v0.2.4

[0.2.4] - 2025-04-23

Added

• Support for FieldExtension of an arbitrary number of square roots. Does not check for relations among square roots. With finite fields, a single square root is sufficient, assuming a branch. This approach does not assume branches, but hides relations.

Changed

• Finite-field object ModP can be instantiated from a rational string, e.g. 2/3, when the prime is also specified.

• Improved parsing of GaussianRational and its string representation.

Fixed

• Fixed issue with hash of ModP and PAdic causing clashes in caches with functools.lru_cache due to hash(integer) = integer. Hashing function now hashes the string representation of the numbers.

• Fixed issue with PAdic instantiation from string, where if the prime p and the number of digits k were supplied it would fail to call the __rstr__ parser expecting a rational number even if the string was already an expansion in p.

• CI doc workflow should now fail if Sphinx autodoc fails.

v0.2.3

[0.2.3] - 2025-02-11

Changed

• Python 3.13 in CI.

Fixed

• Fixed issue with arithmetic operations between FieldExtension object and e.g. numpy.ndarrays.

v0.2.2

[0.2.2] - 2024-06-07

Added

• Continuous deployment of documentation via github pages.
• DOI.

Changed

• Python 3.10, 3.11 and 3.12 are also tested in CI.

v0.2.1

[0.2.1] - 2024-05-04

Added

• Multivariate Newton interpolation algorithm, multivariate_Newton_polynomial_interpolation.

Changed

• Improved compatibility of extended_euclidean_algorithm: output is of same type as input.
• Improved ModP and PAdic constructors to handle a wider variety of inputs, e.g. ModP('+1', 2 ** 31 - 1) is now valid.
• Splitting CI Test and Lint, adding automatic PyPI release workflow.

Fixed

• Rationalisation of $p$-adic non integers Issue 4.
• Fixed naming of extended_euclidean_algorithm (was extended_euclideal_algorithm).

v0.2.0

v0.2.0 - 2024-01-02

Added

• Univariate Newton and Thiele interpolation algorithms, Newton_polynomial_interpolation and Thiele_rational_interpolation.
• Gaussian rationals, GaussianRational, moved from lips.
• This changelog.

Changed

• vec_chained_FF_rationalize optimized for sparse tensors. New keyword optimize_for_sparse_arrays defaults to True.

Fixed

• Precision of PAdic when instantiated from negative integers proportional to the prime Issue 3.
• Recursion issue in rationalizaton ($\mathbb{F}_p \rightarrow \mathbb{Q}$) of tensors using numpy.vectorize.
• Compatibility with numpy.uint32 and numpy.uint64.

v0.1.2

v0.1.2 - 2023-04-09

Added

• Project description in README.md

v0.1.1

v0.1.1