tabmat 4.2.1

Bug fixes:

• Fixed methods _cross_categorical, _cross_sandwich, multiply, tocsr of CategoricalMatrix and function _sandwich_cat_cat_limited_rows_cols to operate on read-only buffers as well.

Full Changelog: 4.2.0...4.2.1

tabmat 4.2.0

4.2.0 - 2026-01-13

New feature:

• tabmat.from_formula now also supports any data frame supported by narwhals (including polars dataframes).

Other changes:

• tabmat now requires Python>=3.10.

4.1.5

4.1.5 - 2025-12-17

Other change:

• Restore building wheels for Intel-based macOS systems.

tabmat 4.1.4

4.1.4 - 2025-12-16

Other changes:

• tabmat.from_df now avoids unnecessary copies of dense arrays, but still ensures that the results are contiguous (C or F order).
• We now use narwhals' v2 API for data frame handling.

Full Changelog: 4.1.3...4.1.4

tabmat 4.1.3

4.1.3 - 2025-11-12

Bug fixes:

• Fixed CategoricalMatrix.transpose_matvec to operate on read-only buffers as well.
• Fixed incorrect calculation of the shape of a CategoricalMatrix when initialized with zero categories and drop_first=True.

tabmat 4.1.2

4.1.2 - 2025-07-17

Bug fixes:

• Fixed a bug which caused issues when constructing tabmat matrices from existing ModelSpecs when they contained categorical columns with all levels dropped.
• We can now treat dedicated pandas string series, which are the defaults for strings since pandas 2.3, as categoricals.

Full Changelog: 4.1.1...4.1.2

tabmat 4.1.1

4.1.1 - 2025-01-30

Bug fix:

• A more robust :meth:DenseMatrix._get_col_stds results in more accurate :meth:StandardizedMatrix.sandwich results.

Other changes:

• Build wheel for pypi on python 3.13.
• Build and test with python 3.13 in CI.

Full Changelog: 4.1.0...4.1.1

tabmat 4.1.0

4.1.0 - 2024-11-07

New feature:

• Added a new function, :func:tabmat.from_df, to convert any dataframe supported by narwhals into a :class:tabmat.SplitMatrix.

Other changes:

• Allow :class:CategoricalMatrix to be initialized directly with indices and categories.
• Added checks for dimension and dtype mismatch in :meth:MatrixBasesandwich.sandwich.

Bug fix:

• Fixed a bug in :meth:tabmat.CategoricalMatrix.standardize that sometimes returned nan values for the standard deviation due to numerical instability if using np.float32 precision.

Full Changelog: 4.0.1...4.1.0

tabmat 4.0.1

4.0.1 - 2024-06-25

Other changes:

• Removed reference to .A attribute and replaced it with .toarray()
• Add support between formulaic and pandas 3.0
• Support pypi release for numpy 2.0

Full Changelog: 4.0.0...4.0.1

tabmat 4.0.0

4.0.0 - 2024-04-23

Breaking changes:

• To unify the API, :class:DenseMatrix does not inherit from :class:np.ndarray anymore. To convert a :class:DenseMatrix to a :class:np.ndarray, use :meth:DenseMatrix.unpack.
• Similarly, :class:SparseMatrix does not inherit from :class:sps.csc_matrix anymore. To convert a :class:SparseMatrix to a :class:sps.csc_matrix, use :meth:SparseMatrix.unpack.

New features:

• Added column name and term name metadata to :class:MatrixBase objects. These are automatically populated when initializing a :class:MatrixBase from a :class:pandas.DataFrame. In addition, they can be accessed and modified via the :attr:MatrixBase.column_names and :attr:MatrixBase.term_names properties.
• Added a formula interface for creating tabmat matrices from pandas data frames. See :func:tabmat.from_formula for details.
• Added support for missing values in :class:CategoricalMatrix by either creating a separate category for them or treating them as all-zero rows.
• Added support for handling missing categorical values in pandas data frames.

Bug fix:

• Added cython compiler directive legacy_implicit_noexcept = True to fix performance regression with cython 3.

Other changes:

• Refactored the pre-commit hooks to use ruff.
• Refactored :meth:CategoricalMatrix.transpose_matvec to be deterministic when using OpenMP.
• Adjusted transformation to sparse format in :func:tabmat.from_pandas to future changes in pandas.

What's Changed

• Pre-commit autoupdate by @quant-ranger in #351
• Bump pypa/gh-action-pypi-publish from 1.8.11 to 1.8.12 by @dependabot in #352
• Bump pypa/gh-action-pypi-publish from 1.8.12 to 1.8.14 by @dependabot in #353
• Bump pypa/cibuildwheel from 2.16.5 to 2.17.0 by @dependabot in #354
• Pre-commit autoupdate by @quant-ranger in #355
• Set fill value of sparse columns in line with SparseDtype.subtype by @MatthiasSchmidtblaicherQC in #356
• Tabmat v4 alpha by @MartinStancsicsQC in #286

Full Changelog: 3.1.14...4.0.0