All notable changes to this project will be documented in this file.
- Migrated from
setup.pytopyproject.tomlfor modern Python packaging. - Updated to follow standard Semantic Versioning (three-part version).
- Improved documentation and clarified usage examples for crystal structure exploration workflows.
- Minor packaging and dependency fixes for smoother local editable installs.
- Preliminary support for compositional space exploration with explicit DFT and interatomic potential integration.
- Early support for parallel or batch candidate evaluations.
- Refined Thompson Sampling and Expected Improvement (EI) acquisition function implementations for more robust convergence.
- Initial integration with external crystal structure prediction modules
- New plotting utilities to visualise acquisition surfaces and posterior distributions.
- Support for multiple acquisition strategies: Expected Improvement (EI) and Thompson Sampling.
- Improved Gaussian Process model initialisation for better uncertainty handling in high-dimensional compositional spaces.
- Basic Bayesian optimisation loop for accelerated compositional exploration.
- Early demonstration examples on phase fields.
- Initial prototype release.
- Core Bayesian optimisation engine with Gaussian Process support.
- Simple compositional candidate generator and initial evaluation interface.
- Support for multi-objective acquisition strategies.
- Integration with active learning for iterative interatomic potential refinement.
- Additional crystal structure filtering and ranking tools.