New features:
TypeError when using unknown argument
non-square panels are now possible
n_points_scatter to set the number of points in the scatter plot
Fixes:
equally spaced automatic ticks
Further improvements:
improved documentation and docstrings
coverage tests
improved tests
cleanup
Fixes:
grid is also plotted in the upper triangle if the plotting order is lower upper lower (and vice versa)
ranges of empty plots are correctly displayed
added warning for people like Arne that use the parameter names incorrectly.
Fix: scatter_kwargs for scatter_density
Fix: RGB(A) colors in plt.contour
New feature:
n_sigma_for_one_sided_tail free parameter to manually set the threshold when to use a 2-sided or 1-sided interval
Fixes:
correct normalization when using samples with prob
avoid upper limits with +- 0.0 and lower limits with – 0.0
Labels showing correctly when using upper triangle. Improved documentation.
New features:
LineChain class
params_from argument to select parameter from chain
add_derived_params to add derived_params to the input
input data can be pandas dataframes
colors follow the matplotlib color cycles by default
when computing bestfits and uncertainties, it automatically detects if to show bestfit +/- uncertainty or a lower/upper limit.
Fixes:
ylimits in 1D plots
tick values when choosing many ticks
Further improvements:
documentation
dependency cleanup
Fixed 1D histograms when using prob.
First release on PyPI. New features:
dictionaries are new possible inputs
progressbar can be disabled
Bug fix: Correct normalization of 1D posteriors (credit to Alexander Charles Tikam)
Bug fix: Number of digits of bestfits and uncertainties are now correctly set.
First public release on Gitlab.