Python API

Browse the modules below for classes, functions, and signatures. Use Get started for installation and a first model.

Compile and run

This example needs only Eqiora and NumPy, which Eqiora installs.

python

import eqiora

model = eqiora.compile(source="""
model decay(parameter rate: 1 / s = 1) {
    state x: 1;
    initial { x = 1; }
    relation flow { derivative(x) + rate * x = 0; }
}
""", filename="decay.eqi")
x = model.field("x")
plan = eqiora.resolve(
    model,
    temporal=eqiora.time.Tsitouras45(
        initial_step_s=0.01,
        relative_tolerance=1e-9,
        absolute_tolerances={(x, 0, 0, False): 1e-11},
    ),
)
result = eqiora.run(
    plan,
    state=eqiora.State.initial(plan),
    until_s=1.0,
    output_times_s=(0.25, 0.5, 1.0),
)
series = result.series(x)
for time_s, value in series:
    print(f"{time_s:.2f} s: {value:.6f}")

The three values are approximately 0.778801, 0.606531, and 0.367879: x is dimensionless, time is in seconds, and the equation gives exponential decay. compile defines the mathematics, resolve chooses its numerical method, and run computes the requested observations.

Inspect, edit, and save

Continue with the objects above:

python

print(model.parameter("rate").value)
print(model.render_equations("flow", profile="plain")[0].text)

faster = model.commit(model.preview_value_edit("rate", 2.0))
print(model.parameter("rate").value, faster.parameter("rate").value)

import numpy as np
np.savetxt(
    "decay.csv",
    np.column_stack((series.time.numpy(), series.values.numpy())),
    delimiter=",", header="time_s,x", comments="",
)

The original rate remains 1.0; the new Model has rate 2.0. Resolve the new Model to simulate it. The CSV contains the observations from the original run, and render_equations prints the model’s equations in readable form.

For spatial models, continue with geometry, meshing, and the complete modeling examples. For gradients, see differentiation.

Modules

Module	Summary
eqiora	Model authoring, numerical plans, execution, and results.
eqiora.geometry	Build geometry and select its regions and boundaries.
eqiora.lang	Author Eqiora Modules, Components, expressions, and equations in Python.
eqiora.units	Shared units for input quantities and explicit presentation.
eqiora.data	Explicit numeric adapters and canonical resolved-array artifacts.
eqiora.meshing	Plan and generate meshes from geometry.
eqiora.formulation	Choose a mathematical formulation and inspect the resolved choice.
eqiora.fem	Finite-element discretization policies.
eqiora.fvm	Finite-volume discretization policies.
eqiora.solve	Linear, nonlinear, and eigenvalue solver settings.
eqiora.time	Time-integration methods and related settings.
eqiora.fluid	Configure incompressible flow and inspect its results.
eqiora.trajectory	Read spatial trajectories and derived fields.
eqiora.solid	Configure linear elasticity and inspect structural results.
eqiora.viewer	Interactive, read-only views of geometry, meshes, and fields.
eqiora.colab	Prepare a Google Colab runtime for Eqiora’s notebook viewer.
eqiora.jupyter	IPython extension that compiles %%eqiora notebook cells.
eqiora.matplotlib	Plot Eqiora results with Matplotlib.
eqiora.diff	Implicit differentiation of Eqiora programs.
eqiora.torch	Use Eqiora differentiable programs with PyTorch autograd.
eqiora.jax	Use Eqiora differentiable programs with JAX.