The CDFChart class is designed for visualizing Cumulative Distribution Functions (CDF), which are essential for analyzing data distributions, latency, and performance variability in technical papers.
from hachimiku import CDFChart
cc = CDFChart(figsize=(10, 6))figsize: Default figure size (width, height) in inches.
| Parameter | Type | Description |
|---|---|---|
log_scale_x |
bool |
Enable logarithmic scale for the X-axis (default: True). |
markers |
bool/list |
Enable markers (True to use default set, or provide a list). Default: None (no markers). |
markevery |
int/list |
Frequency of markers on the line. Essential for CDFs. If markers is enabled but markevery is not set, a sparse default is used. |
grid |
bool |
Show both X and Y grid lines. |
grid_x, grid_y |
bool |
Show only vertical or horizontal grid lines. |
ylabel |
str |
Default is 'CDF (%)'. |
Creates a CDF chart where each series can have its own X-data points. You can use markers to distinguish lines even in black-and-white prints.
cc.create_cdf_chart(
...,
markevery=100, # Show a marker every 100 points
markersizes=8
)
Accepts a list of dictionaries for more flexible configuration of individual curves.
Technical data like latency often spans multiple orders of magnitude. The log_scale_x=True setting (default) provides a clear view of tail behavior.
cc.create_cdf_chart(
x_data_list=[x1, x2],
y_data_list=[y1, y2],
log_scale_x=True,
grid=True
)For datasets with a narrower range, a linear X-axis can be used.
cc.create_cdf_chart(
...,
log_scale_x=False,
grid_y=True
)
See examples/cdf_chart_gallery.py for the complete source code used to generate the gallery above.