Version 0.6

What's Changed

Typing

• Fixes wrong types in u_distance_stats_sqr.
• Add missing types in rowwise.

Documentation

• New documentation theme.
• Added links in the theory.
• Added examples to the documentation.
• Warning added to partial distance correlation/covariance docstrings by @jltorrecilla in #47

Performance

• Improve the computation time of distances for Numpy arrays, which improves performance for energy distance and the naive case of distance covariance/correlation.
• Improve AVL algorithm for distance covariance performance to bring it closer to mergesort.
• Refactor distance covariance to be able to compute distance correlation without additional calls to the covariance function.

New Contributors

• @jltorrecilla made their first contribution in #47

Full Changelog: 0.5.7...0.6

0.5.7

What's Changed

• Fix error with zero denominator. by @vnmabus in #37
• Add first typing support

Full Changelog: 0.5.6...0.5.7

Version 0.5.6

Fix: allow non-contiguous arrays again.

Version 0.5.5

What's Changed

• Parallelize the permutation test with joblib. by @lemiceterieux in #33
• Added compatibility with the Python Array Standard for most functionality.

Full Changelog: 0.5.3...0.5.5

Version 0.5.4

What's Changed

• Parallelize the permutation test with joblib. by @lemiceterieux in #33
• Added compatibility with the Python Array Standard for most functionality.

Full Changelog: 0.5.3...0.5.4

Version 0.5

Add rowwise function to optimize computation of several distance measures over different data.

Version 0.4

• Allow the selection of the method used for computing distance covariance.
• Add mergesort method for computing distance covariance.

First release with Zenodo DOI

Merge branch 'release/0.3'

Add distance correlation t test for independence in high dimension