Skip to main content

pacopy

PyPi Version PyPI pyversions GitHub stars PyPi downloads

Discord

pacopy provides various algorithms of numerical parameter continuation for ODEs and PDEs in Python.

pacopy is backend-agnostic, so it doesn't matter if your problem is formulated with NumPy, SciPy, FEniCS, pyfvm, or any other Python package. The only thing the user must provide is a class with some simple methods, e.g., a function evaluation f(u, lmbda), a Jacobian a solver jacobian_solver(u, lmbda, rhs) etc.

Install with

pip install pacopy

To get started, take a look at the examples below.

Some pacopy documentation is available here.

Examples

Basic scalar example

Let's start off with a problem where the solution space is scalar. We try to solve sin(x) - lambda for different values of lambda, starting at 0.

import math
import matplotlib.pyplot as plt
import pacopy


class SimpleScalarProblem:
    def inner(self, a, b):
        """The inner product in the problem domain. For scalars, this is just a
        multiplication.
        """
        return a * b

    def norm2_r(self, a):
        """The norm in the range space; used to determine if a solution has been found."""
        return a**2

    def f(self, u, lmbda):
        """The evaluation of the function to be solved"""
        return math.sin(u) - lmbda

    def df_dlmbda(self, u, lmbda):
        """The function's derivative with respect to the parameter. Used in Euler-Newton
        continuation.
        """
        return -1.0

    def jacobian_solver(self, u, lmbda, rhs):
        """A solver for the Jacobian problem. For scalars, this is just a division."""
        return rhs / math.cos(u)


problem = SimpleScalarProblem()

lmbda_list = []
values_list = []


def callback(k, lmbda, sol):
    # Use the callback for plotting, writing data to files etc.
    lmbda_list.append(lmbda)
    values_list.append(sol)


pacopy.euler_newton(
    problem,
    u0=0.0,
    lmbda0=0.0,
    callback=callback,
    max_steps=20,
    newton_tol=1.0e-10,
    verbose=False,
)

# plot solution
plt.plot(values_list, lmbda_list, ".-")
plt.xlabel("$u_1$")
plt.ylabel("$\\lambda$")
plt.show()

Simple 2D problem

A similarly simple example with two unknowns and a parameter. The inner product and Jacobian solver are getting more interesting.

import matplotlib.pyplot as plt
import numpy as np
import pacopy
import scipy
from mpl_toolkits.mplot3d import Axes3D


class SimpleProblem2D:
    def inner(self, a, b):
        return np.dot(a, b)

    def norm2_r(self, a):
        return np.dot(a, a)

    def f(self, u, lmbda):
        return np.array(
            [
                np.sin(u[0]) - lmbda - u[1] ** 2,
                np.cos(u[1]) * u[1] - lmbda,
            ]
        )

    def df_dlmbda(self, u, lmbda):
        return -np.ones_like(u)

    def jacobian_solver(self, u, lmbda, rhs):
        A = np.array(
            [
                [np.cos(u[0]), -2 * u[1]],
                [0.0, np.cos(u[1]) - np.sin(u[1]) * u[1]],
            ]
        )
        return np.linalg.solve(A, rhs)


problem = SimpleProblem2D()
# Initial guess and initial parameter value
u0 = np.zeros(2)
lmbda0 = 0.0

# Init lists
lmbda_list = []
values_list = []


def callback(k, lmbda, sol):
    lmbda_list.append(lmbda)
    values_list.append(sol)


pacopy.euler_newton(problem, u0, lmbda0, callback, max_steps=50, newton_tol=1.0e-10)

# plot solution
fig = plt.figure()
ax = fig.add_subplot(111, projection="3d")
ax.plot(*np.array(values_list).T, lmbda_list, ".-")
ax.set_xlabel("$u_1$")
ax.set_ylabel("$u_2$")
ax.set_zlabel("$\\lambda$")
# plt.show()
plt.savefig("simple2d.svg", transparent=True, bbox_inches="tight")
plt.close()

Bratu

Let's deal with an actual PDE, the classical Bratu problem in 1D with Dirichlet boundary conditions. Now, the solution space isn't scalar, but a vector of length n (the values of the solution at given points on the 1D interval). We now need numpy and scipy, the inner product and Jacobian solver are more complicated.

import matplotlib.pyplot as plt
import numpy as np
import pacopy
import scipy.sparse
import scipy.sparse.linalg


# This is the classical finite-difference approximation
class Bratu1d:
    def __init__(self, n):
        self.n = n
        h = 1.0 / (self.n - 1)

        self.H = np.full(self.n, h)
        self.H[0] = h / 2
        self.H[-1] = h / 2

        self.A = (
            scipy.sparse.diags([-1.0, 2.0, -1.0], [-1, 0, 1], shape=(self.n, self.n))
            / h**2
        )

    def inner(self, a, b):
        """The inner product of the problem. Can be np.dot(a, b), but factoring in
        the mesh width stays true to the PDE.
        """
        return np.dot(a, self.H * b)

    def norm2_r(self, a):
        """The norm in the range space; used to determine if a solution has been found."""
        return np.dot(a, a)

    def f(self, u, lmbda):
        """The evaluation of the function to be solved"""
        out = self.A.dot(u) - lmbda * np.exp(u)
        out[0] = u[0]
        out[-1] = u[-1]
        return out

    def df_dlmbda(self, u, lmbda):
        """The function's derivative with respect to the parameter. Used in Euler-Newton
        continuation.
        """
        out = -np.exp(u)
        out[0] = 0.0
        out[-1] = 0.0
        return out

    def jacobian_solver(self, u, lmbda, rhs):
        """A solver for the Jacobian problem."""
        M = self.A.copy()
        d = M.diagonal().copy()
        d -= lmbda * np.exp(u)
        M.setdiag(d)
        # Dirichlet conditions
        M.data[0][self.n - 2] = 0.0
        M.data[1][0] = 1.0
        M.data[1][self.n - 1] = 1.0
        M.data[2][1] = 0.0
        return scipy.sparse.linalg.spsolve(M.tocsr(), rhs)


problem = Bratu1d(51)
# Initial guess and parameter value
u0 = np.zeros(problem.n)
lmbda0 = 0.0

lmbda_list = []
values_list = []


def callback(k, lmbda, sol):
    # Use the callback for plotting, writing data to files etc.
    fig = plt.figure()
    ax1 = fig.add_subplot(111)
    ax1.set_xlabel("$\\lambda$")
    ax1.set_ylabel("$||u||_2$")
    ax1.grid()

    lmbda_list.append(lmbda)
    # use the norm of the currentsolution for plotting on the y-axis
    values_list.append(np.sqrt(problem.inner(sol, sol)))

    ax1.plot(lmbda_list, values_list, "-x", color="C0")
    ax1.set_xlim(0.0, 4.0)
    ax1.set_ylim(0.0, 6.0)
    plt.close()


# Natural parameter continuation
# pacopy.natural(problem, u0, lmbda0, callback, max_steps=100)

pacopy.euler_newton(
    problem, u0, lmbda0, callback, max_steps=500, max_num_retries=10, newton_tol=1.0e-10
)

Ginzburg–Landau

https://user-images.githubusercontent.com/181628/146639709-90b6e6aa-48ba-418d-9aa4-ec5754f95b93.mp4

The Ginzburg-Landau equations model the behavior of extreme type-II superconductors under a magnetic field. The above example (to be found in full detail here) shows parameter continuation in the strength of the magnetic field. The plot on the right-hand side shows the complex-valued solution using cplot.

Metadata

Release files for pacopy 0.2.30

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Built distributions (wheels)

Table of built distributions (wheels) for pacopy 0.2.30
File Actions
pacopy-0.2.30-cp315-none-any.whl CPython 3.15 none any Details for pacopy-0.2.30-cp315-none-any.whl
pacopy-0.2.30-cp314-none-any.whl CPython 3.14 none any Details for pacopy-0.2.30-cp314-none-any.whl
pacopy-0.2.30-cp313-none-any.whl CPython 3.13 none any Details for pacopy-0.2.30-cp313-none-any.whl
pacopy-0.2.30-cp312-none-any.whl CPython 3.12 none any Details for pacopy-0.2.30-cp312-none-any.whl
pacopy-0.2.30-cp311-none-any.whl CPython 3.11 none any Details for pacopy-0.2.30-cp311-none-any.whl
pacopy-0.2.30-cp310-none-any.whl CPython 3.10 none any Details for pacopy-0.2.30-cp310-none-any.whl

Total release size: 245.1 kB

Release files / pacopy-0.2.30-cp315-none-any.whl

Download URL pacopy-0.2.30-cp315-none-any.whl
Size 44.9 kB
Tags CPython 3.15
SHA-256 checksum
How to use checksums
b07dbc02d625189d2d0af67d7fa50f72d26bc5919be503d027287e06b1f1464c
BLAKE2b-256 checksum
How to use checksums
00c3d5434712b1be1504e139d1c045bc6c723fa09c8bde4a1f187ce163ccc470
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 9, 2026.

Transparency log

Release files / pacopy-0.2.30-cp314-none-any.whl

Download URL pacopy-0.2.30-cp314-none-any.whl
Size 44.6 kB
Tags CPython 3.14
SHA-256 checksum
How to use checksums
d1cc076b4558a7201b493b6e7f9e99a11a91c4c78da07e4f6e28536d2d571c9a
BLAKE2b-256 checksum
How to use checksums
547abbf2794af24f93c77ae2438836417781ad012b465b4209f0299e757ae8fe
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 9, 2026.

Transparency log

Release files / pacopy-0.2.30-cp313-none-any.whl

Download URL pacopy-0.2.30-cp313-none-any.whl
Size 41.0 kB
Tags CPython 3.13
SHA-256 checksum
How to use checksums
a48fb0ee69b9438d90d59227a719f1bd4ab8a61697aff64638b959305ed0d2cb
BLAKE2b-256 checksum
How to use checksums
c30f0158569ce97f778a69ed937492b512cb966d81bb8337199e7606fbb09a35
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 9, 2026.

Transparency log

Release files / pacopy-0.2.30-cp312-none-any.whl

Download URL pacopy-0.2.30-cp312-none-any.whl
Size 40.6 kB
Tags CPython 3.12
SHA-256 checksum
How to use checksums
0e4240f333671c743df2d200232fbab54c1c7ffe3d258a5014e3492c886bb00b
BLAKE2b-256 checksum
How to use checksums
2c1a155617d66e756f621b364590501ce43f19cd36a92cbbbaefbbd96c2fe64b
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 9, 2026.

Transparency log

Release files / pacopy-0.2.30-cp311-none-any.whl

Download URL pacopy-0.2.30-cp311-none-any.whl
Size 44.0 kB
Tags CPython 3.11
SHA-256 checksum
How to use checksums
28b14dae05b9ed729cc80aa3373d09cd00cdce6862f2847674596b39ab92dc51
BLAKE2b-256 checksum
How to use checksums
06636f6e5a0d43aaf298fca0d1041c819a959f673c62e160a5c158aa9272dd46
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 9, 2026.

Transparency log

Release files / pacopy-0.2.30-cp310-none-any.whl

Download URL pacopy-0.2.30-cp310-none-any.whl
Size 29.9 kB
Tags CPython 3.10
SHA-256 checksum
How to use checksums
faa559a1d65de3836d7b7f796444d7e8d8688b239d2750da55ba8a7802b3489e
BLAKE2b-256 checksum
How to use checksums
574dbe4ce0a570a5240d2796bcd8cfadd8e11604a15908b616c4f909e122b014
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 9, 2026.

Transparency log

Security for pacopy 0.2.30

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

All 6 files in this release have been uploaded with provenance metadata.

Select from 6 files

Attestation for pacopy-0.2.30-cp315-none-any.whl

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 9, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page