fenicsx-pulse

fenicsx-pulse is a cardiac mechanics solver based on FEniCSx. It is a successor of pulse which is a cardiac mechanics solver based on FEniCS.

• Documentation: https://finsberg.github.io/fenicsx-pulse/
• Source code: https://github.com/finsberg/fenicsx-pulse

Install

You can install the library with pip

python3 -m pip install fenicsx-pulse

or with conda

conda install -c conda-forge fenicsx-pulse

Note that installing with pip requires FEniCSx already installed

We also provide a pre-built docker image with FEniCSx and fenicsx_pulse installed. You pull this image using the command

docker pull ghcr.io/finsberg/fenicsx-pulse:v0.7.0

Getting started

Here is a minimal example of how to use fenicsx-pulse to solve a simple cardiac mechanics problem.

import numpy as np
import dolfinx
import cardiac_geometries
import pulse

# Create a geometry with cardiac-geometries
geo = cardiac_geometries.mesh.lv_ellipsoid(
    outdir="geometry",
    create_fibers=True,
    fiber_space="Quadrature_6",
)
# Convert the geometry to a pulse.Geometry
geometry = pulse.HeartGeometry.from_cardiac_geometries(geo, metadata={"quadrature_degree": 6})

# Create a material model
material_params = pulse.HolzapfelOgden.transversely_isotropic_parameters()
material = pulse.HolzapfelOgden(f0=geo.f0, s0=geo.s0, **material_params)

# Define model for active contraction
Ta = pulse.Variable(dolfinx.fem.Constant(geometry.mesh, dolfinx.default_scalar_type(0.0)), "kPa")
active_model = pulse.ActiveStress(geo.f0, activation=Ta)

# Define model for compressibility
comp_model = pulse.Incompressible()

# Assemble into a cardiac model
model = pulse.CardiacModel(
    material=material,
    active=active_model,
    compressibility=comp_model,
)

# Define boundary conditions
traction = pulse.Variable(
    dolfinx.fem.Constant(geometry.mesh, dolfinx.default_scalar_type(0.0)), "kPa"
)
neumann = pulse.NeumannBC(traction=traction, marker=geometry.markers["ENDO"][0])


def dirichlet_bc(V: dolfinx.fem.FunctionSpace):
    # Find facets for the BASE marker
    facets = geo.ffun.find(geo.markers["BASE"][0])
    # Locate degrees of freedom for the x-component (sub(0)) on these facets
    dofs = dolfinx.fem.locate_dofs_topological(V.sub(0), geo.mesh.topology.dim - 1, facets)
    # Return the Dirichlet BC object
    return [dolfinx.fem.dirichletbc(0.0, dofs, V.sub(0))]


robin_epi = pulse.RobinBC(
    value=pulse.Variable(
        dolfinx.fem.Constant(geometry.mesh, dolfinx.default_scalar_type(1e3)),
        "Pa / m",
    ),
    marker=geometry.markers["EPI"][0],
)
robin_base = pulse.RobinBC(
    value=pulse.Variable(
        dolfinx.fem.Constant(geometry.mesh, dolfinx.default_scalar_type(1e3)),
        "Pa / m",
    ),
    marker=geometry.markers["BASE"][0],
)

bcs = pulse.BoundaryConditions(neumann=(neumann,), dirichlet=(dirichlet_bc,), robin=(robin_base, robin_epi))

# Create a mechanics problem
problem = pulse.StaticProblem(
    model=model,
    geometry=geometry,
    bcs=bcs,
)
# Perform an initial solve
problem.solve()

# Create a file for storing the solution
vtx = dolfinx.io.VTXWriter(geometry.mesh.comm, "displacement.bp", [problem.u], engine="BP4")
vtx.write(0.0)

# Assign a pressure and activation and ramp them up in steps
target_pressure = 5.0  # kPa
target_activation = 5.0  # kPa
num_steps = 5
for i, (pressure, activation) in enumerate(
    zip(np.linspace(0, target_pressure, num_steps), np.linspace(0, target_activation, num_steps))
):
    traction.assign(pressure)  # kPa
    Ta.assign(activation)  # kPa
    problem.solve()
    # Save the displacement field
    vtx.write(i + 1)
vtx.close()

A more realistic example is visualized here:

https://github.com/user-attachments/assets/8e2f5d85-3fbf-4e30-9574-22e7f718230c

Check out the demos in the documentation for more examples.

Using with Claude

This repo ships a Claude Agent Skill at .claude/skills/fenicsx-pulse/SKILL.md that teaches Claude the fenicsx-pulse composition model (CardiacModel, StaticProblem/DynamicProblem, materials, active contraction, boundary conditions, units, ...) so it can write and review code against this library more accurately.

To use it in your own project with Claude Code, copy the skill into your project's .claude/skills/ directory:

mkdir -p .claude/skills/fenicsx-pulse
curl -o .claude/skills/fenicsx-pulse/SKILL.md \
    https://raw.githubusercontent.com/finsberg/fenicsx-pulse/main/.claude/skills/fenicsx-pulse/SKILL.md

Claude will then automatically load it whenever your prompt or code touches fenicsx-pulse/pulse. To install it globally for all your projects, copy it to ~/.claude/skills/fenicsx-pulse/ instead.

Contributing

See https://finsberg.github.io/fenicsx-pulse/CONTRIBUTING.html