v0.2.4

• Replace dijets category with QCD category !170

v0.2.3

• Integrate atlas-ftag-tools package !168
• HLAPI and CI Updates !165
• Extend format saving options !160

v0.2.2

• Cast scores to full precision !159
• Add Xbb Support !157
• Improvements to the high level API !155
• Fixate the python container version !153
• Improve ROC format #146
• Fix for CI !152

v0.2.1

• Change legend label of dijets #146

v0.2.0

• Adding new high level API #128

v0.1.9

• Adding boosted categories for Xbb to utils !138
• Running pylint also for tests #133
• Fix handling of nan values in histograms #125
• Adding support for under- and overflow bins in puma.HistogramPlot #124
• (Documentation) Adding copy-button to code cells in documentation #131

v0.1.8

• Fix set_ylim in puma.PlotBase such that y-limits are correctly modified in cases with a y-offset #119
• Adding example for puma.Line2DPlot to the docs #117
• Adding support for ROC plots without ratio panels (was not possible until now) #114
• Lines with label=None (which is the default) will not appear in the legend anymore #113
• Adding new function puma.utils.get_good_linestyles() for easier linestyle management #116
• Adding the method make_linestyle_legend() which allows to specify an additional legend for linestyles #113

v0.1.7

• Adding new option to place rejection label legend in ROC plots #109

v0.1.6

• Adding support for weighted histograms (puma.Histogram now has an optional argument weights) #86
• Fixing bug where code crashed when histograms with discrete values + ratio panel were drawn #99
• Adding h5py to the Docker images #97
• Adding transparent attribute to PlotObject class. This allows to specify transparent background when initialising the plot #96

v0.1.5

• Add linewidth and alpha to legend handles + set alpha=1 by default (in puma.Histogram) #92
• Decreased default figsize for plots with zero or one ratio panels #90