ngstents 0.1.dev404

Spacetime tent facilities for solving hyperbolic equations

Spacetime Tents in NGSolve

This ngstents package is a c++ extension of the NGSolve finite element library, designed to ease experimentation with solvers based on spacetime tents for hyperbolic systems. A python front-end allows new equations (linear or nonlinear conservation laws) to be solved by easily defining required fluxes and numerical fluxes in a few lines of code.

Install

Binary installers are available for linux, mac, and windows (with python >= 3.9).

• pip install ngstents

But note that the binaries are not generally kept up to date. The recommended option is to build from source (see below).

Build

If you built NGSolve from source, you can build and install ngstents. After cloning this repository, compile the c++ code here and install:

• cd src
• mkdir build && cd build
• cmake -DNGSolve_DIR=<Path2YourNGSolveInstallCMake> ..
• make install

Use

To start using this code, import the module after installation:

import ngstents

It contains classes named Tent and TentSlab used to partition spacetime into tents. It also has a submodule named conslaw which contains specific pre-programmed hyperbolic equations and also facilities to define and solve new hyperbolic systems using the python interface. Use python's help system to query for documentation on each object. The demo folder contains example scripts. The docs folder contains more documentation, explained next.

Documentation

Read the docs online. This documentation is generated from hands-on style tutorial notebooks in the docs folder.

Offline, to build and test the docs locally, navigate to the ngstents/docs folder and run these:

• pip install -r requirements.txt
• sphinx-build -b html . _build_docs

You can then open the docs from ngstents/docs/_build_docs/INDEX.html.

Check

If you extend the code, your new code changes can be cross-checked against a test suite provided in the 'tests' folder. For example, if you have pytest installed, move to the ngstents/tests folder and use pytest:

• pytest .

Organization

• src: c++ source code
• py: python packaging files
• doc: tutorials and explanations
• demo: examples of applications
• tests: test suite