How molmod/stacie works
Project type and purpose
• STACIE is a Python package and command-line algorithm library to compute time integrals of autocorrelation functions, primarily for post-processing molecular-dynamics simulation data.
• It has four main “faces”:
a. Core library modules (src/stacie)
b. CLI/tools (tools/*.py)
c. Documentation (docs/)
d. CI & release pipelines (.github/, pyproject.toml)
High-level components
• Core Library
• Modules under src/stacie: conditioning, cost, cutoff, estimate, model, plot, spectrum, synthetic, utils
• Exposes functions and classes for computing integrals, modeling, and plotting results
• CLI & Helper Scripts
• tools/gensynth.py, quick.py, logo.py – example/script entry points to generate synthetic data or quick runs
• tests/inputs – sample data for regression tests
• Documentation
• docs/source – Sphinx sources, examples, tutorials, theory, references
• docs/compile_*.sh, docs/release_docs.sh – build scripts (HTML/PDF)
• Packaging & Publishing
• pyproject.toml – build configuration for setuptools/poetry
• CITATION.cff, LICENSE files
• Continuous Integration & Release
• .github/workflows/pytest.yaml – test suite on push/PR
• .github/workflows/release.yaml – release to PyPI/Zenodo
• .github/scripts/extract-notes.sh – release-note automation
Interactions and Data Flows
• User → CLI/tools → Core Library
• Core Library ↔ Data Inputs
• Data (numpy .npz) flows into spectrum module, processed through model/cutoff/estimate pipeline
• Core Library → Plot module → Output (graphs/files)
• Tests (pytest) → Core Library & CLI
• tests/conftest.py sets up fixtures, test_*.py validate each module
• Documentation Builder → docs/source → Sphinx → HTML/PDF → GitHub Pages
• CI Pipelines → run tests, build docs, publish package
Architectural Patterns & Design Principles
• Modular design: each concern (estimation, modeling, plotting) in its own module
• Separation of concerns: core logic vs. CLI vs. docs vs. CI
• Extensibility: new estimators or models can plug into core via utils and synthetic data functions
• Standard Python packaging and CI best practices