Abstract
Modepy is a Python library for defining reference elements, equipping them with appropriate approximation spaces, and numerically performing calculus operations (derivatives, integrals) on those spaces. It is written in pure, type-annotated Python 3, offering comprehensive documentation and minimal runtime dependencies (mainly NumPy). modepy focuses on high-order accuracy-given an element size ℎ, this refers to the asymptotic decay of the approximation error as 𝑂(ℎ 𝑛), for 𝑛 ≥ 3, assuming sufficient smoothness of the solution being approximated. For a problem in 𝑑 dimensions, the number of unknowns scales as 𝑂(ℎ-𝑑). Therefore, if accuracy is desired at manageable cost, high-order methods are crucial.

This publication has 14 references indexed in Scilit: