Almost a year has passed since Desbordante’s last release, and now we’re rolling out a new one — 2.5.0.

Desbordante is an open-source, high-performance data profiler that finds and validates complex patterns in data using a bunch of advanced algorithms.

Unlike the black-box ML approaches that dominate today, Desbordante takes a symbolic, explainable approach to data analysis, producing explicit rules or rule-shaped artifacts that are verifiable, such as data dependencies, association rules, or other patterns. It’s a different way to analyze data, and it works well alongside black-box models. Unlike free-form model output, these patterns have precise semantics and can be validated algorithmically against the data. And, importantly, no hallucinations or misclassification.

This release adds 9 new patterns, including support for another dataset type: sequences in data. That brings the total to 4. We also picked up Python 3.15, which just came out. On top of the new patterns, there was a lot of infrastructure work too — mostly invisible to users, but important.

Full changelog: https://github.com/Desbordante/desbordante-core/releases/tag...