ennemi: easy nearest neighbor estimation of mutual information. Mutual information (MI) can be used to find non-linear correlations between variables, and this Python 3 package is designed to fit into your data analysis workflow. This version features internal changes to how the package is distributed and tested. The documentation for this release is available at https://polsys.github.io/ennemi. This release requires at least Python 3.12 NumPy 2.0 SciPy 1.12 (Optional: pandas 2.2+) Changes since 1.5.0 The package now requires NumPy 2.0. This is in accordance to scientific Python ecosystem guidelines on dependency support. The package is tested to work with free-threaded Python 3.14. The support is still officially in "beta". The performance impact of free-threaded mode on ennemi is generally negligible: the package has already used multithreading since the heavy computation is offloaded to SciPy. Large datasets with discrete variables might see speedups. Some changes to the optional type annotations: Some parameters now have more precise types, and some no longer accept Python sequences (use NumPy or Pandas types instead). The package is tested against ty instead of mypy. The package is now built from metadata in pyproject.toml, skipping setup.py completely. There are no new features or algorithm changes. If you cannot upgrade your Python version, ennemi 1.4.0 or 1.5.0 should produce (near-)identical results. Installation This package is available on PyPI. To install/update it, executepip install --upgrade ennemion your Python installation. Contributing Your feedback is very valuable! If you encounter any problems, please [file an issue](https://github.com/polsys/ennemi/issues). Code contributions are welcomed as well.