Development

Project aims

Aestimo is developed as an educational and research tool. Its source should be readable enough to follow a calculation, while numerical methods should be checked for accuracy and convergence. Python input files let users describe layers, material parameters and numerical controls with ordinary Python statements; the desktop interface offers JSON projects for the same broader simulation work.

Contributions to the solver, material database, examples, tutorials and interface are welcome. The source and issue tracker are hosted on GitHub. Development discussion takes place on the Aestimo development group.

Contribute a change

Open an issue to discuss a substantial feature or scientific-model change. For a bug, include the input, Aestimo version, full error and relevant output.
Fork the repository and make a focused change. Explain the resulting behavior and update examples or documentation where needed.
Run the relevant checks. For solver changes, compare representative numerical outputs and include the physical assumptions, units and convergence evidence.
Submit a pull request to aestimosolver/aestimo, including the commands used to verify it. Documentation corrections can be small standalone contributions.

Set up a checkout

Clone your fork or the upstream repository, then install it in an activated virtual environment:

git clone https://github.com/aestimosolver/aestimo.git
cd aestimo
python -m pip install -e .

See Installation for Python and desktop requirements. Editable installation uses the checkout’s code. Package metadata, dependencies, data-file declarations and launch commands are defined in pyproject.toml.

Run software checks

From the checkout root on Linux/macOS:

MPLBACKEND=Agg python -m unittest discover -s examples -p 'test_*.py'

On Windows PowerShell:

$env:MPLBACKEND = "Agg"
python -m unittest discover -s examples -p 'test_*.py'

The runtime workflow checks installed packages on Linux/Python 3.9, 3.12 and 3.13 and Windows/Python 3.12. Linux jobs also run the example regression suite. The installed-package checker verifies launch commands, a representative CLI calculation, plotting and bundled resources without opening a GUI window. See Validation and reference data for interpreting numerical and reference comparisons.

Build the documentation

python -m pip install -r docs/requirements.txt
python -m sphinx -n -W --keep-going -b html docs docs/_build/html

Open docs/_build/html/index.html. The build uses Python 3.12 and renders Markdown with MyST. It does not import the solver or run simulations. Read the Docs uses the configuration in .readthedocs.yaml at the repository root and the pinned dependencies in docs/requirements.txt.

The public guide covers installation, calculations and project information. Release checklists and historical integration records are maintained separately in the repository for contributors.

Build distributions

python -m pip install build twine
python -m build
python -m twine check dist/*

Install the resulting wheel in a fresh environment outside the checkout and run python tools/check_release_install.py --source . with that environment’s interpreter. A local build does not upload the package or create a release. The release preparation checklist records the remaining publication steps.

Background reading

Background references on quantum confinement, computational physics and semiconductor-device simulation:

Paul Harrison, Quantum Wells, Wires and Dots.
Morten Hjorth-Jensen, Computational Physics, University of Oslo lecture notes (2007 and later notes).
Jaan Kiusalaas, Numerical Methods in Engineering with Python, Cambridge (2010).
Jonathan Edward Moussa, The Schrodinger-Poisson Self Consistency in Layered Quantum Semiconductor Structures, MSc thesis (2003).
Nuno Sucena Almeida, Development of a Multi-physics Simulation Framework for Semiconductor Materials and Devices, thesis (2005).
Jeffery Wayne Allen, A Simulation Tool Suite for the Modelling and Optimization of Multiple Quantum Well Structures, dissertation (2005).