Version 0.2.0-alpha

After one year of hard work, we, the minterpy development team, publish minterpy version 0.2.0-alpha. Despite that the code is still marked as experimental, the overall code quality is increased, a lot of bugfixes, and several new features are included compared to the previous version.

Highlights:

New feature: partial derivatives

With version 0.2.0-alpha, we can compute partial derivatives of multivariate polynomials in canonical and Newton bases.
This additional functionality also caused the core code to be changed in order to add the partial derivative as a member function to the base class of all polynomials.

New feature: support of arbitrary lp_degree

With version 0.2.0-alpha, we remove the restriction of the lp_degree to be larger than one. This gives the user more flexibility and opens a wider range of applications for interpolation.

New feature: ordinary regression

With version 0.2.0-alpha, we add functionalities for ordinary regression based on multivariate polynomials implemented in minterpy. This is the first extra feature of minterpy exported with minterpy.extras.regression. Those extra features, i.e. the namespace minterpy.extras is reserved for features, which are tightly bound to the core functionality of minterpy without changing it directly.

Minor changes

Clean up and increase code quality

Besides fixing bugs, a lot of code-cleaning work is slipped in version 0.2.0-alpha. This includes deletions of parts of the code, which are no longer needed. Especially multi_index, grid, tree, jit_utils, and several utils are slimed down to the actually used functionality. Despite that these changes are mostly internal, they increase the readability, quality and performance of the code. Furthermore, a lot of API documentation was written which lowers the learning curve for future contributors.

Enhanced presentation

With this release of version 0.2.0-alpha, we introduce a corporate design to the appearance of the minterpy repository, and documentation. Mostly, this is achieved by including of the new logo and logo banner in the documentation, repositories, README, etc.

v0.1.0-alpha

This is the initial alpha release of minterpy. It contains general structures to perform the polynomial interpolation task in multiple dimensions:

• Multivariate polynomial bases (ABC + concrete implementations)
• Base transformations
• Interpolation schemes

This code is still highly experimental and there is no issuance, that neither everything works as expected, nor if further releases will break the current API.