gp3sequencespy

gp3sequencespy is the Python implementation of gp3sequences 0.3.0, providing transparent, reproducible, and auditable analysis of ordered categorical sequences and scanpaths.

The frozen R 0.3.0 tarball is the behavioral reference. Development is parity-first: preserve the frozen scientific contracts, validate them explicitly, and only then add Python-native extensions.

Version 0.1.0 status

• 81 / 81 frozen R public function counterparts are present and exported.
• 81 / 81 frozen R public signatures have been audited; every difference is either structurally matched or explicitly classified as an R→Python semantic/plotting translation.
• 130 / 130 frozen R test_that() blocks have dedicated Python translation tests recorded in PARITY_TEST_MATRIX.md.
• The current Python suite contains 182 tests, including documentation, release-contract, and R-hclust oracle-regression tests.
• GitHub Actions covers Linux, macOS, and Windows with Python 3.11–3.14, a dedicated Ruff/mypy/frozen-contract quality job, and a fresh-wheel smoke build.
• Fifteen Python-native articles port the full frozen R vignette set, with MkDocs documentation under docs/.
• The deterministic core, hierarchical/PAM, and time-varying-model cross-language oracle tranches have been executed against frozen R 0.3.0; remaining deliberate boundaries are documented in PARITY_EXCEPTIONS.md.

Documentation: https://stefanosbalaskas.github.io/gp3sequencespy/

Version 0.1.0 is the first non-prerelease Python release candidate. Its parity claims are deliberately bounded: backend-object translations, cross-language random-number streams, Matplotlib rendering, and the validated mssm translation of the R mgcv model remain explicitly documented in PARITY_EXCEPTIONS.md.

Installation

For development installs from the repository:

uv sync --extra time

For a built wheel:

pip install gp3sequencespy-0.1.0-py3-none-any.whl

The exact 0.1.0 wheel and sdist are validated before tagging. PyPI publication remains separate from release validation and is not performed by repository workflows.

Reproducibility and release governance

• Frozen-reference details: REPRODUCIBILITY.md
• Known parity boundaries: PARITY_EXCEPTIONS.md
• Release gates: RELEASE_CHECKLIST.md
• Changelog: CHANGELOG.md
• Citation metadata: CITATION.cff
• Contribution guidance: CONTRIBUTING.md

Scientific interpretation guardrail

Sequence structure does not independently establish emotion, cognition, comprehension, personality, intention, deception, diagnosis, or causality. Observational group contrasts are associational unless a valid randomized design supports a causal interpretation.