The dolfiny package provides a set of high-level convenience wrappers for DOLFINx, the FEniCS library.

While DOLFINx involves many low-level operations - from handling ghosted values to interfacing PETSc solvers directly - dolfiny serves as a wrapper around the low-level functionality of DOLFINx and is meant to combine a user-oriented API with the performance and functionality of DOLFINx, FFCx, Basix and UFL.

This library is written exclusively in Python with optional interfacing to user-provided C++ kernels.

pip install dolfiny

Certain functionality (see demos and unit tests) relies on the availability of external packages such as

• Matplotlib (plotting),
• PyVista (scientific visualisation), or
• cppyy (dynamic Python/C++ bindings).

Install dolfiny with these dependencies by running

pip install dolfiny[all]

For ARM-based architectures (aarch64/arm64 on Linux) we recommend to fallback to our custom-compiled binary wheels for vtk and cppyy-cling by setting the local package index repository in the respective environment variable

export PIP_INDEX_URL=https://gitlab.com/api/v4/projects/59503118/packages/pypi/simple

before calling pip.

You may also check the Dockerfile for an up-to-date version of the installation process.

Multi-arch (amd64 and arm64) Docker images with pre-installed dolfiny (and dependencies) are available at DockerHub.

docker pull dolfiny/dolfiny

Some of our demos are available as documented and rendered jupyter notebooks on our website at dolfiny.uni.lu.

Beside this, other references are demos or unit tests.

Presentations about dolfiny functionality:

• dolfiny: Convenience wrappers for DOLFINx at FEniCS 2021 conference,
• Nonlinear analysis of thin-walled structures based on tangential differential calculus with FEniCSx at FEniCS 2022 conference,
• Nonlinear local solver at FEniCS 2022 conference,
• Non-standard formulations of 3D elasticity with FEniCSx at FEniCS 2024 conference,
• PETSc TAO support for optimisation problems with FEniCSx at FEniCS 2025 conference,
• Physical units in UFL at FEniCS 2025 conference,
• Dimensional analysis in UFL at FEniCS 2025 conference,
• Sequential quadratic and convex optimisation with FEniCSx and PETSc at KLAIM 2025 conference.

• Michal Habera, University of Luxembourg / Rafinex, Luxembourg
• Paul T. Kühner, University of Luxembourg, Luxembourg
• Andreas Zilian, University of Luxembourg, Luxembourg

We are always looking for contributions and help with dolfiny. If you have ideas, nice applications or code contributions then we would be happy to help you get them included. We ask you to follow the FEniCS Project git workflow.

For more information and opportunities to get involved, see here.

Please use the GitHub issue tracker to report any issues.

dolfiny is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

dolfiny is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with dolfiny. If not, see http://www.gnu.org/licenses/.