Changelog¶

0.5.1¶

Bug to fix bokeh plots
Adjust save/show/inline behavior slightly. return_filename and print_filename now only work when image is actually saved to disk

Major speed improvements achieved for matplotlib engine. Actual time reduction depends on the plot type and the number of subplots, data sets, etc. Based on the plots included as unit tests:
- ~30% increase for standard xy plots
- ~40-50% increase for barplot, boxplots (sans advanced features like violins), contours, and heatmaps
New plot types added:
- gantt
- imshow (preferred choice for display of image data over heatmap)
- pie
Significant cleanup of legacy code and refactoring of Data and Layout classes
Full documentation of kwargs API
Support added for:
- upper- and lower-control limit area shading
- marker size definition via a column in the DataFrame (allows emphasis of specific points)
- cdf and pdf conversion with fcp.hist
- rolling mean in fcp.bar
Numerous bug fixes

fix heatmap but when DataFrame column names are of type object but are actually int

legend:
- allow placing legend below plot (legend_location=='below' or legend_location==11)
- fixes for figure grouping and multi-column plots
heatmap:
- font and font size fixes
themes:
- beautification of “white” theme
- allow switching of themes via keyword at the plot call
- allow overloading of default color list in theme file
other:
- fixed requirements so it will import after install in clean environment
- deal with iterticks warning from matplotlib versions >= 3.1
- added save_data keyword to dump a subset of the plotted data only from the original pandas.DataFrame
- file extension bug fix
- filter improvements (allow “not in” list)
- added option to disable alphabetical sorting of data and plot based on the order in which data appears in the original pandas.DataFrame