Biopsykit

New Features
- `biopsykit.classification`: `SklearnPipelinePermuter` now supports regression
Bugfixes
- further small bugfixes

New Features
- `biopsykit.stats`: Added new statistic features:
- `biopsykit.stats.multicoll`: Functions to handle multicollinearity in data
- `biopsykit.stats.regression`: Functions for performing more "complex" regression analysis
(currently: stepwise backward multiple linear regression)
- `biopsykit.classification`: `SklearnPipelinePermuter` instances can now be saved as pickle files and imported again
-
Bugfixes
_ `biopsykit.utils.time`: Renamed `time_to_datetime` to `time_to_timedelta` (wrong function name)

Bugfixes
- `biopsykit.stats.StatsPipeline`: Fixed bug of failing to generate significance brackets from within-subjects
stats data
- `biopsykit.protocols.plotting.saliva_plot`: Fixed wrong plot styling

New Features
- `biopsykit.plotting.feature_boxplot`: Added `legend_title` parameter
- `biopsykit.classification.model_selection.nested_cv`: Train and test indices are now saved for each fold of the nested cv
- `biopsykit.classification.model_selection.SklearnPipelinePermuter`:
- Updated "scoring" and "refit" behavior
- Added confusion matrix, true labels, predicted labels, train and test indices per fold and flattened

Bugfixes
- `biopsykit.signals.imu.static_moment_detection`: Now returning an empty dataframe with correct column names when
detecting no static moments
- `biopsykit.classification.model_selection.SkleanPipelinePermuter`: Pipelines are now correctly sorted by their
mean test scores
- `biopsykit.protocols.plotting.saliva_plot`: Fixed bug when attempting to plot saliva data without sample times
_ `biopsykit.plotting.feature_boxplots`: Updated plotting color configuration: If an alpha value smaller than 1.0 is
passed, the saturation is by default set to 1.0 to still have a nice color shading despite the transparency
- `biopsykit.protocols.TSST`: Changed default TSST test time from 20 to 15 min

New Features
- `biopsykit.stats.StatsPipeline`:
- added new functions `StatsPipeline.results_to_latex_table` and `StatsPipeline.stats_to_latex`
to convert statistical analysis results to LaTeX tables and to inline LaTeX code, respectively.
- changed argument `stats_type` to `stats_effect_type`.
**WARNING: This is a breaking change!** The use of `stats_type` is deprecated and will be removed in `0.5.0`!
- `biopsykit.colors`: Removed `biopsykit.colors` module and replaced it with colors from the new `fau_colors` package.
**WARNING: This is a breaking change!**
Bugfixes
- `biopsykit.signals.imu.static_moment_detection`: end indices of static moments are inclusive instead of exclusive
Misc
- `biopsykit.utils.array_handling`: if `overlap_percent` > 1 it is converted to a fraction [0.0, 1.0]

Misc
- Updated dependencies: now using scikit-learn `>=1.0`