Question 1

How do I install scikit-fem with all optional features?

Accepted Answer

Run 'pip install scikit-fem[all]' to include meshio for mesh I/O and matplotlib for visualization. Minimal install uses only NumPy and SciPy, but examples often require extras.

Question 2

scikit-fem vs FEniCS: which is better for research?

Accepted Answer

scikit-fem is lighter and Python-focused, ideal for quick prototyping and educational use, while FEniCS offers a more comprehensive suite with C++ backend for performance. Choose based on your need for simplicity versus advanced features.

Question 3

How to visualize results in scikit-fem?

Accepted Answer

Use matplotlib via methods like mesh.plot() or save to VTK files with meshio. Visualization requires optional dependencies and is basic compared to dedicated FEM post-processors.

Question 4

Can scikit-fem handle 3D meshes and complex geometries?

Accepted Answer

Yes, it supports tetrahedral and hexahedral elements for 3D, and meshes can be loaded via meshio or generated with built-in constructors like MeshTet.init_tensor().

Question 5

How to set boundary conditions in scikit-fem?

Accepted Answer

Use the enforce() function on assembled matrices, as shown in the Poisson example. Boundary nodes are identified via mesh.boundary_nodes(), and conditions are applied by modifying the linear system.

Question 6

Is scikit-fem suitable for parallel or distributed computing?

Accepted Answer

Limited native support; optional PETSc integration allows distributed assembly via petsc4py, but it's not a core feature and requires additional setup and dependencies.

Question 7

How to cite scikit-fem in a paper?

Accepted Answer

Cite the JOSS article provided in the README, with the BibTeX entry including DOI 10.21105/joss.02369. This ensures proper attribution for academic use.

scikit-fem

What is scikit-fem?

Overview

Use Cases

Best For

Found a gem we're missing?

Not Ideal For

Pros & Cons

Pros

Cons

Frequently Asked Questions