v0.20.0

The highlight this release is the addition of optimized simultaneous ECDF confidence bands. It also includes support for idata["new_group"] = dataset directly and several bug fixes and documentation improvements.

For users of arviz.from_pytree it will now be necessary to install dm-tree manually as it was only used in this one function and has been made an optional dependency.

Full changelog available on GitHub.

v0.19.0

This release highlights are: Work with Bokeh3, uses revised Pareto k threshold, refactor plot_ecdf arguments, expose new features as preview submodule.

Full Changelog: https://github.com/arviz-devs/arviz/blob/main/CHANGELOG.md#v0190-2024-jul-19

v0.18.0

This release highlights are support for nested dictionaries and pytrees in from_dict converter and data examples updates: it adds a new example data rugby_field with 4d variables and updates the rugby example to include log_prior and unconstrained_posterior groups.

It also has bugfixes and addresses multiple deprecation warnings from dependencies that were being triggered.

Full changelog available on GitHub.

v0.17.1

Patch release with a couple bug fixes, mainly plot_forest didn't work when latest numba was installed which has now been fixed.

Full changelog available on GitHub.

v0.17.0

This new release adds a couple new features, mainly support for prior/likelihood sensitivity checks via arviz.psens and several bugfix and documentation improvements.

Full changelog available on GitHub.

v0.16.1

This patch release includes numba related fixes to avoid triggering deprecations and improve behaviour overall.

Full changelog available on GitHub.

ArviZ 0.16.0

This new release adds several new features, mostly related to InferenceData I/O and maintenance and bugfix improvements.

Full changelog available on GitHub.

v0.15.1 (2023 Mar 06)

v0.15.1 (2023 Mar 06)

New features

Maintenance and fixes

• Fix memory usage and improve efficiency in from_emcee (2215)
• Lower pandas version needed (2217)

Deprecation

Documentation

• Update documentation for various plots (2208)

v0.15.0 (2023 Feb 19)

What's Changed

New features

• Adds Savage-Dickey density ratio plot for Bayes factor approximation. (2037, 2152)
• Add CmdStanPySamplingWrapper and PyMCSamplingWrapper classes (2158)
• Changed dependency on netcdf4-python to h5netcdf (2122)

Maintenance and fixes

• Fix reloo outdated usage of ELPDData (2158)
• plot_bpv smooth discrete data only when computing u_values (2179)
• Fix bug when beanmachine objects lack some fields (2154)
• Fix gap for plot_trace with option kind="rank_bars" (2180)
• Fix plot_lm unsupported usage of np.tile (2186)
• Update _z_scale to work with SciPy 1.10 (2186)
• Fix bug in BaseLabeller when combining with with NoVarLabeller (2200)

Deprecation

Documentation

• Add PyMC and CmdStanPy sampling wrapper examples (2158)
• Fix docstring for plot_trace chain_prop and compact_prop parameters (2176)
• Add video of contributing to ArviZ webinar in contributing guide (2184)

Maintenance and fixes

• Fix dimension ordering for plot_trace with divergences (2151)

Full Changelog: v0.14.0...v0.15.0

0.14.0

What's Changed

• Fix dimension ordering for traceplot with divergences by @sethaxen in #2151
• Add Bayes Factor plot via Savage-Dickey density ratio by @orduek in #2037
• Refactor plot_bf by @aloctavodia in #2152
• Add step about conda forge recipe to how-to release notes by @OriolAbril in #2159
• rename arviz-doc to arviz-docgrid and make arviz-doc a non-grid style by @aloctavodia in #2157
• Add weight_predictions function by @aloctavodia in #2147
• add observed_rug argument to plot_ppc by @aloctavodia in #2161
• copy coords before modifying by @tspeng in #2160
• add dev flag to version by @OriolAbril in #2162
• Version bump to 0.13.1 by @zaxtax in #2163

New Contributors

• @orduek made their first contribution in #2037
• @tspeng made their first contribution in #2160

Full Changelog: v0.13.0...0.14.0