Technical Details

Dependencies and Platforms

PyPEEC is entirely programmed in Python 3 and has the following dependencies:

  • SciLogger and SciSave (logging and serialization)

  • NumPy, SciPy, and Joblib (basic numerical computing libraries)

  • Shapely and Rasterio (only used for the mesher, 2D shape manipulation)

  • Pillow (only used for the mesher, 2D image manipulation)

  • VTK and PyVista (only used for the mesher, 3D shape manipulation)

Additionally, the following libraries are used for the plotter and viewer:

  • Matplotlib and QtPy (2D plots)

  • VTK, PyVista, PyVistaQt, and QtPy (3D plots)

The following optional packages can be used to speed up the solver:

  • PyAMG (Python Algebraic Multigrid)

  • MKL/FFT (available through mkl_fft)

  • MKL/PARDISO (available through Pydiso)

  • FFTW (available through pyFFTW)

  • CuPy (using GPUs through CUDA)

The following optional packages are required for Jupyter notebooks:

  • JupyterLab / Server Proxy

  • IPyWidgets / ipympl

  • Trame / Vuetify / VTK

PyPEEC is known to work with the following platforms:

  • Linux on x86/x64/X11/glibc

  • Linux on ARM64/X11/glibc

  • Apple macOS on x86/x64

  • Apple macOS on ARM64

  • MS Windows on x86/x64

Some more details on the supported platforms:

  • The main target platform of PyPEEC is Linux on x86/x64/X11/glibc.

  • PyPEEC is a pure python package and should work on any platform.

  • The following Python versions are compatible: 3.10, 3.11, 3.12, and 3.13.

  • Any CUDA and CuPy compatible GPU is supported.

Logger and Data Serialization

For the logging, PyPEEC is using SciLogger:

  • The logger configuration files can be set with the SCILOGGER environment variable.

  • More information on the logging module: scilogger.

For the serialization, PyPEEC is using SciSave:

  • The input/configuration files are either JSON or YAML files.

  • The output/data files are either JSON, MessagePack or Pickle files.

  • Pickle/MessagePack files are faster than JSON for large output/data files.

  • More information on the serialization module: scisave.

Packaging and Environment

The following files are describing the packages and dependencies:

  • Definition of the Python package: pyproject.toml.

  • Dockerfile for building the Docker image: Dockerfile.

  • Feedstock recipe for conda-forge: feedstock.

The development environment is defined in conda.yaml:

  • Contains all the required and optional dependencies.

  • Contains the packaging/development tools.

  • All the version numbers are pinned.

  • Tested with Linux x86/x64/X11/glibc.

The following scripts are used for the packaging and releasing processes:

  • scripts/run_clean.sh - Clean all the autogenerated and temporary data.

  • scripts/run_build.sh - Build the Python package and build the HTML documentation.

  • scripts/run_release.sh - Create a release (tag, release, package, and documentation).

Tests and Coverage

PyPEEC is using different tests to check for potential regressions:

  • The tests are located in the tests folder.

  • The tests are using the unittest framework.

  • The tests are covering all the main functionalities.

  • The tests are using the examples and the tutorial to check the code.

  • Only integration tests are currently implemented (no unit tests).

These files are used to run the code quality checks and the tests:

  • scripts/run_ruff.sh - Run the code quality checks.

  • scripts/run_tests.sh - Run all the integration tests.

  • scripts/run_coverage.sh - Run a code coverage analysis.

The following automated tests are performed:

  • Test the latest version of the released Conda and PyPI packages.

  • Test the Conda and PyPI environments with the current Git code.

The following platforms are considered for the automated tests:

  • Linux / Ubuntu on x86/x64

  • Microsoft / Windows on x86/x64

  • Apple / macOS on ARM64

The following Python versions are considered for the automated tests:

  • CPython 3.10

  • CPython 3.11

  • CPython 3.12

  • CPython 3.13

Contributing and Bug Report

PyPEEC is gladly accepting contributions (code, benchmark, packages, or tests). Non-code contributions (documentation, examples, or tutorials) are particularly welcomed. For large contributions, please first discuss the changes in the issue tracker.

For the bug reports, please report the following information:

  • The version of PyPEEC and Python.

  • The operating system/platform/hardware.

  • A clear and concise description of the bug.

  • A minimal working example for the bug.

  • For PyVista-related bugs, please include the pyvista.Report output.

  • For NumPy-related bugs, please include the numpy.show_config output.

  • For SciPy-related bugs, please include the scipy.show_config output.