ggrust¶

The grammar of graphics, with a Rust engine and a lets-plot API.

pythonrustgrammar of graphics interactivearrowpolars

ggrust is a plotting library for Python. You describe a plot with ggplot, aes, geom_*, scale_* and theme, joined with +, exactly as in lets-plot. Rust computes the stats, scales and layout and draws the picture. Hover the plot above for tooltips: every plot on this site is a live ggrust plot.

Install¶

pipuvpoetry

pip install ggrust

uv add ggrust

poetry add ggrust

Wheels are published for Linux, macOS and Windows, for Python 3.11 and newer. Nothing else is needed: no Java, no browser, no JavaScript build.

Why ggrust¶

The lets-plot API

from ggrust import * instead of from lets_plot import *. Specs compose with + and as_dict() gives the same layout, so most code ports as an import swap.

Migrating from lets-plot
Fast, in Rust

Stats, scales, layout and rendering run in Rust. Data crosses from Python once, as Arrow: polars, pandas, pyarrow and dicts of numpy arrays all work.

Layers
Interactive by default

Notebook and HTML output has tooltips, and zoom and pan with ggtb(). Dense point layers are drawn on a canvas, so a hundred thousand points stay smooth.

Interactivity
SVG, PNG, PDF, HTML

One embedded font makes text measurement, and every PNG, identical on every machine. Export without a browser or any system library.

Export

A taste¶

import polars as pl
from ggrust import *

DATA = "https://raw.githubusercontent.com/datavil/ggrust/main/docs/data/"
cities = pl.read_csv(DATA + "cities.csv")

plot = (
    ggplot(cities, aes("month", "temperature", color="city"))
    + geom_ribbon(aes(ymin="low", ymax="high", fill="city"), alpha=0.12, color="rgba(0,0,0,0)")
    + geom_line(size=1)
    + geom_point(size=2.5)
    + scale_x_continuous(
        breaks=[1, 3, 5, 7, 9, 11], labels=["Jan", "Mar", "May", "Jul", "Sep", "Nov"]
    )
    + labs(x="", y="Temperature (°C)", color="", fill="")
    + theme_minimal()
    + ggsize(480, 320)
)
plot

ggrust was built for cellestial, a single-cell plotting library, after lets-plot's Kotlin notebook support ended. It is a general plotting library: the gallery uses everyday data.