kwplot 0.4.13

The main webpage for this project is: https://gitlab.kitware.com/computer-vision/kwplot

The kwplot module is a wrapper around matplotlib and can be used for visualizing algorithm results.

The top-level API is:

from .auto_backends import (autompl, autoplt, set_mpl_backend,)
from .draw_conv import (make_conv_images, plot_convolutional_features,)
from .mpl_3d import (plot_surface3d,)
from .mpl_color import (Color,)
from .mpl_core import (distinct_colors, distinct_markers, ensure_fnum, figure,
                       imshow, legend, next_fnum, set_figtitle,
                       show_if_requested,)
from .mpl_draw import (draw_boxes, draw_boxes_on_image, draw_clf_on_image,
                       draw_line_segments, draw_text_on_image, plot_matrix, draw_points,)
from .mpl_make import (make_heatmask, make_orimask, make_vector_field,)
from .mpl_multiplot import (multi_plot,)
from .mpl_plotnums import (PlotNums,)

One of the key features is the kwplot.autompl function, which is able to somewhat intelligently set the notorious matplotlib backend. By default it will attempt to use PyQt5 if it is installed and a DISPLAY is available. Otherwise it will ensure the backend is set to Agg.

The kwplot.multi_plot function is able to create line and bar plots with multiple lines/bars in a labeled axes using only a single function call. This can dramatically reduce the code size needed to perform simple plot visualizations as well as ensure that the code that produces the data is decoupled from the code that does the visualization.

The kwplot.imshow and kwplot.figure functions are extensions of the matplotlib versions with slightly extended interfaces (again to help reduce the density of visualization code in research scripts).