v0.5.3

Summary

This release introduces multi-backend plotting with support for matplotlib, bokeh, and altair backends (in addition to the existing plotly backend), and several new optimizer libraries including PySwarms, PyGAD, and gradient-free-optimizers. It also adds lazy loading for optional dependencies to improve import times.

Many contributions in this release were made by Google Summer of Code (GSoC) 2025 contributors.

Pull Requests

• #665 Skips nag_dfols tests when DFO-LS is not installed (@Swayam-maurya).
• #664 Adds from __future__ import annotations to constraints.py to fix annotations issue with Python 3.13 and NumPy 2.4 (@timmens).
• #660 Renames the bayes_opt parameter n_iter to stopping_maxiter (@spline2hg).
• #659 Removes None as a valid option for stopping_criterion in convergence_plot and updates the docstring (@szd5654125).
• #658 Enhances documentation and minor fixes in backend plotting (@r3kste).
• #654 Implements the altair plotting backend (@r3kste).
• #653 Adds llms.txt and llms-full.txt to documentation (@mostafafaheem).
• #652 Implements the bokeh plotting backend (@r3kste).
• #649 Implements backend plotting for slice_plot (@r3kste).
• #647 Implements backend plotting for convergence_plot (@r3kste).
• #645 Implements backend plotting for profile_plot (@r3kste).
• #644 Adds a how-to guide for changing plotting backends (@r3kste).
• #643 Skips doctest that fails due to negative signed zero handling (@r3kste).
• #641 Implements backend plotting for params_plot (@r3kste).
• #639 Adds optimizers from PySwarms (@spline2hg).
• #637 Adds note about __future__ import (@spline2hg).
• #636 Wraps population-based optimizers from gradient-free-optimizers (@gauravmanmode).
• #633 Migrates bayesian-optimizer docs to new documentation style (@spline2hg).
• #632 Migrates nevergrad optimizers to new documentation style (@gauravmanmode).
• #631 Migrates iminuit docs to new documentation style (@spline2hg).
• #624 Wraps local optimizers from gradient-free-optimizers (@gauravmanmode).
• #621 Implements lazy loading for optional dependencies (@spline2hg).
• #619 Adopts the NumFOCUS code of conduct (@timmens).
• #616 Adds optimizers from PyGAD (@spline2hg).
• #600 Separates data preparation and plotting for criterion_plot() (@r3kste).
• #599 Implements the matplotlib backend for criterion_plot() (@r3kste).
• #581 Adds 3D visualizations of optimization problems (@shammeer-s).
• #554 Improves documentation of algorithm options (@janosg).