FigureSpec¶

spec.figure is a FigureSpec — figure-level properties (size, resolution, backgrounds, font). Used when the spec creates its own figure.

import behaviz as bv

spec = bv.PlotSpec(
    figure=bv.FigureSpec(figsize=(8, 5), dpi=150, face_color="#101010", axes_color="#181818"),
    text_color="#e0e0e0",
)
bv.plot_line("t", "v", data=df, spec=spec)

Fields¶

Field	Default	Meaning
`figsize`	`(12, 8)`	figure size, inches
`dpi`	`120`	dots per inch
`tight`	`True`	call `tight_layout()` automatically
`style`	`"default"`	a `plt.style` name, a preset, or a raw rcParams dict
`face_color`	`None`	figure background (`None` → backend default)
`axes_color`	`None`	plot-area background
`font_family`	`None`	font family for all text

Examples¶

Size & resolution¶

bv.FigureSpec(figsize=(6, 4), dpi=300)   # print-ready

Dark backgrounds, without a preset¶

face_color/axes_color are first-class fields and they work on bokeh too. Pair with PlotSpec.text_color so labels stay legible.

import behaviz as bv
import numpy as np
import polars as pl

x = np.linspace(0, 2 * np.pi, 100)
y = np.sin(x)

df = pl.DataFrame({"t":x,"v":y})

spec = bv.PlotSpec(
    figure=bv.FigureSpec(face_color="#ffebb1e0",
    axes_color="#1e1e1e"),
    text_color="#061fad",
    x=bv.AxisSpec(grid_color="#c42828"),
    y=bv.AxisSpec(grid_color="#299803"),
)

bv.plot_scatter("t","v",data=df, spec=spec)

ugly_plot

Raw rcParams via `style`¶

For matplotlib/seaborn you can still drop a full rcParams dict on style (line widths, marker sizes, fonts). bokeh has no rcParams, so it honours background and text colour only from a style — which is why the discrete fields above exist.

bv.FigureSpec(style={"lines.linewidth": 3, "axes.facecolor": "#fafafa"})

Presets set these for you

The built-in presets (presentation, presentation_dark, print, …) populate face_color/axes_color/font_family/spine/tick fields from their rcParams, so a preset renders the same on all three backends. See Presets.