optimagic wants to enable domain experts to solve difficult nonlinear optimizationproblems without having to become algorithm experts. Since there isno universally efficient optimizer, this requiresaccess to many different algorithms as well as diagnostic tools to compare how wellan algorithm works on a given problem.To achieve this, we wrap optimizers from SciPy, NLopt, Pygmo, and many other packagesand give them a unified and familiar minimize interface. Since our minimize functionis a superset of SciPy’s, the switch to optimagic is effortless for many users.Besides a seamless switch between different optimizers, optimagicprovides the following features:Plotting tools to visualize the history of one or several optimizers on a givenproblem.Parameters don’t have to be flat arrays. Use (nested) dictionaries, NamedTuples orwhatever is best for your problem.Simple parallelization on the algorithm level or during the calculation of numericalderivatives.Error handling and persistent logging to deal with or diagnose numerical instabilitiesin objective functions.Compatibility with automatic differentiation via JAX.For more details and examples, have a look at this tutorial.optimagic was formerly called estimagic, because it also provides functionality toperform statistical inference on estimated parameters. estimagic is now a subpackageof optimagic.Install withpip install optimagicor via conda-forge withconda install -c conda-forge optimagicFor details and instructions, check out the documentation.RoadmapWe are currently working on the following topics:Enhancement Proposal 2: Static typingThis enhancement proposal explains the adoption of static typing in optimagic.It has three major goals:Users will benefit from IDE tools such as easier discoverability of options andautocompletion.Developers and users will find code easier to read due to type hints.The codebase will become more robust due to static type checking and use of strictertypes in internal functions.The detailed enhancement proposal can be found here. It containsseveral improvements that go beyond the adoption of static typing such as a betterintegration with JAX and a better interface for least-squares problems.All breaking changes of this enhancement proposal have been implemented with version0.5.0. The remaining changes will be implemented until version 0.6.0.Enhancement Proposal 3: Alignment with SciPyThis enhancement proposal describes the changes needed to become a superset ofscipy.optimize.minimize. After it is completed, all code that was written for SciPy’sminimize function will also run with optimagic. This lowers switching costs and letsusers gradually adopt the advanced features of optimagic.The detailed enhancement proposal can be found hereAll breaking changes of this enhancement proposal have been implemented with version0.5.0. The remaining changes will be implemented until version 0.6.0.Wrapping more optimizersWe are open for requests to add any optimizer with Python bindings. We are currentlyplanning to wrap the following libraries or optimizers:Acknowledgmentsoptimagic originated in the Open Source Economics group at theUniversity of Bonn in 2019. Since 2022 it is NumFocus affiliated. Optimagic hasreceived funding from the University of Bonn, TRA Modeling,and Hoover Institution.
