py-cvxpy has an 'XXX: needs port for clarabel' above its tdeps.
Here's a port for that which is completely straightforward now that we
have the awesome MODPY_PYBUILD = maturin.

(The ugly HOMEPAGE regex dance is due to the fact that they don't seem
to have documentation for v0.11.1 (perhaps it's what's under /stable))

Tested on amd64 with numpy 2.2 and scikit 0.13.1 and
on arm64 with numpy 2.4 and scikit 0.16.3. All tests pass.

Without py3-clarabel, py3-cvxpy-1.6.2's tests look like this:

==== 198 failed, 931 passed, 348 skipped, 234 warnings, 12 errors in 56.98s ====

with it, they look much better.

========= 35 failed, 1126 passed, 328 skipped, 252 warnings in 49.33s ==========

This helps raise my confidence in a pending py3-cvxpy update to 1.7.5.

===

Python bindings for Clarabel.rs

Clarabel.rs is a Rust implementation of an interior point numerical
solver for convex optimization problems using a novel homogeneous
embedding.

Clarabel.rs solves linear programs, quadratic programs, second-order
cone programs and semidefinite programs. It also solves problems with
exponential, power cone and generalized power cone constraints.

Unlike interior point solvers based on the standard homogeneous
self-dual embedding (HDSE, Clarabel.rs handles quadratic objectives
without requiring any epigraphical reformulation of the objective. It
can therefore be significantly faster than other HSDE-based solvers
for problems with quadratic objective functions.

Infeasible problems are detected using a homogeneous embedding technique.

Maintainer: The OpenBSD ports mailing-list <[email protected]>

WWW: https://clarabel.org/v0.11/

Attachment: py-clarabel-0.11.1.tgz
Description: application/tar-gz

Reply via email to