New release [0.3.2] - 2025-07-21

Changed

• Documentation updated following Arlin Stoltzfus feedback and comments. Motivation and history of the package
  added to the introductory page together with links to other studies using gpmap-tools and online talks. Getting started
  section shows the basic tasks one can do with gpmap-tools. Installation instructions simplified to use pip alone for users and download github repository and test for developers. Docstrings updated for completeness for API in rtd.

New release [0.3.1] - 2025-03-31

Added

• Computation of posterior covariance under the Minimum epistasis interpolation model. If a is not provided by the user, it can be inferred from the MEI solution before computing the posterior covariance. Uniqueness of the solution is now checked before calculations are done.
• Gaussian likelihood implemented for consistency and sampling from the likelihood when simulating data as in SeqDEFT

Changed

• Generalization of Gaussian process class so that all the models have the same set of basic methods under the more general class, including sample_prior, simulate, fit, predict and make_contrast.

• Homogeneization of inference classes to all work under the same logic with the same methods for defining and fitting hyperparameters, making predictions, computing the posterior of linear combinations and sampling from the prior.

This first release includes functionality to:

• Infer sequence-function relationships from experimental data
• Infer sequence-function relationships from natural sequences
• Generate visualization coordinates for complete combinatorial landscapes
• Plot those visualizations