Predictr

Fixed
- MRR based NPBB are now computed correctly. A typo in the code lead to a bug that computed solely MLE based bounds, no matter if 'mrr' was set as an attribute. A big THANK YOU to William Gandler who reported this bug and various spelling mistakes!
- Minor other spelling mistakes pointed out by Antonia, Sarahi and Edna

Added
- added new style argument plot in PlotAll.contour_plot(style='...'): Currently scatter and angular line plots are supported

Fixed
- Boosted LRB computation by fully vectorizing each computation step and using algeabric tricks (as a tip: don't use np.power(), because it is way too slow for big computations). Now, it takes half the amount of time to have more precise results.

Added
- added show argument in PlotAll.contour_plot()
- added customizable x-axis limits in PlotAll.mult_weibull: x_bounds=[start, end]

Added
- In previous versions of predictr, saving Weibull plots required to plot and show them. Now, saving plots and actually showing them are independent from each other and can be called separately.

Fixed
- Computation of one-sided Likelihood Ratio Bounds had an error, which always resulted in a 95% one-sided bound; no matter what was set by the user.

Added
- Legend fontsize can now be set in mult_weibull(legend_fontsize=9)
- Bootstrap bounds now fully support censored data
- Non-parametric boostrap bias-correction now fully supports censored data (parametric already did)

Fixed
- Units in axis labels are now according to the International System of Units (SI):e.g. '[%]' -> 'in %'