Roadmap

v0.0.x

• Initial project setup
• Implement & optimize unconstrained sampling (randint)
• Implement benchmarking

v0.1.0

• First version with constrained sampling (randint_constrained) + benchmarking functionality

v0.1.x

• Tweaks & optimizations to constrained sampling

v0.2.0

• First version of the MaxDivSolver framework all core classes (metrics, distances, constraints, solver state, ...)

v0.2.x

• Implement all remaining parts of framework
• multi-component prioritized constraints
• soft constraint handling
• tie-breaker diversity scores
• full MaxDivSolverBuilder class for MaxDivSolver instances
• probability selectivity modification functionality
• support for SolverSteps with scheduled (=progress-dependent) parameters

v0.3.0

• First version of the MaxDivSolver framework & all core classes (metrics, distances, constraints, solver state, scheduling, ...)

v0.3.x

• Implementation of various initialization & optimization strategies
• Implementation of convenient strategy presets in MaxDivSolverBuilder class.

v0.4.0

• First version with full-featured MaxDivSolver, including a variety of initialization & optimization strategies.

v0.4.x

• Additional tweaks and completions
• add 'smart' presets
• add preset benchmarking

v0.5.0

• First version with finalized presets & optimization strategies with satisfactory generic performance characteristics.

v0.5.x

• Documentation push
• Getting started / Tutorial
• Use cases with benchmarks & visualizations
• Package architecture

v0.9.0

• First full-featured release with full docs

v0.9.x

• Bug fixes & minor improvements based on initial usage

v1.0.0

• First stable release
• To be released before downstream use cases start relying on the package