pytristan is a Python package, based on NumPy, that provides tools for numerical solution of partial differential equations, such as

• Multi-dimensional meshes
• (coming soon) Multi-dimensional differential operators (DO), constructed using different approximation techniques, such as
  • Fourier spectral method
  • Chebyshev collocation method
  • Finite-differences method

Main features

Meshes and differential operators are NumPy arrays

The dedicated mesh's (Grid) and DO's (FourMat, ChebMat, FinDiffMat) objects subclass numpy.ndarray, benefitting from all of its powerful tools and allowing for convenient and intuitive usage.

Simple API to treat cases of special geometry

pytristan provides simplified interfaces to treat some specific cases, such as, for instance, polar geometry. Besides, there are pre-defined mapping functions in case the user aims to build a non-uniform mesh. It is also possible to define custom mappers and apply them as easily as the pre-defined ones.

For complete flexibility and control over the program, the user can opt for a more manual approach to construct the same objects using a generic interface.

Re-usage of once allocated objects

If it's necessary to re-use the same mesh or DO multiple times in the same program, pytristan spares a user the need to reconstruct them or repeatedly pass their variables between functions. Once allocated, they can be extracted from memory by calling a dedicated getter function with a simple interface anywhere in the same program.

Installation

You can install pytristan using pip:

pip install pytristan

Alternatively, the source code is available on GitHub.

Dependencies

• NumPy - the fundamental package for scientific computing with Python

License

MIT