorithms and tests
- Implement arch(), bpl(), gq(), and wlm() functions for heteroscedasticity tests.
- Create heteroscedasticity() function to select and call appropriate tests based on algorithm.
- Introduce is_heteroscedastic() function to evaluate test results and return structured output.
- Update imports to include new functions and necessary modules. (by chrimaho)
dcfedb4: Implement initial structure for heteroscedasticity module
- Add arch(), bpl(), gq(), and wlm() functions as placeholders
- Create initial test structure with heteroscedasticity() and is_heteroscedastic() functions (by chrimaho)
9a0347e: Add framework for heteroscedasticity module
- Introduce heteroscedasticity.py algorithm implementation
- Create test_heteroscedasticity.py for unit tests
- Update mkdocs.yml to include Heteroscedasticity documentation
- Add heteroscedasticity.md documentation file (by chrimaho)

v0.7.0

v0.7.0 - Add Linearity Module🔗

v0.7.0
2026-01-17
data-science-extensions/ts-stat-tests/releases/v0.7.0

Release Notes

🚀 Overview🔗

Introduce the linearity module, the sixth major analytical pillar of the ts-stat-tests library. This release provides a suite of econometric tests designed to validate the linear specification of regression models, including the Harvey-Collier, Lagrange Multiplier, Rainbow, and Ramsey RESET tests. These tools enable researchers to identify neglected non-linearity and functional form misspecifications with high precision. Alongside the new algorithms, this version reaches a new quality milestone with 228 passing tests and 100% code coverage across all modules.

🛠️ Implementation details🔗

Linearity module implementation🔗

Implement core linearity algorithms in src/ts_stat_tests/algorithms/linearity.py:
- hc(): Harvey-Collier test based on recursive residuals.
- lm(): Lagrange Multiplier test for checking functional form.
- rb(): Rainbow test for linearity using subset comparisons.
- rr(): Ramsey RESET test for neglected non-linearity.
Provide standardised Boolean dispatchers in src/ts_stat_tests/tests/linearity.py:
- linearity(): Unified high-level interface for all linearity tests.
- is_linear(): Boolean checker returning a consistent "result" key, standardised against the rest of the library’s API.
Integrate strict type-checking via @typechecked and handle statsmodels signature inconsistencies using internal type-narrrowing helpers.

Create a comprehensive theoretical guide in docs/code/linearity.md, featuring LaTeX definitions for all implemented tests.
Update README.md and docs/code/index.md status tables to mark the Linearity module as fully supported.
Optimise navigation in mkdocs.yml to include the new diagnostic tools.
Standardise documentation cross-references to ensure seamless transitions between algorithmic theory and code examples.

Quality assurance and testing🔗

Expand the package test suite to 228 passing tests, incorporating 11 new comprehensive unit tests for the Linearity module.
Maintain 100% path coverage across the entire ts_stat_tests codebase.
Enforce 10.00/10 Pylint scores and 0 Pyright errors for all new implementations.

✅ Checklist🔗

Implement hc(), lm(), rb(), and rr() algorithms.
Implement linearity() and is_linear() dispatchers.
Standardise is_linear() return keys for API consistency.
Author theoretical documentation for linearity diagnostics.
Update project metadata and status tables in README.md.
Pass 100% coverage and linting quality gates.

📊 Changes🔗

Metric	Value
Files Changed	7
Lines Added	799
Lines Deleted	8
Contributors	1 (@chrimaho)
Commit Count	11

🆕 Pull Requests🔗

Implement Linearity Module by @chrimaho in https://github.com/data-science-extensions/ts-stat-tests/pull/30

Full Changelog: https://github.com/data-science-extensions/ts-stat-tests/compare/v0.6.2...v0.7.0

Updates

b78c7dc: Refactor is_linear() function to return keys consistent with the rest of the package
- Change return key from "is_linear" to "result" in is_linear() function.
- Update related test assertions in test_linearity.py to reflect the new key.
- Ensure documentation examples are consistent with the updated return structure. (by chrimaho)
a16b017: Update progress
- Mark linearity tests as implemented in the README.md
- Update test status for Harvey Collier, Lagrange Multiplier, Rainbow, and Ramsey's RESET tests to ✅
- Reflect changes in the index.md documentation for linearity algorithms (by chrimaho)
63f7847: Refactor options structure in linearity documentation
- Move options under extras for clarity
- Ensure consistent formatting across sections (by chrimaho)
457d175: Add documentation for linearity modules
- Introduce an overview of linearity testing and its significance
- Detail the implementation and source library used
- Provide examples of source modules related to linearity tests and algorithms (by chrimaho)
1bef1a7: Add unit tests for linearity algorithms
- Implement TestLinearity class for testing various linearity algorithms.
- Include tests for Harvey-Collier, Lagrange Multiplier, Rainbow, and Ramsey RESET tests.
- Add dispatcher tests for algorithm selection in linearity checks.
- Validate results for linear and non-linear datasets. (by chrimaho)
4aae6e3: Refactor linearity algorithms for improved robustness and clarity.
- Remove unnecessary type hints for variables.
- Replace direct attribute access with getattr() for safer attribute retrieval.
- Ensure consistent handling of return values from linearity functions. (by chrimaho)
ce23d16: Add linearity tests and convenience functions
- Implement linearity() function to perform linearity tests on fitted regression models.
- Introduce is_linear() function to check if the relationship in a fitted model is linear.
- Include detailed docstrings for both functions with usage examples.
- Add necessary imports and define module exports. (by chrimaho)
87f9b3b: Add docstrings for linearity algorithms
- Include summaries, parameters, return types, and examples for the hc(), lm(), rb(), and rr() functions.
- Enhance documentation clarity for users implementing linearity tests. (by chrimaho)
03dfb72: Add linearity algorithms
- Introduce hc(), lm(), rb(), and rr() functions for linearity tests.
- Define constants VALID_RR_TEST_TYPE_OPTIONS and VALID_RR_COV_TYPE_OPTIONS.
- Include necessary imports from statsmodels and numpy. (by chrimaho)
e9eb401: Add framework for linearity modules
- Introduce linearity.py algorithm implementation.
- Create test_linearity.py for unit tests.
- Add linearity.md documentation file.
- Update mkdocs.yml to include Linearity in navigation. (by chrimaho)

v0.6.2

v0.6.2 - Add Stability Module🔗

v0.6.2
2026-01-17
data-science-extensions/ts-stat-tests/releases/v0.6.2

Release Notes

🚀 Overview🔗

Introduce the stability module, the fifth major analytical component of the ts-stat-tests library, providing advanced variance-based metrics for time-series consistency. This release focuses on "Stability" and "Lumpiness"—critical indicators for data quality and forecasting feasibility. Alongside these algorithmic additions, this version implements significant documentation infrastructure improvements, including multi-column layouts for enhanced readability and a comprehensive audit of internal cross-referencing to ensure seamless navigation across the API.

🛠️ Implementation details🔗

Stability module implementation🔗

Implement two core stability algorithms in src/ts_stat_tests/algorithms/stability.py:
- stability(): Measures the variance of the trend component within a time series, providing a quantified metric for structural consistency.
- lumpiness(): Calculates the variance of the variance across tiling windows, identifying intermittent "lumpy" patterns in the data.
Provide standardised Boolean dispatchers in src/ts_stat_tests/tests/stability.py:
- is_stable(): Evaluates if a series maintains trend stability below a specified significance threshold.
- is_lumpy(): Determines if a series exhibit excessive variance variability (lumpiness).

Documentation architecture and UX🔗

Introduce multi-column CSS layouts to the documentation site to optimise information density and improve the presentation of API references.
Standardise and audit all internal documentation links, ensuring consistent navigation between algorithm implementations and their corresponding test dispatchers.
Simplify docstring filtering logic to resolve cross-referencing ambiguities and ensure "Summary" and "Details" sections render correctly under strict dfc compliance.
Update README.md and docs/code/index.md implementation tables to reflect the 100% completion of the Stability module and the expanded test suite.

Quality assurance and testing🔗

Expand the global unit test suite to 146 tests, maintaining 100% code coverage across all modules.
Audit and correct threshold logic in Boolean tests to ensure standard p-value interpretation (rejecting the null hypothesis when statistics are below alpha).
Standardise all internal variables and public signatures to use NDArray[np.float64] for improved type specificity and compatibility with pyright strict mode.

✅ Checklist🔗

Implement core stability and lumpiness algorithms.
Provide standardised Boolean convenience functions for stability assessments.
Achieve 100% coverage for the new stability module.
Implement multi-column layout support in the documentation site via custom CSS.
Complete comprehensive audit and fix of internal docstring cross-references.
Update project progress metrics to reflect 146 total tests across 6 modules.

📊 Changes🔗

Metric	Value
Files Changed	19
Lines Added	639
Lines Deleted	69
Commit Count	20
Contributors	@chrimaho

💪 Pull requests🔗

Implement stability module by @chrimaho in https://github.com/data-science-extensions/ts-stat-tests/pull/29

Full Changelog: https://github.com/data-science-extensions/ts-stat-tests/compare/v0.5.3...v0.6.2

Updates

8861517: Update purpose description for stability tests
- Change purpose from "Unit tests for stability and lumpiness algorithms." to "Convenience functions for stability algorithms." (by chrimaho)
2b76084: Fix docs descriptions
The documentation states that stability limits are "between 0 and 1", but this is incorrect. According to the implementation and examples (e.g., airline data has stability of 13428.67), stability can have unbounded positive values as it measures variance. This documentation is misleading and should be corrected to reflect that stability is a non-negative value without an upper bound. (by chrimaho)
8ff12e1: Fix typo
Grammatical error: "this modules" should be "these modules" to match plural usage. (by chrimaho)
74c8ccc: Add missing docs justifications
The placeholder text "REASONS" should be replaced with an actual explanation of why the tsfeatures package was chosen. This appears to be incomplete documentation left from a template. (by chrimaho)
a132892: Fix module description
The module docstring description is inaccurate. This module contains convenience functions and boolean checkers (is_stable, is_lumpy), not unit tests. It should describe the module's purpose similar to other test modules, such as "This module contains convenience functions and tests for stability measures, allowing for easy access to stability and lumpiness algorithms." (by chrimaho)
61e84c5: Update progress for all docs
- Update stability tests to reflect accurate results.
- Modify README to include stability in statistical tests.
- Enhance documentation for clarity on stability algorithms. (by chrimaho)
50e3e30: Restructure navigation pages in docs
- Change 'Modules' section to be nested under 'Code'
- Improve organisation of documentation for better clarity (by chrimaho)
45ee3fc: Add CSS for multi-column layouts and button styles
- Introduce styles for two, three, and four-column layouts
- Implement mobile responsiveness for column layouts
- Define button widths and alignment for various button types (by chrimaho)
43e8ce3: Standardise links in documentation for modules
- Update links in correlation.md, normality.md, regularity.md, seasonality.md, stability.md, and stationarity.md to use reference-style links.
- Modify stability.py and tests/stability.py to standardise link formatting for stability() and lumpiness() functions. (by chrimaho)
87eb94b: Update algorithm titles for consistency across modules
- Change titles to include 'Algorithms' for clarity in correlation.py, normality.py, regularity.py, seasonality.py, and stationarity.py.
- Update purpose descriptions for better understanding in seasonality.py test file. (by chrimaho)
0e73702: Update module docstrings for stability and tests
- Enhance docstrings in the stability module to include a summary.
- Add detailed descriptions for the stability algorithms.
- Improve docstrings in the stability tests module to clarify their purpose.
- Include a summary note in both modules for better documentation clarity. (by chrimaho)
c1dfa07: Remove unnecessary filters from docs pages
- Eliminate filters for is and specific test names
- Simplify documentation structure for clarity (by chrimaho)
5711842: Fix stability docs formatting config
- Update source module links to remove redundant prefixes
- Adjust options formatting for stability tests and algorithms (by chrimaho)
f6e516f: Refactor stability and lumpiness functions for type consistency
- Update type hints in stability() and lumpiness() functions to use NDArray[np.float64].
- Modify is_stable() and is_lumpy() functions to accept NDArray[np.float64] as input type.
- Adjust expected result in test_stability_2() to reflect correct stability assessment. (by chrimaho)
de3a1f0: Add docstrings to the stability module
- Separate imports for lumpiness and stability for clarity.
- Add detailed docstrings for stability() and lumpiness() functions.
- Implement is_stable() and is_lumpy() functions with comprehensive examples and explanations.
- Ensure consistency in parameter descriptions and return values. (by chrimaho)
7347910: Add stability documentation
- Introduce a new document for testing the stability of time-series datasets.
- Include an introduction explaining the concept of stability and its importance.
- Describe the tests for stability and lumpiness, including their significance and scoring.
- Provide references to source libraries and modules related to stability. (by chrimaho)
ff5ebd8: Add stability unittests
- Implement TestStability class to test stability and lumpiness
- Set up class method to initialise test data and results
- Add tests for stability and lumpiness calculations
- Validate results against expected outputs (by chrimaho)
5261964: Implement stability and lumpiness algorithms
- Add stability() function to calculate stability from input data.
- Add lumpiness() function to calculate lumpiness from input data.
- Introduce type checking using @typechecked decorator for both functions.
- Update imports to include necessary libraries and algorithms. (by chrimaho)
294a38d: Add framework for stability module
- Create stability.py algorithm file for stability tests
- Add test_stability.py for unit tests related to stability
- Include stability.md documentation for stability algorithms
- Update mkdocs.yml to include stability in navigation (by chrimaho)

v0.5.3

v0.5.3 - Add Seasonality Module🔗

v0.5.3
2026-01-16
data-science-extensions/ts-stat-tests/releases/v0.5.3

Release Notes

🚀 Overview🔗

Introduce the seasonality module, the fourth major component of the ts-stat-tests library, while simultaneously achieving a significant architectural milestone: the total elimination of typing.cast() and the attainment of 100% global code coverage. This release transforms the library into a strictly typed, fully validated suite for time-series analysis. By integrating six new seasonality algorithms, standardising on NDArray type hints, and implementing MathJax for professional mathematical rendering, this version provides the most robust and accessible interface for statistical testing to date.

🛠️ Implementation details🔗

Seasonality module implementation🔗

Implement six core seasonality algorithms in src/ts_stat_tests/algorithms/seasonality.py:
- qs(): The Ljung-Box test for seasonality.
- ocsb(): The Osborn-Chui-Smith-Birchenhall test for seasonal differencing.
- ch(): The Canova-Hansen test for seasonal stability.
- seasonal_strength() and trend_strength(): Measures based on STL decomposition.
- spikiness(): A measure of variance in the remainder component.
Provide a unified dispatcher in src/ts_stat_tests/tests/seasonality.py:
- seasonality(): Central entry point for all seasonality-related measurements.
- is_seasonal(): Standardised Boolean checker returning result dictionaries with interpreted statistics and alpha thresholds.

Type safety and architectural hardening🔗

Remove all usages of typing.cast() throughout the codebase, replacing them with safer type-narrowing patterns and intermediate Any assignments to satisfy strict pyright checks without sacrificing runtime safety.
Standardise internal data structures to use NDArray[np.float64] for improved type specificity and interoperability with numpy ecosystem.
Refactor normality and stationarity modules to use protocol-based result handling for third-party library outputs (e.g., scipy and statsmodels).
Implement output rounding across all statistical tests to ensure consistent return values and stable documentation examples.

Documentation and visual enhancements🔗

Integrate MathJax into the mkdocs pipeline to enable high-quality rendering of mathematical equations in the documentation site.
Introduce custom CSS-based implementation progress bars to the documentation landing pages.
Standardise all docstrings to follow the strict Google-style format with full dfc (Docstring Format Checker) compliance.
Update README.md and docs/index.md status tables to reflect the completed seasonality module.

✅ Checklist🔗

Implement core seasonality algorithms and unified dispatcher.
Achieve 100% code coverage across the entire ts_stat_tests package.
Remove all typing.cast() calls to satisfy strict Pylint 10/10 and Pyright requirements.
Integrate MathJax for KaTeX math rendering in the documentation.
Standardise all return types and implement numerical rounding for consistency.
Update data utility functions with improved type safety and expanded synthetic generators.

📊 Changes🔗

Metric	Value
Files Changed	34
Lines Added	2645
Lines Deleted	532
Commit Count	86
Contributors	Chris Mahoney

💪 Pull requests🔗

Implement seasonality module by @chrimaho in https://github.com/data-science-extensions/ts-stat-tests/pull/28

Full Changelog: https://github.com/data-science-extensions/ts-stat-tests/compare/v0.4.5...v0.5.3

Updates

e54021f: Refactor return types and handling in normality and seasonality tests
- Update return type of normality() function to support variable-length tuples.
- Simplify handling of results in is_normal() function for better clarity.
- Enhance return structure in seasonality() function to accommodate ARIMA objects.
- Improve type handling in internal helper functions for robustness. (by chrimaho)
699a777: Refactor return type and structure of is_stationary() function
- Update return type hint to use STATIONARITY_ITEM for clarity.
- Explicitly define return dictionary to match the new type hint.
- Adjust handling of pvalue to return None when applicable. (by chrimaho)
dd5445a: Add tests for seasonality and normality algorithms
- Implement fallback tests for Anderson-Darling in is_normal() function.
- Add coverage for list-based result branches in is_seasonal() function.
- Test fallback for non-numeric types in is_seasonal() function.
- Test fallback for non-numeric scalar types in is_seasonal() function. (by chrimaho)
7b6daf2: Enhance pytest configuration and update check functions.
- Add --doctest-modules and --doctest-continue-on-failure options to pytest in pyproject.toml.
- Modify check_pytest() function to dynamically include Python files for testing.
- Comment out check_doctest() in the check() function. (by chrimaho)
0c1da06: Refactor type hints in stationarity module
- Remove unnecessary Optional from return type of is_stationary() function.
- Simplify type hints for stat and pvalue variables to Any. (by chrimaho)
8bf8b36: Refactor type hints to remove cast() usage
- Remove cast() from normality mofule.
- Remove cast() from seasonality module.
- Remove cast() from stationarity module. (by chrimaho)
74bf5f8: Remove cast() from stationarity module
- Remove unnecessary cast() in the is_stationary() function.
- Simplify the return type assignment to directly use the result of stationarity() function. (by chrimaho)
aa51207: Enhance normality test return types and type hints.
- Update return type of normality() to include NormaltestResult, ShapiroResult, and AndersonResult.
- Replace cast() with direct type annotations for improved clarity.
- Adjust variable types in is_normal() for consistency and readability. (by chrimaho)
49203f7: Refactor type hints to remove cast() in stationarity and seasonality modules
- Update type hints to use Any for better flexibility.
- Remove unnecessary cast() calls to simplify code.
- Ensure consistent handling of return types across functions. (by chrimaho)
2ccd1bd: Round numerical outputs in the adf() function for improved readability.
- Update output formatting to display rounded values.
- Enhance clarity of example output in documentation. (by chrimaho)
e6fe3af: Round numerical outputs in entropy and regularity functions for improved readability.
- Update example outputs in entropy() function to show rounded values.
- Update example outputs in regularity() function to show rounded values. (by chrimaho)
d9ac474: Refactor numpy type hints to use NDArray instead of np.ndarray
- Update type hints in seasonality.py to use NDArray[np.float64]
- Modify return types in correlation.py to use NDArray[np.float64]
- Adjust type casting in normality.py to reflect new type hints (by chrimaho)
b2ffd98: Refactor data types in test setup for consistency
- Change data types of test data attributes to use NDArray[np.float64]
- Update cls.data_random, cls.data_sine, cls.data_line, cls.data_noise, and cls.data_2d attributes (by chrimaho)
9a75913: Fix import issue for VarianceRatio in the stationarity module
- Remove explicit type annotation for res in vr() function.
- Simplify assignment to res by directly using _vr(). (by chrimaho)
bff495e: Refactor return types in data utility functions for improved type safety
- Change return type of get_random_numbers() to NDArray[np.float64]
- Change return type of get_random_numbers_2d() to NDArray[np.float64]
- Change return type of get_sine_wave() to NDArray[np.float64]
- Change return type of get_normal_curve() to NDArray[np.float64]
- Change return type of get_straight_line() to NDArray[np.float64]
- Change return type of get_trend_data() to NDArray[np.float64]
- Change return type of get_uniform_data() to NDArray[np.float64]
- Change return type of get_noise_data() to NDArray[np.float64]
- Update variable types in data loading to reflect new return types (by chrimaho)
13d0c18: Refactor type handling in stationarity functions
- Introduce type aliases for improved clarity and maintainability.
- Update function signatures to use new type aliases.
- Simplify return statements in is_stationary() function. (by chrimaho)
0cec5c4: Refactor imports in stationarity.py for consistency and clarity.
- Standardise import statements for VarianceRatio.
- Group related imports together for better readability. (by chrimaho)
a15a06c: Add examples to adf() function docs
- Include example usage for storing results with adf(x=airline, store=True)
- Add example for using adf() without autolag: adf(x=airline, autolag=None, maxlag=5) (by chrimaho)
1d300e9: Refactor return statements in stationarity functions for clarity and efficiency.
- Simplify return statements in kpss(), rur(), and za() functions by removing unnecessary cast().
- Enhance readability by directly returning results from _kpss(), _rur(), and _za().
- Maintain type safety while improving code clarity in the vr() function. (by chrimaho)
feb0ace: Refactor adf() function for improved type handling and return structure
- Change return type handling to use Any for compatibility with statsmodels stubs
- Simplify return statements based on store and autolag parameters
- Ensure consistent float and integer conversions for returned values (by chrimaho)
5479ac7: Add tests for ADF stationarity function
- Introduce test_stationarity_adf_store() to validate ADF with store=True
- Introduce test_stationarity_adf_autolag_none() to validate ADF with autolag=None (by chrimaho)
aaba18f: Refactor type annotations in the regularity module
- Update type annotations in regularity.py and tests/regularity.py to use NDArray where appropriate.
- Simplify return statements in approx_entropy(), sample_entropy(), permutation_entropy(), spectral_entropy(), and svd_entropy() functions.
- Ensure all functions return consistent types, enhancing type safety and readability. (by chrimaho)
8f90a8c: Refactor normality module for improved type annotations in all functions
- Remove unnecessary class definitions and simplify return types.
- Update function signatures to use specific return types like ShapiroResult and NormaltestResult.
- Streamline the implementation of statistical tests by returning results directly from their respective functions. (by chrimaho)
f8d4c7e: Fix typo (by chrimaho)
d9f458d: Refactor example outputs in qs() function
- Adjust precision of output values in the docstring.
- Ensure clarity in the representation of ARIMA model output. (by chrimaho)
340f63c: Fix typo (by chrimaho)
74fec0b: Refactor seasonality tests for clarity and consistency.
- Rename test methods for better readability.
- Remove debug print statements.
- Update assertions to ensure accuracy.
- Improve documentation for test cases. (by chrimaho)
006a7b2: Refactor correlation tests and add new data tests
- Remove unused imports and redundant tests from test_correlation.py
- Create test_data.py with tests for load_airline() and load_macrodata()
- Enhance error handling tests in test_errors.py (by chrimaho)
ece6d42: Refactor covariance type handling in lm() function
- Standardise cov_type parameter to use VALID_LM_COV_TYPE_OPTIONS
- Remove redundant overloads for lm() function
- Update docstring for clarity on cov_type options (by chrimaho)
46fdb01: Refactor type hints for correlation algorithms
- Update type hints to use NDArray[np.float64] for better type clarity.
- Modify return types in acf(), pacf(), ccf(), lb(), lm(), and bglm() functions.
- Standardise covariance type options in VALID_LM_COV_TYPE_OPTIONS. (by chrimaho)
41ddcd0: Update seasonality tests and documentation for improved readability
- Modify doctest_optionflags in pyproject.toml to include "ELLIPSIS"
- Update expected output in qs() function examples for clarity
- Shorten numerical outputs in qs(), seasonal_strength(), trend_strength(), and spikiness() functions
- Adjust expected output in seasonality() test cases for consistency (by chrimaho)
0138dd4: Enhance equation formatting in normality algorithms
- Improve readability of equations in the Jarque-Bera, D'Agostino's K², Shapiro-Wilk, and Anderson-Darling tests.
- Update equations to use consistent formatting with $$ for better display. (by chrimaho)
4219a53: Update pytest command in check_doctest() functions to include doctest options.
- Add --doctest-modules and --doctest-continue-on-failure flags to enhance doctest execution. (by chrimaho)
5e776f6: Fix type checking for load_macrodata() function
- Introduce TypeCheckError for more specific error handling.
- Update error assertions to check for expected messages.
- Enhance test coverage for type errors in data loading. (by chrimaho)
a8a4ec3: Add print_label functionality for improved checking output readability
- Introduce print_label() function to format command output labels.
- Enhance linting functions with descriptive labels for better user feedback.
- Update command usage messages for clarity and consistency. (by chrimaho)
382bafe: Add tests for is_correlated() function with various algorithms
- Implement tests for Ljung-Box, LM, and Breusch-Godfrey algorithms.
- Validate results and types for correlation checks.
- Ensure proper error handling for unsupported algorithms.
- Test correlation logic for both correlated and non-correlated data.
- Refactor failure assertions in assert_almost_equal() method tests. (by chrimaho)
a1fa2eb: Refactor error message assertions in test_generate_error_message() method.
- Update assertions to check for specific error message components.
- Ensure clarity in error reporting for invalid parameters. (by chrimaho)
4fcee28: Refactor error message formatting in assert_almost_equal() function.
- Update error message to use clearer variable names.
- Change tol: p={places}, d={delta} to ({places=}, {delta=}) for consistency. (by chrimaho)
7ade057: Add ELLIPSIS option to doctest flags
- Include # "ELLIPSIS" in the doctest_optionflags section of pyproject.toml
- Enhance flexibility for doctest output matching (by chrimaho)
0a86247: Improve error message for invalid algorithm parameter in stationarity() function.
- Clarify the options available for the algorithm parameter in the error message.
- Enhance user experience by providing detailed feedback on valid options. (by chrimaho)
47a14a5: Update example outputs in seasonality tests
- Update expected output for qs() function to reflect precise values:
  qs(data, freq=12) now returns (194.46928920877465, 5.9092232580147965e-43)
- Update expected output for seasonality() function using seasonal strength:
  seasonality(x=data, algorithm="ss", m=12) now returns 0.7787219427520644 (by chrimaho)
3fc4004: Fix typo (by chrimaho)
e3d5993: Add parameter and return type documentation for stationarity() function
- Document parameters and return types for clarity.
- Include examples for better understanding of usage. (by chrimaho)
8cfe9df: Add parameter and return type documentation for seasonality() function
- Document parameters func and args with types.
- Specify return type as a union of possible result types.
- Include example usage in the docstring. (by chrimaho)
8a4cd9d: Add example to AndersonResult protocol documentation
- Include an example of how to use the AndersonResult protocol.
- Enhance clarity for users implementing the protocol.
- Provide a code snippet demonstrating usage:
```python
def my_func(res: AndersonResult) -> None:
print(res.statistic) (by chrimaho)
73fa21d: Refactor is_correlated() function to enhance correlation testing
- Update function signature to accept parameters for correlation algorithm and significance level.
- Implement logic for various correlation tests including Ljung-Box, LM, and Breusch-Godfrey.
- Raise ValueError for unsupported algorithms.
- Add detailed docstring with examples and parameter descriptions. (by chrimaho)
e52debc: Enhance docs structure for all correlation functions
- Add summary and details for the acf(), pacf(), ccf(), and lb() functions.
- Update references formatting for clarity.
- Improve readability by adding line breaks in the documentation. (by chrimaho)
5a183f7: Refactor load_macrodata() function to improve data loading
- Update data types for year and quarter columns to int.
- Simplify data loading process using pd.read_csv().
- Adjust index creation to use a combined string format for quarterly periods. (by chrimaho)
a740fd6: Update doctest examples in data utilities
- Move the data = ... section from the Example chunk to the Setup chunk for each function in the utils/data module (by chrimaho)
bfd3af1: Add doctest config to pyproject.toml
- Include --doctest-modules and --doctest-continue-on-failure options in pytest configuration.
- Update test command in check_doctest() and check_doctest_module() functions to use --config-file=pyproject.toml.
- Standardise data type outputs in docstrings from np.float64 to float64 for consistency. (by chrimaho)
a5bb4e8: Update error message formatting in assert_almost_equal() function
- Modify the traceback output for clarity in error reporting.
- Ensure the assertion error message is displayed correctly. (by chrimaho)
2a3ec92: Add __init__.py module to the utils namespace (by chrimaho)
d03c7f0: Refactor data utility documentation for clarity and consistency.
- Standardise docstring formatting across functions.
- Add detailed examples and references for better understanding.
- Introduce type checking for improved type safety.
- Enhance descriptions to clarify functionality and usage. (by chrimaho)
0754d8b: Refactor error utility documentation for clarity and consistency.
- Update docstrings to provide clearer summaries and details.
- Enhance examples for better understanding of usage.
- Standardise formatting across all functions. (by chrimaho)
ab48087: Add utils docs pages to MkDocs navigation
- Create data.md file for data utilities documentation
- Create errors.md file for error utilities documentation
- Update mkdocs.yml to include new documentation in navigation (by chrimaho)
0050943: Modify 'calculation' of docstrings section to be optional
- Change 'required' attribute of 'calculation' from true to false
- Ensure flexibility in documentation requirements (by chrimaho)
26174f2: Refactor table formatting for clarity in implementation progress
- Adjust column headers for better alignment
- Enhance readability of the implementation progress table (by chrimaho)
d4c3fcb: Update README.md for improved link formatting and clarity
- Correct link formatting for Statistical Tests references
- Add direct links for [Normality], [Stationarity], and [Correlation] (by chrimaho)
d080631: Add navigation link to Data Science Extensions page
- Include a link to Data Science Extensions in the navigation menu. (by chrimaho)
77204b4: Add additional code features to MkDocs configuration
- Include content.code.select feature for enhanced code selection
- Include content.code.copy feature for improved code copying
- Ensure content.code.annotate feature is available for code annotations (by chrimaho)
1f29834: Add MathJax integration for enhanced mathematical rendering
- Introduce mathjax.js for MathJax configuration
- Subscribe to document changes to clear cache and typeset
- Update mkdocs.yml to include MathJax scripts (by chrimaho)
f0c979c: Add progress bar styles and integrate into MkDocs configuration
- Create progress_bar.css with styles for progress bars and labels
- Include progress bar styles in mkdocs.yml under extra_css (by chrimaho)
27b17bd: Move CSS styles from stylesheets sub-dir to the css sub-dir
- Introduce styles for various admonition types including pro-tip, deprecation, observation, and calculation
- Implement styles for code chunk filenames with appropriate icons
- Update mkdocs.yml to reference new CSS file paths (by chrimaho)
4df11e7: Update README and Code docs pages
- Enhance README structure and badge visibility
- Update seasonality test results in the implementation progress table
- Correct overall test statistics for seasonality (by chrimaho)
9d43aa2: Refactor pytest imports for clarity
- Remove unused fixture and mark imports from pytest
- Simplify import statements for better readability (by chrimaho)
155879f: Refactor assertions and enhance seasonality tests
- Update assert_almost_equal() calls to use places parameter for precision.
- Add detailed test cases for seasonality() and is_seasonal() functions.
- Include summaries for test_seasonality() and test_is_seasonal() methods.
- Improve test coverage for seasonal algorithms: QS, OCSB, CH, seasonal strength, trend strength, and spikiness. (by chrimaho)
c6de989: Update seasonality documentation for clarity and accuracy
- Correct quote reference to Rob Hyndman and George Athanasopoulos.
- Refine definitions of trend, seasonal, and cyclic patterns for better understanding.
- Enhance information on source libraries and modules used in the implementation.
- Improve overall structure and readability of the documentation. (by chrimaho)
5631240: Fill out the seasonality() and is_seasonal() functions
- Introduce seasonality() function as a dispatcher for various seasonality algorithms.
- Enhance is_seasonal() function to provide a standardized output.
- Update docstrings for better understanding and usage examples.
- Implement type checking for function parameters. (by chrimaho)
c6a8d15: Enhance documentation requirements for sections
- Require 'params', 'examples', and 'calculation' sections to be mandatory.
- Add 'warning' section with optional requirement. (by chrimaho)
a289db6: Refactor seasonality documentation for clarity and consistency
- Improve descriptions and examples for the qs(), ocsb(), ch(), seasonal_strength(), and spikiness() functions.
- Standardise parameter descriptions and default values.
- Enhance mathematical equations and their explanations.
- Update example usage for better clarity and accuracy. (by chrimaho)
c6afda0: Refactor seasonality documentation paths for consistency
- Update paths in the seasonality tests section to remove the src. prefix
- Adjust paths in the seasonality algorithms section for uniformity (by chrimaho)
b716147: Refactor spikiness function to handle NaN values
- Replace np.mean() with np.nanmean() for robust calculation of mean
- Ensure proper handling of missing values in seasonal decomposition (by chrimaho)
739cc65: Refactor type hints and input handling for seasonality functions
- Change input type from np.ndarray to ArrayLike for improved flexibility.
- Update input conversion to np.asarray(x, dtype=float) for consistent handling of input data. (by chrimaho)
91dfc47: Specify exceptions in ARIMA model fitting error handling
- Replace generic exception handling with specific exceptions: ValueError, RuntimeError, and IndexError
- Improve error handling clarity and robustness in the qs() function (by chrimaho)
4f6a078: Refactor type hints across unit tests for improved clarity and consistency
- Update type hints in setup.py to use object instead of Any
- Modify return type of data_dict() function for better type specificity
- Ensure consistency in type annotations across various functions (by chrimaho)
c0f5e6f: Add comprehensive documentation pages for seasonality analysis
- Introduce an overview of seasonality in time series datasets
- Define key concepts such as trend, seasonal, and cyclic patterns
- Provide references for further reading and source libraries
- Include links to relevant source modules for clarity (by chrimaho)
6c9e2e2: Add additional pylint disable rules for improved code quality
- Include "R0903" for too-few-public-methods
- Add "R0912" for too-many-branches
- Add "R1719" for simplifiable-if-expression (by chrimaho)
cb9529f: Implement unit tests for seasonality algorithms
- Add TestSeasonality class to test various seasonality functions
- Include tests for qs(), ocsb(), ch(), seasonal_strength(), trend_strength(), and spikiness()
- Validate expected outputs and handle failure cases for qs()
- Ensure proper setup with setUpClass() method (by chrimaho)
5d3b1a1: Refactor stationarity module header documentation
- Remove unnecessary comments and clean up the docstring.
- Enhance the clarity of the purpose and usage of stationarity tests.
- Include a relevant article link for further reading on ADF & KPSS tests. (by chrimaho)
9bbc9cb: Enhance docs for seasonality module
- Introduce the qs() function for the Ljung-Box test to detect seasonality in time series data.
- Implement the ocsb() function for the Osborn, Chui, Smith, and Birchenhall test for seasonal differencing.
- Add the ch() function for the Canova-Hansen test to assess seasonal stability.
- Include the seasonal_strength() function to measure the strength of seasonality in time series data.
- Implement the trend_strength() function to evaluate the strength of the trend component.
- Add the spikiness() function to detect spikes or volatility in time series data.
- Enhance docstrings with detailed descriptions, parameters, return values, and examples for each function. (by chrimaho)
cde3717: Add detailed headers and descriptions for seasonality modules
- Include a summary of seasonality tests in the seasonality.py module.
- Document the purpose and functionality of each implemented algorithm.
- Enhance clarity for users regarding the statistical tests available. (by chrimaho)
ebfeac7: Add AndersonResult protocol and improve type casting
- Introduce AndersonResult protocol for better type safety.
- Enhance type casting in ad() function to use cast(AndersonResult, ...).
- Import Protocol from typing and numpy for type annotations. (by chrimaho)
341686b: Implement seasonality algorithms and tests
- Introduce algorithms for seasonality analysis including qs(), ocsb(), ch(), seasonal_strength(), trend_strength(), and spikiness().
- Add placeholder functions seasonality() and is_seasonal() for future implementation.
- Include necessary imports and exports for the new functionality. (by chrimaho)
ee30457: Add seasonality modules and related tests & docs
- Introduce seasonality.py for seasonality analysis
- Create test_seasonality.py for unit testing seasonality features
- Update mkdocs.yml to include seasonality documentation
- Add seasonality.md documentation file (by chrimaho)

v0.4.5

v0.4.5 - Add Stationarity Module🔗

v0.4.5
2026-01-13
data-science-extensions/ts-stat-tests/releases/v0.4.5

Release Notes

🚀 Overview🔗

Introduce the stationarity module, completing the third major component of the statistical test suite. This release provides a unified interface for seven core unit-root and stationarity tests, expanding the library's capability to assess time-series properties. Alongside the new module, this release implements a comprehensive overhaul of documentation standardisation, introduces new synthetic data generators, and achieves a milestone of 100% global code coverage through enhanced unit tests and doctest validation logic.

🛠️ Implementation details🔗

Stationarity module implementation🔗

Introduce the src/ts_stat_tests/algorithms/stationarity.py module, implementing seven essential stationarity tests:
- adf(): Augmented Dickey-Fuller test.
- kpss(): Kwiatkowski-Phillips-Schmidt-Shin test.
- pp(): Phillips-Perron test (utilising arch for improved reliability).
- za(): Zivot-Andrews test for structural breaks.
- ers(): Elliott-Rothenberg-Stock GLS test.
- vr(): Variance Ratio test.
- rur(): Range Unit Root test.
Implement a unified dispatcher in src/ts_stat_tests/tests/stationarity.py:
- stationarity(): Provide a consistent interface to execute any supported stationarity test via string aliases.
- is_stationary(): Return a standardised dictionary containing the Boolean result, test statistic, and interpreted p-value.
Standardise all algorithm outputs to native Python types (float, int, dict, tuple), ensuring consistent behaviour across the API.

Data utility enhancements🔗

Expand the src/ts_stat_tests/utils/data.py module with new synthetic data generators:
- get_normal_curve(): Generate data following a normal distribution.
- get_trend_data(): Generate data with deterministic linear trends.
- get_uniform_data(): Generate data following a uniform distribution.
Expose these as global variables (data_normal, data_trend, data_uniform) to facilitate reproducible examples and testing.
Refactor all docstring examples across the library (correlation, normality, regularity, stationarity) to use these standardised data imports.

Documentation standardisation🔗

Standardise all function docstrings to follow strict Google-style formatting and DFC compliance.
Move mathematical formulations into dedicated equation dropdowns using KaTeX notation for improved readability.
Update terminology to Australian English (e.g., standardise, summarise, recognise) and adopt the imperative mood across all descriptions.
Enhance the API reference by adding docs/code/stationarity.md and updating the implementation progress tables in README.md and docs/index.md.

✅ Checklist🔗

Implement seven core stationarity algorithms in the algorithms layer.
Develop a unified stationarity dispatcher and Boolean check in the tests layer.
Achieve 100% global code coverage across the ts_stat_tests package.
Standardise all docstrings and mathematical equations across the entire library.
Expand data utilities with new synthetic generators and integrate them into examples.
Implement robust cross-platform doctest validation for floating-point precision.

🧪 Testing and quality assurance🔗

Achieve 100% code coverage by adding comprehensive unit tests for the stationarity module and addressing edge cases in utils/errors.py.
Implement a custom doctest runner in src/utils/scripts.py to enable targeted validation of documentation examples.
Standardise numeric outputs in doctests using explicit string formatting to ensure stability across different operating systems and Python versions.
Pass all strict pylint (10/10), pyright (strict), and complexipy checks.

💪 Pull requests🔗

Implement stationarity module by @chrimaho in https://github.com/data-science-extensions/ts-stat-tests/pull/27

Full Changelog: https://github.com/data-science-extensions/ts-stat-tests/compare/v0.3.1...v0.4.5

Updates

ef5d67a: Fix docstring for the get_normal_curce() function (by chrimaho)
ee52e6d: Fix sys.argv length in the check_doctest_cli() function
The function accesses sys.argv[2] but the length check only ensures at least 2 elements exist (indices 0 and 1). This will cause an IndexError when exactly 2 arguments are provided. The check should be len(sys.argv) < 3 or the access should be sys.argv[1]. (by chrimaho)
ba4df2c: Add test_errors unit tests (by chrimaho)
9d3ca5c: Fix unit tests (by chrimaho)
df27be0: Update unit tests (by chrimaho)
5610b08: Add check_doctest() function to list of checks (by chrimaho)
667ac56: Standardise and modernise docstring examples for all algorithms and tests
- Replace legacy or inconsistent data preparation in docstring examples with unified imports from ts_stat_tests.utils.data
- Use consistent variable names such as normal, airline, random, trend, noise, and seasonal to improve clarity in all code examples
- Remove redundant or outdated example blocks, and add new, focused examples for every major algorithm to better illustrate expected output
- Update expected output values in example code to reflect current results from the associated algorithms
- Ensure all examples use the proper pycon code block style and follow a logical, repeatable structure for easier maintenance
- Improve exception and error reporting in usage notes, including explicit ValueError tracebacks for invalid algorithm parameters
- Align with Australian English spelling and imperative mood in documentation (by chrimaho)
95b107b: Expand doc filters to include algorithm-specific patterns
- Improve documentation clarity by adding new filter patterns for algorithm-specific members such as ^correlation, ^entropy, ^regularity, and ^stationarity
- Standardise filtering approach to ensure relevant functions and methods are included in docs for each algorithm section
- Reduce noise and improve discoverability for users browsing documentation (by chrimaho)
3ce6ff6: Standardise test docstring examples and update expected outputs
- Replace .values usage with direct assignment for test data examples to align with actual return types and avoid confusion
- Update expected numeric outputs in code examples for accuracy and consistency with current algorithm behaviour
- Clarify example titles and comments for improved readability and instructional value
- Ensure test documentation reflects standard dataset and normality testing patterns (by chrimaho)
ee2eea8: Fix bug in README links (by chrimaho)
ae14c7f: Fix typos (by chrimaho)
bf06810: Remove redundant test documentation sections from API docs
- Streamline documentation by removing duplicate ts_stat_tests.tests sections
- Focus on algorithm documentation for clarity and consistency
- Improve maintainability by reducing unnecessary content (by chrimaho)
5e9fe9f: Add doctest checking functions for targeted modules
- Introduce check_doctest() function to enable running doctests on specific files containing ts_stat_tests
- Add check_doctest_module() function to allow module-specific doctest execution
- Implement check_doctest_cli() function for command-line doctest invocation using arguments
- Comment out doctest check in check() function to prevent default execution until ready
- Facilitate easier validation of documentation examples and improve test coverage for doctests (by chrimaho)
a02994a: Clarify and standardise normality test return values
- Update the return type documentation to clarify that the Anderson-Darling algorithm yields a tuple of (stat, critical_values, significance_level) instead of a library-specific object
- Unpack the result from _jb() and return a tuple of test statistic and p-value for consistency with other algorithms
- Improve code clarity and maintainability by standardising interface and documentation (by chrimaho)
84e144e: Standardise and simplify normality test docstring examples
- Replace lengthy, varied examples with concise, uniform usage across all normality test function docstrings
- Use standardised datasets (data_airline, data_noise, data_normal, data_uniform) to ensure examples are reproducible and consistent
- Remove unrelated or overly detailed explanations, focusing on clarity and direct usage of test interfaces
- Improve clarity for end users by providing setup code blocks and clear example output
- Reduce maintenance burden by avoiding external library dependencies (e.g. remove references to sktime, numpy sin data) (by chrimaho)
243adad: Add usage examples for correlation() function
- Clarify usage by including practical examples
- Demonstrate different algorithms such as "acf" and "lb"
- Help users understand input requirements and expected outputs (by chrimaho)
5f8f675: Standardise usage of example data in docstrings
- Replace legacy data loading functions with direct data imports in example code for consistency
- Update example code to use data_airline and data_macrodata variables rather than call load_airline() or load_macrodata()
- Align example usage with updated data utilities to improve clarity and reproducibility
- Use from statsmodels import api as sm import style to match internal standards (by chrimaho)
6a3cf35: Add normal, trend, and uniform data generators
- Expand data utilities to include get_normal_curve(), get_trend_data(), and get_uniform_data() functions for generating additional synthetic datasets
- Expose new sample arrays such as data_normal and data_trend for easier access in tests and examples
- Support reproducibility by seeding random generators for new functions
- Enhance flexibility for statistical testing and demonstration scenarios (by chrimaho)
57ed6df: Standardise correlation algorithm documentation and references
- Update algorithm table to improve clarity and consistency, including renaming 'Breusch-Godfrey LM Test' and adding 'Lagrange Multiplier Test'
- Remove redundant library column entries for algorithms that are all from statsmodels
- Replace external reference with direct link to official statsmodels documentation for improved relevance
- Correct grammar and style for improved readability, including module names and package references
- Add documentation configuration block for ts_stat_tests.tests.correlation to align with existing structure
- Apply consistent filtering for member listing by excluding methods starting with is and _ (by chrimaho)
8ae9069: Standardise return types for normality test functions
- Ensure all normality test functions consistently return tuples of floats or lists, improving type clarity and downstream compatibility
- Refine type casting and result extraction in tests to align with updated function outputs
- Enhance robustness of result handling in both normality and stationarity test utilities (by chrimaho)
215a2af: Standardise normality test function return types
- Ensure all normality test functions return plain tuples for consistency and predictability
- Remove use of union types and result classes from function signatures to simplify API and reduce ambiguity
- Facilitate easier downstream usage and type checking by unifying return types across statistical test functions (by chrimaho)
cbeac37: Standardise and clarify documentation for statistical tests
- Clarify and condense docstrings for normality and stationarity tests, removing repetitive explanations and focusing on essential details
- Standardise notation in mathematical equations, consistently using LaTeX formatting and correct mathematical symbols (e.g., \Delta, \epsilon_t)
- Enhance explanations of test statistics, including concise mathematical definitions and parameter descriptions
- Add, correct, and clarify credit and reference sections for external libraries and sources, ensuring proper attribution
- Improve consistency in section headers and callout formatting (e.g., ??? success "Credit", ??? tip "See Also")
- Add an abstract and usage details to the correlation test dispatcher, listing supported algorithms
- Improve clarity and accuracy of parameter descriptions and summary sections
- Update docstring language to Australian English (e.g., 'standardise', 'summarise'), and use imperative mood throughout (by chrimaho)
8e78a81: Standardise exception docstrings for validation errors
- Replace inconsistent exception documentation with uniform Raises: sections detailing (ValueError) for invalid input across normality and regularity algorithms
- Remove redundant Raises: notes specifying no exceptions raised, clarifying documentation
- Improve clarity and maintainability of API reference for users and contributors (by chrimaho)
9bab801: Standardise and clarify docstrings for return types
- Update and clarify return type annotations and descriptions in docstrings for statistical test functions
- Remove outdated or ambiguous type hints, ensuring consistency and improved readability
- Adopt consistent dictionary and tuple return type formatting using Python type hinting
- Enhance explanations for result dictionaries, clearly listing expected fields and value types
- Improve maintainability and reduce confusion for users and contributors by standardising documentation (by chrimaho)
eee99be: Standardise parameter documentation formatting
- Align docstrings for parameter default values and descriptions across functions
- Remove inconsistent HTML tags and unify wording to improve readability
- Clarify default values using Default: value for easier reference
- Improve consistency for users consuming API documentation (by chrimaho)
c298449: Standardise and expand statistical test docstring examples
- Align all docstring code examples to a clear, consistent format using stepwise "Setup" and labelled example sections
- Replace outdated or confusing usage with current, project-specific data loader and methods
- Expand coverage for key statistical tests to demonstrate more parameter combinations and result formats
- Replace pprint with print for clarity, and showcase array slicing for readable outputs
- Add new examples for confidence intervals, advanced options, and more realistic outputs for acf(), pacf(), ccf(), lb(), lm(), and bglm() functions
- Improve readability for end users and facilitate easier learning by showing expected results and recommended invocation patterns (by chrimaho)
01926a1: Refactor test data setup to use shared utility functions
- Standardise test data access by importing shared utility functions
- Remove duplicate data generation logic from test setup file
- Simplify test base class to use direct references to shared data
- Improve maintainability and consistency of test data across the project (by chrimaho)
bc95f0f: Add utility functions and datasets for synthetic and real data
- Enable generation of random, sine wave, straight line, and fractional Gaussian noise datasets via dedicated utility functions for reproducible testing and experimentation
- Introduce caching for data generators using @lru_cache to improve performance and ensure consistent outputs
- Add direct access to classic real-world datasets by loading airline passenger and macroeconomic data as pandas objects
- Standardise data access patterns for downstream code and unit tests by exposing curated data objects
- Improve code clarity by listing public exports and grouping related logic (by chrimaho)
bafa581: Standardise return value documentation for functions in the correlation module
- Align function docstrings to consistently document return types and conditional outputs
- Improve clarity on optional outputs based on input parameters for core statistical test functions
- Enhance maintainability and user comprehension of function behaviour (by chrimaho)
6b19d4b: Standardise docstring formatting and improve references
- Replace ??? admonition syntax with !!! for credits and adjust headings to follow a consistent style
- Add hyperlinks to referenced resources for clarity and ease of access
- Update reference block styles to better distinguish between credits and citations
- Clarify return type documentation by removing redundant text and consolidating explanations (by chrimaho)
fa557b8: Refactor and standardise entropy docstrings and return types
- Refactor and condense docstrings for all entropy-related functions to improve clarity, consistency, and maintainability
- Standardise parameter descriptions and sections, ensuring consistent formatting and mathematical notation using LaTeX
- Add or update examples and notes for better user guidance, using realistic data and reproducible snippets
- Remove redundant or overly verbose explanations and outdated links, and clarify equations in dedicated sections
- Standardise all entropy function return types to explicitly cast outputs to float for consistency
- Update references to external libraries and methods, and ensure proper credit is given to the original sources
- Adopt imperative mood and Australian English spelling throughout documentation (by chrimaho)
f913bb3: Standardise docstring param and return sections
- Replace non-standard ???+ itemized markers with standard Params: and Returns: sections in function docstrings
- Improve clarity and consistency of documentation to align with typical Python docstring conventions
- Facilitate better integration with documentation tools and enhance readability for maintainers (by chrimaho)
1cc3d1e: Refactor and clarify statistical docstrings for correlation tests
- Reorganise and clarify docstring explanations for autocorrelation, partial autocorrelation, cross-correlation, Ljung-Box, Lagrange Multiplier, and Breusch-Godfrey tests
- Move mathematical definitions and formulas into dedicated equation dropdowns for better readability and accessibility
- Standardise notation (e.g. use $nlags$ , $n_{obs}$ , and proper LaTeX formatting) for consistency across documentation
- Replace raw Markdown examples with example dropdowns and improve usage clarity
- Fix minor typos and wording for Australian English, e.g. 'standardise', 'summarise', and ensure use of imperative mood in documentation
- Enhance explanation of test parameters, especially default values and deprecation warnings, for improved user guidance
- Remove redundant or duplicated information and modernise docstring structure with collapsible sections for details, examples, and references
- Credit the statsmodels library more consistently and visibly in each section (by chrimaho)
313e185: Expand and standardise stationarity test coverage
- Add tests for invalid algorithm handling in stationarity() function to ensure ValueError is raised as expected
- Standardise test results from statistical libraries by converting outputs to native Python types for consistency and easier assertions
- Improve branch coverage in is_stationary() function, including logic for boolean and non-numeric p-value handling
- Explicitly cover all algorithm branches in stationarity dispatcher for comprehensive test coverage
- Replace pmdarima's PPTest().should_diff() usage with direct arch.unitroot.PhillipsPerron() calls for result consistency across libraries
- Enhance readability and maintainability by grouping related imports and adding whitespace for class separation (by chrimaho)
35aa684: Increase test coverage for normality checks
- Add tests to cover branch scenarios in is_normal() function where result has statistic and pvalue attributes, or is a simple float
- Use unittest.mock.patch() to simulate different return types from normality() function
- Ensure more robust handling and validation of diverse outputs for improved reliability (by chrimaho)
c0ecca4: Add summary tables of key statistical tests to docs
- Improve discoverability by listing major correlation, normality, and regularity tests in markdown tables
- Standardise documentation with clear algorithm names, abbreviations, import scripts, and links for each method
- Aid users in identifying and accessing relevant statistical tools quickly for time series analysis (by chrimaho)
fe7c283: Standardise documentation references and configuration keys
- Remove redundant src. prefix from module references for consistency across documentation
- Update configuration blocks to use extra: root key, aligning with documentation standards
- Improve clarity and maintainability for future documentation updates (by chrimaho)
dcdbd26: Add missing numpy import to the stationarity module (by chrimaho)
c3093aa: Refactor stationarity result handling and typing
- Improve clarity and robustness of the is_stationary() function by explicitly casting result types and handling tuple returns safely
- Standardise type checking and assignment for statistics and p-values, supporting multiple possible types
- Cast return dictionary to expected types, ensuring outputs are consistently formatted and type-safe
- Enhance maintainability and reduce ambiguity in interpreting stationarity() function results (by chrimaho)
1b27b60: Introduce internal helper to standardise function calls
- Introduce _call() function to centralise keyword argument handling when invoking test functions
- Improve maintainability and reduce code duplication by standardising how stationarity test functions are called
- Lay groundwork for future changes that may require consistent invocation logic (by chrimaho)
e26a2e7: Expand supported types for algorithm keyword arguments
- Allow stationarity() function and is_stationary() function to accept ArrayLike as a valid type for keyword arguments
- Enable passing complex argument types directly to underlying algorithms, improving flexibility for users specifying parameters
- Standardise type annotations in docstrings to match updated function signatures (by chrimaho)
335aaa1: Expand keyword argument types to support array-like inputs
- Allow pass array-like types as values for **kwargs in the correlation() function overloads
- Enable greater flexibility for algorithms requiring non-scalar keyword arguments
- Improve API consistency and support for advanced use cases (by chrimaho)
985bd06: Add usage notes and credit to docstring for rur() function
- Clarify that p-values are interpolated and out-of-range statistics generate a warning
- Document that missing values are not handled
- acknowledge the statsmodels library as the basis for the implementation (by chrimaho)
546a1c9: Fix typo (by chrimaho)
2df720e: Standardise output types and add explicit casting in tests
- Ensure all test functions return standardised output formats by explicitly casting results
- Improve type safety and compatibility across statistical test functions
- Remove inline usage examples from docstrings to declutter documentation
- Adjust lag handling logic for the Phillips-Perron test to avoid invalid lag values
- Enhance maintainability and clarity for future extension or integration (by chrimaho)
d293fa1: Relocate and standardise example blocks in docstrings
- Move all example blocks to consistent positions at the end of each function docstring for improved documentation clarity
- Use uniform markdown formatting for all example code, ensuring consistent presentation across tests including kpss(), rur(), za(), pp(), ers(), and vr()
- Add missing "See Also" section to the Zivot-Andrews test to align with other stationarity tests and aid discoverability
- Improve documentation structure to aid user navigation and future maintenance (by chrimaho)
4fd8c98: Add missing Credit section to the docstring for the adf() function (by chrimaho)
714073e: Standardise type annotations and return values in stationarity algorithms
- Align type annotations for parameters and return types to be more explicit and consistent, using domain-specific type aliases where appropriate
- Update docstrings to accurately reflect parameter types and improve clarity on optional arguments
- Adjust return signatures of kpss(), rur(), pp(), za(), and ers() functions for consistency with value structure and expected outputs
- Clarify the use of store parameter and its effect on returned result tuples
- Reduce ambiguity and potential confusion for users and static analysis tools, improving maintainability and type safety throughout the statistical test API (by chrimaho)
fe5eae8: Refactor and standardise docstrings for consistency
- Improve consistency and clarity in docstring formatting across all stationarity test functions
- Replace scattered "Notes", "Credits", and "Examples" sections with standardised locations and Markdown formatting
- Move or consolidate important notes into ??? note "Notes" blocks for better visibility
- Relocate all credit attributions to the end of docstrings in a uniform !!! success "Credit" block
- Remove redundant or duplicate content and streamline explanations, focusing on essential mathematical and usage details
- Enhance maintainability by ensuring all test documentation sections follow a clear, predictable structure (by chrimaho)
e613d7c: Standardise docstring admonitions for clarity
- Replace all occurrences of !!! summary with !!! note to improve consistency in documentation style
- Change ???+ info to ???+ abstract for detailed sections to better reflect their purpose
- Update !!! example to ???+ example for improved formatting and alignment with documentation tools
- Amend ??? info to ??? note to standardise note sections across docstrings
- Improve readability and maintainability of docstring formatting for future contributors (by chrimaho)
9ec60ac: Refactor normality result handling for robustness
- Improve robustness of result extraction by using getattr() and type casting to handle diverse result types from statistical tests
- Ensure consistent float conversion and type safety for statistic and p_value fields
- Enhance code clarity around Anderson-Darling test result parsing by separating tuple unpacking and array conversions
- Reduce risk of runtime errors from unexpected result formats, standardise output structure, and improve maintainability (by chrimaho)
c2fdd32: Add calculation section to docstring configuration
- Introduce a calculation section with equation admonition and unique order in the docstring structure
- Standardise documentation by enabling explicit space for calculation or formula details
- Improve clarity by shifting the order of existing sections to accommodate calculation (by chrimaho)
afabea1: Expand pylint disable list for improved code maintainability
- Add R0911, R0914 to the pylint disable list for warnings related to too many return statements and too many local variables
- Support more flexible code patterns without triggering style violations
- Facilitate maintainability by reducing unnecessary linting noise and focusing on critical issues (by chrimaho)
27aca4e: Fix various typos
- Standardise test naming and refactor variable names
- Replace inconsistent test method names to use test_stationarity_za_* format for clarity and consistency
- Update variable names from indx to index for readability and alignment with standard conventions
- Change import statement to use tests.setup for improved test structure organisation
- Improve code maintainability and ease future test extensions (by chrimaho)
1735768: Standardise typing and signatures for keyword arguments
- Replace generic Any keyword argument types with explicit Union[float, int, str, bool, None] to improve static type checking and code clarity
- Update function return type annotations for accuracy, clarifying output types for consumers and future maintainers
- Remove unused Any imports to clean up namespace and avoid confusion
- Apply cast(int, delay) instead of cast(Any, delay) in permutation_entropy() function to enforce type correctness
- Use object instead of Any in generate_error_message() function parameters to further restrict accepted types
- Improve docstrings to accurately reflect updated parameter and return types, supporting clearer auto-generated documentation (by chrimaho)
f90d46b: Standardise and update stationarity test documentation
- Update statistical test links to point to relevant documentation pages for improved navigation
- Standardise stationarity test implementation indicators from 🔲 to ✅ in the README to reflect completed Python support
- Revise progress table to show full implementation and unit-test coverage for stationarity tests
- Clarify test status and improve transparency of documentation for users (by chrimaho)
826dd68: Update stationarity docs to clarify test coverage
- Standardise terminology and references for time series stationarity
- Reference authoritative external sources for definitions and further reading
- Refine library comparison table to focus on statsmodels and arch for numerical reliability
- Clarify rationale for selecting statsmodels and arch packages, highlighting their comprehensive suite of tests
- Update source module references to point to correct stationarity modules (by chrimaho)
49eb822: Standardise hypothesis logic and update PP test implementation
- Clarify and align null hypothesis handling for stationarity tests, ensuring consistent interpretation of p-values across algorithms
- Replace pmdarima PP test implementation with arch library version, refactor pp() function signature and return values
- Improve documentation with explicit references, trend and test type options for PP test, and fix docstring examples
- Update VR, ERS, RUR test hypotheses in docstrings for accuracy
- Enhance stationarity determination logic to robustly support all test types
  Relates to improved consistency and accuracy in statistical testing APIs (by chrimaho)
17dd78a: Add documentation for time-series stationarity tests
- Provide an overview of key stationarity and seasonality tests for time-series data, comparing major Python libraries and their relevant algorithms
- List supported statistical tests with import instructions and direct links to documentation
- Reference implementation modules and usage for both algorithmic and test layers
- Facilitate easier selection and understanding of statistical tools for assessing time-series suitability (by chrimaho)
5acad0c: Add comprehensive unit tests for stationarity algorithms
- Expand test coverage for statistical stationarity tests by introducing new test classes and cases
- Standardise testing by comparing custom implementations to third-party libraries, verifying consistency of results
- Include tests for major time series stationarity checks such as ADF, KPSS, PP, ERS, ZA, RUR, and VR
- Use selective assertions and rounding to address minor numerical discrepancies between implementations
- Improve reliability of statistical test code and facilitate future refactoring or algorithm enhancements (by chrimaho)
0a1feb9: add comprehensive docstrings for stationarity test functions
- improve clarity and usability by providing detailed docstrings for stationarity() and is_stationary() functions
- include algorithm options, parameter descriptions, return values, and example usage in the documentation
- guide users on the interpretation of test results and clarify statistical assumptions (by chrimaho)
a1da846: add module header and summary docstring for clarity
- provide an overview section and description to improve documentation
- include a summary docstring for the module to explain its purpose
- enhance code readability and maintainability by standardising file headers (by chrimaho)
e497a9c: Add stationarity test wrappers with unified interface
- Provide stationarity() function to standardise access to multiple stationarity tests via a single interface
- Implement is_stationary() function to streamline stationarity assessment using consistent logic across algorithms
- Simplify selection of test algorithms with flexible naming and error handling
- Improve code maintainability and readability by centralising algorithm dispatch and result interpretation (by chrimaho)
68fbec0: Add comprehensive docstrings for all stationarity test functions
- Improve clarity and usability by introducing detailed docstrings for each statistical test function
- Include mathematical formulations, parameter descriptions, expected return values, and practical code examples
- Standardise documentation style with markdown formatting and KaTeX math for equations
- Credit original libraries and add references for further reading to facilitate user understanding
- Enhance discoverability with "See Also" sections linking related tests (by chrimaho)
1eba0d9: Add module overview and detailed documentation for stationarity algorithms
- Introduce a comprehensive module header describing the purpose and scope of stationarity tests in time series analysis
- Add a detailed summary discussing the statistical rationale and importance of stationarity tests, including ADF, KPSS, PP, ERS, and VR
- Reference useful external resources for further reading to aid understanding
- Improve code readability and maintainability by clarifying context and intent for future contributors (by chrimaho)
40d087b: Add stationarity test wrappers with type validation
- Provide standardised function interfaces for multiple stationarity tests
- Use type validation via @typechecked decorator for input safety
- Expose consistent argument options and overloads for flexible usage
- Centralise imports and constants to simplify maintenance and readability
- Improve usability by mirroring third-party library signatures (by chrimaho)
0c6c10b: Add stationarity test scaffolding to codebase
- Prepare groundwork for future stationarity feature implementation
- Standardise documentation and navigation by including stationarity references
- Enable future unit testing via placeholder test files
- Facilitate expansion of statistical algorithms with initial stationarity structure (by chrimaho)

v0.3.1

v0.3.1 - Add Normality Module🔗

v0.3.1
2026-01-08
data-science-extensions/ts-stat-tests/releases/v0.3.1

Release Notes

🚀 Overview🔗

Introduce the normality module, providing a suite of statistical tests to assess the distribution of time series data. Significantly expand the documentation and test coverage for the regularity and normality modules, update the implementation progress, and improve the clarity and accuracy of references across the project. Notably, introduce new documentation pages, update the README.md and progress tables to reflect completed work, and add a comprehensive test suite for normality algorithms.

🛠️ Implementation details🔗

Normality module implementation🔗

Introduce the src/ts_stat_tests/algorithms/normality.py module, implementing five core normality tests: jb(), ob(), sw(), dp(), and ad().
Implement a unified interface in src/ts_stat_tests/tests/normality.py with the following standardised functions:
- normality(): Act as a dispatcher to run a selected normality test.
- is_normal(): Return a standardised dictionary containing the test result (True/False), statistic, and p-value.
Support various nan_policy options ("propagate", "raise", "omit") across algorithms to handle missing data gracefully.
Provide detailed mathematical documentation, including LaTeX equations and bibliographic references, in the docstrings of all normality test functions.

Regularity module enhancements🔗

Expand the src/ts_stat_tests/algorithms/regularity.py module by exposing additional entropy measures: permutation_entropy() and svd_entropy().
Update the entropy(), regularity(), and is_regular() functions in src/ts_stat_tests/tests/regularity.py to support these new algorithms via intuitive string aliases.
Standardise parameter handling and improve error messaging using generate_error_message().
Enhance docstrings with comprehensive examples, mathematical context, and internal cross-references.

Documentation and progress tracking🔗

Create new documentation pages to provide detailed guidance:
- docs/code/normality.md: Provide an overview of normality testing, rationale, algorithm details, and API references.
- docs/code/regularity.md: Detail entropy-based regularity testing and provide a comprehensive API guide.
Update docs/code/index.md and README.md progress tables to mark the normality and regularity modules as 100% complete.
Refactor docs/code/correlation.md to correct and clarify module references and formatting.
Register the normality section in mkdocs.yml to improve site navigation and information architecture.

Testing and quality assurance🔗

Add a comprehensive unit test suite in src/tests/test_normality.py, covering all implemented normality algorithms and edge cases.
Implement a specific mock-based scenario to trigger the fallback path in is_normal(), ensuring 100% code coverage for the normality module.
Adhere to Australian English spelling (e.g. standardise, optimise, recognise) and imperative coding style throughout the codebase.
Ensure all code changes pass linting and type checks to maintain high standards for reliability, usability, and transparency.

✅ Checklist🔗

Implement core normality test algorithms in src/ts_stat_tests/algorithms/normality.py.
Create unified normality test interfaces in src/ts_stat_tests/tests/normality.py.
Add comprehensive unit tests in src/tests/test_normality.py and achieve 100% coverage.
Expand regularity algorithms and update convenience wrappers.
Generate detailed documentation for both normality and regularity modules.
Update project-wide progress tables in README.md and docs/code/index.md.
Standardise spelling to Australian English and ensure consistent formatting.

💪 Pull Requests🔗

Implement normality module and enhance regularity documentation by @chrimaho in https://github.com/data-science-extensions/ts-stat-tests/pull/26

Full Changelog: https://github.com/data-science-extensions/ts-stat-tests/compare/v0.2.0...v0.3.1

Updates

90cd570: Fix grammar error (by chrimaho)
b708fe1: Remove documentation
Duplicate dictionary key "entropy" in the return type documentation. The second entry on line 341 should be removed as it's identical to the one on line 340. (by chrimaho)
4f554e5: Incomplete documentation
The text "because REASONS" is a placeholder that should be replaced with actual reasoning for why the AntroPy package was selected. (by chrimaho)
1427c5e: Document additional entropy algorithms in regularity tests
- Expand documentation to include permutation_entropy() and svd_entropy() functions as supported algorithms
- Clarify algorithm selection options and aliases for improved user guidance
- Ensure consistency across docstrings and abstract sections for easier reference (by chrimaho)
1782d7b: Update progress docs for completed regularity and normality
- Mark completed antropy regularity algorithms and add svd_entropy() function to documentation
- Mark all normality algorithms as implemented and tested
- Reflect full test and unit test coverage for correlation, regularity, and normality modules in progress tables
- Improve accuracy of documentation for current feature and test status (by chrimaho)
d006b1c: Add documentation for normality tests and algorithms
- Provide an overview of normality testing in time-series analysis, including rationale and references
- Detail the use of scipy.stats and statsmodels libraries for statistical tests
- Link to relevant source modules for implementation details
- Outline available tests and algorithms via code documentation structure
- Help users understand when and why to perform normality checks on residuals (by chrimaho)
4584b3f: Add unit test for fallback branch in normality check
- Improve test coverage by introducing a scenario that triggers the fallback path in the is_normal() function
- Ensure behaviour when the normality() function returns an object that is neither a tuple/list nor has a pvalue attribute
- Confirm that the is_normal() function handles unexpected return types gracefully and returns expected results (by chrimaho)
ec8d327: Introduce unified normality test interface and refactor docs
- Add normality() and is_normal() functions to standardise access to multiple normality test algorithms
- Refactor and expand documentation for all normality algorithms, improving clarity and consistency of usage examples and equations
- Replace scattered summary/info/example blocks with a unified doc structure and imperative notes, using Australian English spelling
- Update and extend test coverage for new interfaces, ensuring comprehensive behaviour for all supported normality tests
- Improve parameter handling and error messaging for invalid algorithm selection using generate_error_message()
- Align with latest scipy and statsmodels result object conventions for type safety and compatibility (by chrimaho)
ad67e9d: Add detailed docstrings for normality test functions
- Improve documentation by adding comprehensive docstrings to all normality test functions
- Include summaries, parameter descriptions, return types, example usages, equations, references, notes, and related function links
- Standardise documentation style and formatting for clarity and consistency
- Enhance usability for end users by providing practical guidance and mathematical context, referencing the relevant statistical literature and library sources (by chrimaho)
25c4823: Add module-level summary and structured docstrings
- Provide a clear overview and description for the normality testing algorithms
- Improve future maintainability by standardising documentation structure
- Clarify module purpose for statistical analysis and forecasting workflows
- Facilitate easier onboarding for new contributors (by chrimaho)
8f21f94: Add statistical normality test algorithm implementations
- Provide comprehensive implementations for assessing data normality via jb(), ob(), sw(), dp(), and ad() functions.
- Include detailed docstrings, equations, practical examples, and references to support correct use and interpretation.
- Standardise parameter types, outputs, and documentation for consistent usage across different normality tests.
- Facilitate robust statistical validation in time series forecasting workflows by exposing multiple established tests. (by chrimaho)
a4cfa2a: Add normality test documentation and stubs
- Create initial documentation for normality tests
- Add normality test file stubs for future implementation
- Update navigation to include normality section
- Prepare codebase for standardisation of normality test approach (by chrimaho)
c0e61c5: Fix docs reference bugs (by chrimaho)
1083496: Add regularity documentation
- Introduce comprehensive overview and rationale for regularity testing using entropy-based algorithms
- Standardise documentation with references, examples, and usage guidelines for ts_stat_tests modules
- Clarify differences between approximate entropy and sample entropy in forecasting context
- Link to further resources on time-series analysis methodology and data quality
- Detail regularity algorithms and tests, including filtering logic for code navigation (by chrimaho)

v0.2.0

v0.2.0 - Add Regularity Algorithms🔗

v0.2.0
2026-01-06
data-science-extensions/ts-stat-tests/releases/v0.2.0

Release Notes

🚀 Overview🔗

This release introduces a suite of regularity algorithms for time series analysis, including several entropy-based measures. It also features significant performance optimisations for project maintenance utilities and ensures the package achieves 100% unit test coverage. These enhancements stabilise the codebase and provide a robust foundation for assessing the complexity and regularity of time series data.

🛠️ Implementation details🔗

Regularity and entropy algorithms🔗

Implement the entropy() function as a unified dispatcher for multiple entropy calculation algorithms, supporting Approximate, Sample, Permutation, Spectral, and SVD entropy.
Introduce the regularity() function to provide a standardised wrapper for assessing time series regularity.
Add the is_regular() function to determine if a time series meets a defined regularity threshold.
Provide comprehensive wrappers for the approx_entropy(), sample_entropy(), permutation_entropy(), spectral_entropy(), and svd_entropy() functions.
Utilise the antropy library for underlying calculations and apply the @typechecked decorator for rigorous runtime parameter validation.
Integrate mathematical documentation within docstrings to describe the logic behind regularity measures.

Performance and maintenance utilities🔗

Refactor the get_all_files() function in the scripts.py module to use system-level git ls-files and find commands.
Optimise directory traversal for project maintenance tasks, reducing execution time from 41 seconds to near-instantaneous.
Update the Pylint configuration to ignore R0801 (duplicate code) warnings, facilitating specialised test case implementation.
Fix documentation build failures by resolving mkdocs-material extension inconsistencies in the development environment.

Quality assurance and environment🔗

Achieve 100% unit test coverage across the library by adding comprehensive test cases for regularity algorithms, data loading, and error handling.
Enhance the assert_almost_equal() and is_almost_equal() utility functions to support precise verification of statistical results.
Update the load_airline() function to include stricter data type validation and error reporting.
Restrict the numpy version to < 2.4 in pyproject.toml to maintain compatibility with Numba() and antropy.
Drop support for Python 3.9 in the CI/CD workflows to align with modern Python lifecycle standards.

✅ Checklist🔗

Implement core regularity algorithms and entropy measures.
Optimise file discovery performance in maintenance scripts.
Reach 100% unit test coverage across all modules.
Update package dependencies and Python support.
Standardise documentation and verify build success.

💪 Pull Requests🔗

Implement regularity algorithms by @chrimaho in https://github.com/data-science-extensions/ts-stat-tests/pull/25

Full Changelog: https://github.com/data-science-extensions/ts-stat-tests/compare/v0.1.0...v0.2.0

Updates

df1172b: Fix f-string syntax in documentation
The example code contains syntax errors with mismatched quotes. Lines 665, 670, and other locations use straight double quotes inside the f-string which will cause a Python syntax error. The method parameter value should use single quotes or escaped double quotes. (by chrimaho)
388713e: Fix documentation typo
The examples show calls to "sample_entropy", "approx_entropy", and "spectral_entropy" functions, but the surrounding documentation indicates these should be calls to "is_regular()" with different algorithm parameters. The function names in the examples should be corrected to match the function being documented. (by chrimaho)
49ecb5a: Fix documentation consistency
There's an inconsistency in the documentation examples. The function calls in the examples use "approx_entropy", "sample_entropy", and "spectral_entropy" directly, but the surrounding text indicates these should be calls to the "regularity()" function with different algorithm parameters. The examples should match the function being documented. (by chrimaho)
579c51d: Fix redundant docstring comment
The docstring contains duplicate quote marks in the "Summary" admonition. It should be either !!! note "Summary" or just !!! note Summary without the extra quotes. (by chrimaho)
9b125fb: Fix another duplication
The documentation lists "approx" twice in line 337 which appears to be a typo. It should likely list the valid string options as: ["app", "approx"] based on the implementation in the entropy function. (by chrimaho)
b8461aa: Fix duplication
The documentation lists "approx" twice in line 99 which appears to be a typo. It should likely list the valid string options as: ["app", "approx"] based on the implementation in the options dictionary on line 181. (by chrimaho)
90ce8c2: Update src/ts_stat_tests/algorithms/regularity.py
The example output shows a raw array representation "array([0.9426, 0.9383, 0.9411, 0.9375])" instead of formatted output. Since this is in an f-string print statement, the output format is inconsistent with the other examples which show formatted floats. Consider using .tolist() or formatting the array elements consistently. (by chrimaho)
43954ef: Remove redundant docs comment
The docstring contains duplicate quote marks in the "Summary" admonition. It should be either !!! note "Summary" or just !!! note Summary without the extra quotes. It should not have duplicate "Summary" sections. (by chrimaho)
81defae: Remove duplicates
The documentation lists "approx" twice in line 227 which appears to be a typo. It should likely list the valid string options as: ["app", "approx"] based on the implementation in the entropy function. (by chrimaho)
b2c8d65: Fix numpy cast process
The cast to np.ndarray on line 429 may be incorrect if the input x is actually an ArrayLike that is not a numpy array (e.g., a list or pd.Series). The np.std() function already accepts ArrayLike inputs, so this cast is both unnecessary and potentially misleading. Consider removing the cast() or verifying that x has been converted to a numpy array first. (by chrimaho)
a344cca: Fix numpy version constraint
The numpy version constraint "numpy<2.4" lacks a lower bound, which could allow installation of very old numpy versions (including 1.x) that may not be compatible with the codebase. Consider adding a lower bound like "numpy>=2.0.0,<2.4" to ensure compatibility, matching the override-dependencies specification on line 45. (by chrimaho)
541a996: Expand test coverage for utils and entropy logic
- Include unit tests for the load_airline() function to verify data type validation and error handling.
- Add coverage for the is_almost_equal() and assert_almost_equal() utility functions to testing parameter validation and failure messages.
- Minimise gaps in coverage for the svd_entropy() and entropy() functions by testing direct calls and algorithm selections.
- Verify tolerance logic in the is_regular() function for None and string-based input values.
- Bring test coverage up to 100% (by chrimaho)
63ebd14: Add permutation and SVD entropy options
- Expose the permutation_entropy() and svd_entropy() functions within the regularity testing module.
- Extend the entropy() function logic to include support for permutation and SVD algorithm types.
- Update the options mapping to recognise new aliases and standardise selection of these entropy measures. (by chrimaho)
369a321: Optimise file discovery performance
- Use the find system command within the get_all_files() function to accelerate directory traversal.
- Implement a fallback mechanism to the .glob() method on the Path() class if the system command is unavailable.
- Prune .venv and hidden directories in the search path to minimise processing time.
- Standardise the output by applying the sorted() function to the list of discovered file paths.
- Add a docstring to the get_all_files() function to document the dual-method execution logic. (by chrimaho)
4dc7cad: Update the Pylint configuration to suppress R0801 duplicate-code check
- Ignore duplicate-code warnings to reduce linting noise
- Allow code repetition where refactoring for deduplication is not desirable (by chrimaho)
5f6b9e5: Refine documentation and update test baselines
- Update baseline numerical values for approx_entropy(), sample_entropy(), permutation_entropy(), and spectral_entropy() functions in the test suite to match revised calculations.
- Standardise the type hint for the metric parameter in entropy(), regularity(), and is_regular() functions to use VALID_KDTREE_METRIC_OPTIONS.
- Refine docstrings by removing redundant admonition titles in svd_entropy(), entropy(), regularity(), and is_regular() functions. (by chrimaho)
89ce566: Improve regularity test type safety and documentation
- Standardise docstring admonition blocks to use consistent note and abstract labels.
- Add a comprehensive docstring for the svd_entropy() function including parameters and return types.
- Enhance type safety by using the cast() function and specific type aliases for metric and method arguments.
- Refactor logic in the entropy() function to replace nested branches with early returns.
- Update parameter documentation for tolerance and metric in the approx_entropy() and sample_entropy() functions.
- Ensure the is_regular() function returns predictable types through explicit dictionary value casting.
- Add module-level headers and summary documentation to the regularity test utility source code. (by chrimaho)
3e13d54: Restrict NumPy version for Numba compatibility
- Limit numpy version to less than 2.4 to ensure compatibility with Numba
- Synchronise the version cap across project dependencies and uv overrides (by chrimaho)
c0d065d: Add unit tests for regularity and entropy metrics
- Implement the TestRegularity() class to provide comprehensive unit testing for regularity and entropy algorithms.
- Add test cases for the approx_entropy(), sample_entropy(), permutation_entropy(), and spectral_entropy() functions across multiple data types including noise, sine waves, and linear trends.
- Verify that the is_regular() function validates return keys and value types while handling invalid algorithm parameters.
- Validate numerical accuracy for various scenarios such as multidimensional arrays and different distance metrics used within the algorithms.
- Utilise the assert_almost_equal() function to ensure precise verification of calculated entropy values. (by chrimaho)
f71614b: Standardise documentation and fix docstring typos
- Update code block labels to pycon to support interactive example rendering
- Fix a typographical error in the import statement for the spectral_entropy() function
- Standardise spacing in arithmetic expressions within the sample_entropy() and permutation_entropy() functions
- Improve docstring layout by adding vertical spacing before example blocks in the approx_entropy() and spectral_entropy() functions (by chrimaho)
700a063: Add regularity and entropy statistical tests
- Introduce entropy() function as a unified interface for several entropy calculation algorithms
- Add regularity() function to provide a pass-through wrapper for assessing time series regularity
- Implement is_regular() function to determine if a dataset is regular by comparing entropy against a threshold
- Apply typechecked decorator to ensure robust parameter validation for all new functions
- Include helper logic to calculate a default tolerance based on the standard deviation of the input data
- Support approx_entropy(), sample_entropy(), and spectral_entropy() functions via a central entry point
- Provide internal parameter validation to normalise algorithm selection strings (by chrimaho)
50b5405: Add regularity test file
- Initialise the test_regularity.py file to provide a structure for upcoming test cases (by chrimaho)
f33cbf7: Add documentation for regularity functions
- Provide extensive docstrings for approx_entropy(), sample_entropy(), permutation_entropy(), and spectral_entropy() functions.
- Include mathematical formulations such as $ApEn(m, r, N) = \phi_m(r) - \phi_{m+1}(r)$ to describe regularity logic.
- Add practical examples demonstrating how to apply entropy functions to various datasets.
- Document parameter configurations such as the order and metric arguments for the sample_entropy() function.
- Detail the steps required to normalise results within the permutation_entropy() and spectral_entropy() functions.
- Incorporate academic references and credit the AntroPy library for the underlying implementations.ts (by chrimaho)
759edd6: Refine typing and expand entropy functions
- Introduce Literal type hints to restrict valid options for metric and method parameters.
- Add tolerance parameter to approx_entropy() and sample_entropy() functions.
- Standardise parameter definitions in approx_entropy(), sample_entropy(), and spectral_entropy() functions using VALID_KDTREE_METRIC_OPTIONS and VALID_SPECTRAL_ENTROPY_METHOD_OPTIONS constants. (by chrimaho)
5c48940: Add svd_entropy() function for regularity testing
- Include svd_entropy in the __all__ list to expose the new algorithm.
- Implement the svd_entropy() function as a type-checked wrapper for the antropy implementation to provide Singular Value Decomposition entropy calculations. (by chrimaho)
a51404c: Add documentation header and module docstring
- Add a descriptive header to the regularity module to clarify its purpose.
- Include a module-level docstring that summarises functionality for computing regularity measures. (by chrimaho)
1796788: Add regularity entropy algorithms
- Implement approx_entropy(), sample_entropy(), permutation_entropy(), and spectral_entropy() functions
- Utilise the antropy library for underlying entropy calculations
- Apply the typechecked decorator to ensure parameter type safety
- Define the __all__ list to expose the public API of the module (by chrimaho)
ffb08f7: Add regularity module and documentation
- Provide regularity statistical tests within the algorithms library
- Ensure code quality with a new test suite
- Initialise the documentation and update the site navigation (by chrimaho)
dd00a96: Drop Python 3.9 support from CD workflow
- Remove 3.9 from the environment matrix to align with modern support standards.
- Optimise the deployment pipeline by focusing on more recent releases. (by chrimaho)

v0.1.0

v0.1.0 - Initial release of Time Series Statistical Tests🔗

v0.1.0
2026-01-05
data-science-extensions/ts-stat-tests/releases/v0.1.0

Release Notes

🚀 Overview🔗

This is the initial release of ts-stat-tests, a Python package dedicated to providing a unified and standardised interface for time series statistical testing. This release establishes the foundational project infrastructure, implements a comprehensive suite of correlation algorithms, and provides extensive documentation and automated CI/CD pipelines. It aims to bridge the gap between R and Python for time series analysis by offering a single, robust library for standard time series statistical tests.

🛠️ Implementation details🔗

Core correlation algorithms🔗

Introduce the acf() function, pacf() function, and ccf() function to estimate autocorrelation, partial autocorrelation, and cross-correlation functions.
Implement residual diagnostic tests including the lb() function for Ljung-Box testing, the lm() function for Lagrange Multiplier tests, and the bglm() function for Breusch-Godfrey tests.
Provide a unified correlation() dispatcher function to simplify algorithm selection via string aliases.
Utilise the statsmodels library as the underlying engine to ensure robust and standardised statistical results.
Include detailed docstrings for all algorithms, featuring mathematical definitions, parameter details, and practical usage examples.

Project infrastructure and automation🔗

Initialise the project structure using uv for efficient dependency management and build orchestration.
Implement a centralised scripts.py module to provide a suite of automation utilities for linting, checking, and maintenance.
Configure pre-commit hooks to enforce code quality, standardise formatting, and prevent direct commits to the main branch.
Establish GitHub Actions workflows for continuous integration (ci.yml) and continuous delivery (cd.yml) to automate testing and releases.
Integrate Dependabot to automate updates for GitHub Actions dependencies on a weekly schedule.

Documentation and presentation🔗

Launch a comprehensive documentation site using mkdocs with the material theme and custom stylesheets.
Expand the README.md with project badges, a motivation section, and a detailed feature implementation table.
Implement custom documentation hooks in shortcodes.py to support dynamic content and enhanced rendering.
Add placeholder pages and "To Do" notes for CHANGELOG.md and CONTRIBUTING.md to guide future documentation efforts.

Type safety and quality assurance🔗

Enforce runtime type safety by applying the @typechecked decorator to all core algorithm functions.
Integrate pyright for rigorous static type analysis and resolve complex type-hinting issues.
Implement utility functions for error reporting and float comparison, including generate_error_message(), is_almost_equal(), and assert_almost_equal().
Achieve 100% unit test coverage for the correlation module to ensure reliability and correctness.

✅ Checklist🔗

Establish foundational project structure and configuration.
Implement core correlation algorithms and diagnostic tests.
Create a centralised automation and maintenance script.
Launch a comprehensive documentation site and expand README.md.
Configure CI/CD pipelines and pre-commit hooks.
Enforce runtime and static type safety across the package.
Achieve high unit test coverage for the initial release.

💪 Pull Requests🔗

Initialise project structure and configuration by @chrimaho in https://github.com/data-science-extensions/ts-stat-tests/pull/15
Add correlation algorithms for time series analysis by @chrimaho in https://github.com/data-science-extensions/ts-stat-tests/pull/16
Set up unit tests and enhance development infrastructure by @chrimaho in https://github.com/data-science-extensions/ts-stat-tests/pull/17
Establish CI/CD pipelines and automate release workflows by @chrimaho in https://github.com/data-science-extensions/ts-stat-tests/pull/18
Bump actions/checkout from 5 to 6 by @dependabot[bot] in https://github.com/data-science-extensions/ts-stat-tests/pull/19
Bump actions/setup-python from 5 to 6 by @dependabot[bot] in https://github.com/data-science-extensions/ts-stat-tests/pull/23
Bump astral-sh/setup-uv from 6 to 7 by @dependabot[bot] in https://github.com/data-science-extensions/ts-stat-tests/pull/22
Bump actions/upload-artifact from 4 to 6 by @dependabot[bot] in https://github.com/data-science-extensions/ts-stat-tests/pull/21
Bump actions/download-artifact from 5 to 7 by @dependabot[bot] in https://github.com/data-science-extensions/ts-stat-tests/pull/20
Enhance project documentation and presentation by @chrimaho in https://github.com/data-science-extensions/ts-stat-tests/pull/24

🆕 New Contributors🔗

@chrimaho made their first contribution in https://github.com/data-science-extensions/ts-stat-tests/pull/15
@dependabot[bot] made their first contribution in https://github.com/data-science-extensions/ts-stat-tests/pull/19

Full Changelog: https://github.com/data-science-extensions/ts-stat-tests/commits/v0.1.0

Updates

085a4c0: Re-enable key components in the cd workflow, ready for first deployment (by chrimaho)
a722504: Ensure coverage reports are staged even when directory paths are ignored.
- Use the --force flag in the git_add_coverage_report() function.
- Bypass .gitignore restrictions to guarantee documentation updates. (by chrimaho)
85f68e6: Hide some lines in cd to check workflow during release (by chrimaho)
16c9288: Standardise type hints and remove dependencies
- Replace custom type hints with native Python equivalents for the __all__ variable
- Remove the unnecessary import from the toolbox_python.collection_types module
- Update the replace() function to use the re.Match() class directly (by chrimaho)
8e76143: Fix missing justifications in docs (by chrimaho)
a3085ff: Fix typos (by chrimaho)
aef58db: Fix typo (by chrimaho)
90c953b: Fix incorrect percentages (by chrimaho)
15202ad: Re-enable mkdocs checks (by chrimaho)
728c24a: Add overview of module progress (by chrimaho)
02d36d1: Add correlation module docs page (by chrimaho)
60c7127: Fix some typos on the correlation module docs (by chrimaho)
fe9003d: Add docs landing page and home page structures (by chrimaho)
37fab51: Add custom hooks for docs to utilise (by chrimaho)
83dd55f: Initial commit of package icons (by chrimaho)
8c3f0cf: Initial commit of package stylesheets (by chrimaho)
950c1fb: Fill in README file (by chrimaho)
e2c62bd: Initial commit of standard package docs (by chrimaho)
1f1a9d6: Initial commit of mkdocs config file (by chrimaho)
c7839a4: Set UTF-8 encoding for Python I/O in CI
- Define the PYTHONIOENCODING environment variable to ensure consistent character handling across CI environments (by chrimaho)
cc368ff: Fix typo (by chrimaho)
0b2d590: Standardise variable naming convention
- Rename the repository variable to lower-case to follow standard naming conventions for local instances within the main() function.
- Update references to the variable when calling the .get_releases() method and .get_commits() method.
- The variable name REPO is defined in uppercase following constant naming conventions, but it's not a constant - it's a dynamically created repository object. Consider using lowercase repo to follow Python naming conventions where uppercase names are reserved for constants. (by chrimaho)
ed5404d: Initialise the output file before starting the main process
- Call the prepare_output_file() function to ensure the destination is ready for writing.
- Set up the file environment prior to opening the GitHub and file context managers. (by chrimaho)
6060627: Define constant for short SHA length
- Introduce SHORT_SHA_LENGTH constant to replace hardcoded magic numbers
- Update add_commit_info() function to use defined constant
- Standardise representation of short commit identifiers for improved maintainability (by chrimaho)
aa243b7: FixPotential bug
If commit.author is None, the code constructs an incomplete markdown link []() with empty values. This would result in broken links in the changelog. Consider providing a default fallback text like "Unknown" for the author name and omitting the link altogether when the author is unavailable. (by chrimaho)
d1e0800: Refactor correlation tests to use pytest
- Replace .assertRaises() method with raises() function to standardise test assertions
- Update .test_is_correlated() method to expect NotImplementedError() class from is_correlated() function (by chrimaho)
01d3544: Create script to automate changelog generation
- Introduce the src/utils/changelog.py script to automate the creation of CHANGELOG.md by retrieving data via the GitHub API.
- Implement the prepare_output_file() function to ensure the document is recreated from scratch on each run.
- Define the add_page_styling() function to embed CSS for better navigation in Markdown viewers.
- Utilise the main() function to fetch releases via the .get_releases() method and filter out irrelevant commits to maintain a clean history.
- Use the Github() class to authenticate and manage repository interactions.
- Standardise the output format for release notes and commit details to improve readability. (by chrimaho)
702d902: Add continuous delivery workflow for releases
- Introduce cd.yml to automate the release cycle triggered by published GitHub releases.
- Orchestrate a multi-stage pipeline including testing, building, PyPI deployment, and documentation generation.
- Utilise uv to synchronise dependencies and manage the build process for improved performance.
- Execute the git_update_version_cli() function to handle versioning during the build stage.
- Verify package integrity using an installation matrix across multiple operating systems and Python versions.
- Automate changelog generation via the changelog.py script and push updates to the repository.
- Upload distribution assets to GitHub releases and coverage data to Codecov. (by chrimaho)
3e014f6: Increase CI job parallelism
- Increase the max-parallel limit to 30 to allow for more concurrent jobs
- Optimise the CI pipeline performance by utilising more available runners for the matrix build (by chrimaho)
cd1f7ad: Add CI workflow for automated code validation
- Implement GitHub Actions to automate validation on push and pull request events.
- Configure a check job for non-main branches to facilitate early issue detection.
- Define a matrix strategy to ensure cross-platform compatibility across various OS types.
- Support Python versions 3.9 through 3.14 to maintain broad environment stability.
- Utilise uv for efficient dependency management and execution of the validation script. (by chrimaho)
1f68794: Add Dependabot for GitHub Actions
- Initialise the configuration file to manage github-actions dependencies.
- Schedule weekly updates to ensure actions remain current.
- Assign a default reviewer and label to simplify the pull request review process. (by chrimaho)
c059c7c: Raise error for unimplemented correlation test
- Replace the None return value in the is_correlated() function with a NotImplementedError to explicitly signal that this logic is a placeholder and has not yet been implemented. (by chrimaho)
470fbfa: Clarify type ignore reasons
- Document rationale for # type: ignore on calls to the acorr_lm() function and acorr_breusch_godfrey() function
- Note that statsmodels type hints are incomplete or incompatible with internal RegressionResults() class types (by chrimaho)
8ea4c9d: Standardise equality check utility functions
- Set default value for places parameter in is_almost_equal() function and assert_almost_equal() function overloads
- Refactor is_almost_equal() function to prioritise argument validation over value equality checks
- Standardise type annotations for the params dictionary in assert_almost_equal() function
- Simplify logic for error messages in assert_almost_equal() function by defaulting to precision-based comparison (by chrimaho)
00a6029: Initialise parent classes in test suites
- Call the .setUpClass() and .tearDownClass() methods of the superclass in the BaseTester() class
- Implement the .setUp(), .tearDown(), and .tearDownClass() methods in the TestCorrelation() class to ensure proper lifecycle management (by chrimaho)
3b018b4: Remove redundant lines in config file (by chrimaho)
58163f0: Update documentation headers and module summaries
- Add header comments to the setup.py script explaining its role in unit tests.
- Standardise documentation in the errors.py module regarding error generation and data equality checks. (by chrimaho)
9c7206a: Optimise data generation and loading via caching
- Apply the @lru_cache decorator to the get_random_generator() function and various data_*() functions to reduce redundant test overhead.
- Remove the global seed() function call to favour the use of independent random generators.
- Enhance the load_airline() function with result caching to improve data retrieval performance. (by chrimaho)
891edd2: Refactor correlation logic and remove redundant code
- Remove commented-out code from the correlation() function to clean up the source.
- Update the is_correlated() function to return None instead of raising a NotImplementedError class to provide a neutral placeholder. (by chrimaho)
0d078f3: Add implementation error to the is_correlated function
The function is_correlated is defined but returns None and serves no purpose. It's marked as a placeholder in the docstring, but placeholder functions should either be removed or raise NotImplementedError to indicate they're not yet implemented. Returning None with no implementation can lead to confusion. (by chrimaho)
335045d: Refine correlation APIs and improve type safety
- Introduce @overload() function signatures for acf(), pacf(), ccf(), lm(), bglm(), and correlation() functions to improve static analysis and developer experience.
- Standardise parameter validation using Literal() class aliases for algorithm options within the acf() and pacf() functions.
- Update the ccf() function to support nlags and alpha parameters for better consistency with other correlation tools.
- Enforce keyword-only arguments for the store parameter in lm() and bglm() functions and the alpha and qstat parameters in the acf() function.
- Refine the correlation() function to act as a unified interface with specific return type hints based on the provided algorithm string.
- Ensure the lb() function return type is correctly annotated as a pd.DataFrame() class. (by chrimaho)
6176186: Validate data type in load_airline()
- Ensure the load_airline() function returns a pd.Series() class by introducing a type check
- Raise a TypeError() class if the result of the .squeeze() method is not a series to prevent errors when initialising the pd.PeriodIndex() class
- Resolve PyRight errors (by chrimaho)
d668fa7: Organise and clean up check function calls
- Remove the redundant commented out check_mypy() function call from the check() function
- Move and disable the check_mkdocs() function call to execute after the check_build() function call within the check() function (by chrimaho)
b9a9c1a: Add pyright for static type analysis
- Include pyright in the development dependencies
- Define the check_pyright() function to analyse code types
- Integrate the check_pyright() function into the validation check() function (by chrimaho)
d7b0669: Ignore complexipy cache directory
- Prevent .complexipy_cache/* from being tracked in the repository (by chrimaho)
15b4f17: Expand correlation tests and organise imports
- Import correlation() and is_correlated() functions to expand test coverage for the correlation module.
- Add .test_correlation_ccf_raises() and .test_correlation_bglm() methods to validate algorithm error handling and behaviour.
- Simplify the .setUpClass() method by removing redundant type annotations from class attributes and local variables.
- Reformat import statements to improve code structure and maintainability. (by chrimaho)
f61bbf8: Update type hints to use abstract base classes
- Replace Dict() and List() with Mapping() and Collection() classes in the generate_error_message() function to support broader input types
- Standardise the use of the built-in dict() class within the assert_almost_equal() function
- Simplify complex Union type annotations to improve code maintainability (by chrimaho)
c77a089: Introduce correlation test dispatcher
- Implement the correlation() function to provide a unified interface for various statistical correlation algorithms.
- Map string aliases to internal implementations such as the _acf(), _pacf(), and _ccf() functions.
- Utilise the generate_error_message() function to handle unsupported algorithm selections with descriptive feedback.
- Ensure the y parameter is provided when the ccf algorithm is requested via the correlation() function.
- Add a placeholder is_correlated() function to define the intended package structure. (by chrimaho)
47ff459: Refine type annotations for correlation functions
- Standardise return type signatures for the lm() function and the bglm() function.
- Introduce overload() definitions for the lm() function to improve static analysis and type safety.
- Replace np.float64, np.ndarray, and complex Union types with specific float tuples to ensure consistency. (by chrimaho)
4f783a0: Add error handling and float comparison utilities
- Add generate_error_message() function to standardise error reporting for invalid parameter options.
- Implement is_almost_equal() function to provide flexible float comparison using either decimal places or a specific delta.
- Include assert_almost_equal() function to facilitate testing by raising descriptive AssertionError messages when float values deviate beyond tolerances. (by chrimaho)
aa0c816: Disable invalid-name linting rule
- Disable the C0103 rule to standardise the project's naming convention exceptions. (by chrimaho)
8b1424e: Standardise docstrings and improve type hinting
- Define VALID_PACF_METHODS constant to centralise allowed methods for the pacf() function.
- Update docstring callouts to use !!! note and ???+ abstract admonitions for improved documentation consistency.
- Refine parameter type descriptions in docstrings to use ArrayLike and Optional types for the acf(), pacf(), ccf(), lb(), and lm() functions.
- Reformat code examples to use pycon syntax and ensure correct line continuation for better rendering.
- Clean up !!! deprecation blocks and standardise internal formatting. (by chrimaho)
02fd5d2: Document and standardise data utility module
- Add module-level docstrings and structural headers to improve navigation
- Enhance the load_airline() function with comprehensive documentation and academic references
- Remove unused numpy import to tidy the imports section
- Standardise the file layout using consistent section markers (by chrimaho)
f6824ad: Suppress specific Pylint linting warnings
- Introduce configuration to relax linting constraints
- Minimise noise by disabling checks for line length and file length
- Disable warnings for argument counts and redefined built-ins (by chrimaho)
716db6d: Expand and standardise docstring sections
- Add credit, references, see also, and deprecation sections to the configuration
- Standardise the internal key order for all section definitions to maintain consistency
- Align configuration properties to improve visual clarity and maintenance (by chrimaho)
48abfe1: Update package name and add script documentation
- Add a header comment block to the utility script to document usage and purpose.
- Update the PACKAGE_NAME constant to reflect the project identity. (by chrimaho)
9ba6915: Introduce a centralised scripts module
- Provide a centralised suite of automation utilities for project maintenance.
- Implement a run_command() function to handle shell execution and argument expansion.
- Include a lint() function to standardise code formatting across the repository.
- Define a check() function to aggregate quality assurance tests and build verification.
- Automate git workflows for versioning, tagging, and documentation deployment.
- Expose utilities via a CLI entry point to simplify development and CI/CD pipelines. (by chrimaho)
c6021f8: Introduce pre-commit hooks for code quality
- Standardise repository-wide formatting and linting rules
- Automate file-level sanitisation for whitespace and line endings
- Enforce Python style consistency using black, isort, and pyupgrade
- Prevent direct commits to the main branch via no-commit-to-branch
- Validate configuration files with check-json, check-toml, and check-yaml
- Integrate uv lockfile and synchronisation checks to maintain environment integrity
- Sanitise documentation and check spelling with blacken-docs and codespell
- Implement a local ty-check hook for type safety (by chrimaho)
ccc7aed: Add unit tests for the correlation module
- Update llvmlite and numba packages to stable versions in pyproject.toml
- Add stochastic package to the test dependencies
- Introduce BaseTester() class to standardise test data setup via the .setUpClass() method
- Implement load_airline() function to retrieve sample datasets for time series analysis
- Add unit tests for correlation functions including acf() and pacf()
- Configure uv dependency overrides to maintain numpy package version consistency (by chrimaho)
f18bc49: Fix the PACF formula clarity
The PACF formula is incorrectly formatted and unclear. The notation "Corr(Y_t, Y_{t-k} / Y_{t-1}, Y_{t-2}, ..., Y_{t-k+1})" uses division where conditional notation should be used. It should be "Corr(Y_t, Y_{t-k} | Y_{t-1}, Y_{t-2}, ..., Y_{t-k+1})" with a vertical bar (|) to denote conditioning, not a division symbol. (by chrimaho)
9de39de: Correct ACF formula in acf() function docstring
- Update the mathematical expression in the acf() function docstring to correctly show square roots in the denominator.
- Add the simplified version of the ACF formula for stationary time series to the acf() function documentation. (by chrimaho)
984fb1d: Fix typo issue
Inconsistent terminology: The documentation states "Ljung-Box and Box-Pierce statistic differ" but should use "statistics" (plural) since there are two statistics being discussed. (by chrimaho)
04dbd4b: Fix p-value formatting issue (by chrimaho)
9631202: Fix self-referential documentation issue
The text "See q_stat for more information" is circular and unhelpful since this IS the documentation for the qstat parameter. This reference should either be removed or point to relevant external documentation or the Returns section. (by chrimaho)
044f8a0: Fix typo (by chrimaho)
cbf0fb7: Fix duplicate word (by chrimaho)
feba34c: Standardise formatting of p-value in docs (by chrimaho)
ac283f5: Correct the formula provided in the lm() function documentation
- Clarify the relationship between the LM test, Engle's ARCH test, and the Breusch-Godfrey test within the lm() function
- Refine the mathematical definition of the test statistic to focus on the auxiliary regression $R^2$ , the number of observations, and degrees of freedom
- Provide a clearer step-by-step procedure for fitting the time series model and running the auxiliary regression to obtain the test statistic
- Improve the explanation of the null hypothesis and the resulting asymptotic chi-squared distribution of the LM statistic
- Standardise document formatting by using KaTeX for math equations and back-ticks for parameters like resid and nlags (by chrimaho)
1d32293: Add correlation algorithms for time series analysis
- Introduce acf() function, pacf() function, and ccf() function to compute temporal dependencies.
- Implement lb() function, lm() function, and bglm() function for residual autocorrelation testing.
- Utilise statsmodels library to ensure robust and standardised statistical results.
- Include detailed docstrings with mathematical formulas and usage examples.
- Apply @typechecked decorator to ensure runtime type safety for all algorithm inputs. (by chrimaho)
fa3e3fd: Remove duplicate function declaration (by chrimaho)
94efa8b: Initialise project structure and configuration
- Define project metadata and dependencies within pyproject.toml using uv_build.
- Configure linting, formatting, and testing tools to standardise development.
- Implement strip_ansi_codes() function in setup.py to facilitate environment-agnostic CLI output verification.
- Add test naming helper functions like name_func_flat_list() function to setup.py.
- Configure sys.path in the tests __init__.py file to ensure local module resolution.
- Extend .gitignore to exclude uv.lock, .vscode/, and other local environment directories. (by chrimaho)
b963f05: Initial commit (by chrimaho)

Changelog for ts-stat-tests🔗

v1.0.1 - Refresh documentation structure, add new pages, update package infrastructure and enhance repo security🔗

🚀 Key Highlights🔗

Core Infrastructure & Security🔗

Documentation Overhaul🔗

Mathematical & Visual Enhancements🔗

Bug Fixes & Refinements🔗

📊 Change Metrics🔗

🛠️ Internal Changes🔗

💪 Pull Requests🔗

v1.0.0 - The First Major Release: A Unified Foundation for Time Series Analysis🔗

🚀 Overview🔗

🆕 Getting Started🔗

Basic Usage🔗

🧩 Module Synopses🔗

📊 Stationarity🔗

📈 Normality🔗

📏 Linearity🔗

📉 Heteroscedasticity🔗

♻️ Correlation🔗

🔄 Seasonality🔗

🏗️ Stability🔗

🕸️ Regularity🔗

🏛️ Architectural Refactor🔗

💪 Quality Assurance🔗

✅ Checklist🔗

📊 Changes🔗

🔗 Pull Requests🔗

v0.8.0 - Add Heteroscedasticity Module🔗

🚀 Overview🔗

🛠️ Implementation details🔗

Heteroscedasticity module implementation🔗

Infrastructure and documentation🔗

Quality assurance and refactoring🔗

✅ Checklist🔗

📊 Changes🔗

🆕 Pull Requests🔗

v0.7.0 - Add Linearity Module🔗

🚀 Overview🔗

🛠️ Implementation details🔗

Linearity module implementation🔗

Documentation and navigation🔗

Quality assurance and testing🔗

✅ Checklist🔗

📊 Changes🔗

🆕 Pull Requests🔗

v0.6.2 - Add Stability Module🔗

🚀 Overview🔗

🛠️ Implementation details🔗

Stability module implementation🔗

Documentation architecture and UX🔗

Quality assurance and testing🔗

✅ Checklist🔗

📊 Changes🔗

💪 Pull requests🔗

v0.5.3 - Add Seasonality Module🔗

🚀 Overview🔗

🛠️ Implementation details🔗

Seasonality module implementation🔗

Type safety and architectural hardening🔗

Documentation and visual enhancements🔗

✅ Checklist🔗

📊 Changes🔗

💪 Pull requests🔗

v0.4.5 - Add Stationarity Module🔗

🚀 Overview🔗

🛠️ Implementation details🔗

Stationarity module implementation🔗

Data utility enhancements🔗

Documentation standardisation🔗

✅ Checklist🔗

🧪 Testing and quality assurance🔗

💪 Pull requests🔗

v0.3.1 - Add Normality Module🔗

🚀 Overview🔗

🛠️ Implementation details🔗

Normality module implementation🔗

Regularity module enhancements🔗

Documentation and progress tracking🔗

Testing and quality assurance🔗