Changelog

All notable changes to this project are documented here.

[0.1.1] - 2025-11-03

bbstat.evaluate.BootstrapSummary: summarises a bootstraped distribution of estimates.
bbstat.evaluate.BootstrapDistribution: container for a bootstraped distribution of estimates as replacement for bbstat.evaluate.BootstrapResult.
bbstat.plot.plot uses a new optional parameter precision to control rounding.

Moved bbstat.evaluate.BootstrapResult.ndigits to bbstat.utils.get_precision_for_rounding and changed parameter from interval bounds to width.
Moved bbstat.evaluate.credibility_interval to bbstat.utils.compute_credibility_interval.
Moved bbstat.evaluate.BootstrapResult.plot to bbstat.plot.plot.
Renamed all instances "credibility" -> "credible" and "coverage" -> "level".
bbstat.bootstrap.bootstrap returns bbstat.evaluate.BootstrapDistribution.
bbstat.plot.plot expects boostrap_distribution: bbstat.evaluate.BootstrapDistribution as as first and level: float as (no longer optional) second parameter.

Core logic and selected statistic functions.

Bayesian bootstrapping function bbstat.bootstrap.
Dirichlet weights generator bbstat.resample.
Bootstrap results container bbstat.BootstrapResult and credibility interval calculator bbstat.credibility_interval.
Module with initial set of weighted univariate and bivariate statistic functions bbstat.statistics (entropy, eta_square_dependency, log_odds, mean, median, mutual_information, pearson_dependency, percentile, probability, quantile, self_information, spearman_dependency, std, sum, variance).
Registry to look up included statistics by name.
Documentation, tests, and packaging.