Skip to content

Installation

From PyPI

pip install pyplotrs

Pre-built wheels ship the compiled Rust core and the bundled fonts, so there is no build step and no Rust toolchain required on the common platforms (Linux, macOS, Windows; x86-64 and arm64).

pyplotrs supports Python 3.9+ and has no required runtime dependencies — not even NumPy. It accepts plain Python lists (and anything iterable), and works with NumPy arrays or pandas/polars columns if you have them.

Virtual environments

As always, prefer installing into a virtual environment:

python -m venv .venv
source .venv/bin/activate      # Windows: .venv\Scripts\activate
pip install pyplotrs

Verify

import pyplotrs as pp

fig, ax = pp.subplots()
ax.line([0, 1, 2], [0, 1, 4])
fig.save("test.png")
print(pp.resolved_font_name())   # the body font resolved on this machine

From source

Building from source needs a Rust toolchain (1.92 or newer) and maturin. The project uses uv in development, but any PEP 517 frontend works.

What a source build actually does

It compiles ~110 crates with lto = true and codegen-units = 1, which takes a few minutes and needs network access to fetch the registry the first time. The build is entirely offline afterwards. You only need any of this on a platform with no wheel — pip install pyplotrs downloads a pre-built one on Linux, macOS and Windows, x86-64 and arm64.

git clone https://github.com/tkclam/pyplotrs
cd pyplotrs
pip install .

git clone https://github.com/tkclam/pyplotrs
cd pyplotrs
uv venv
uv pip install maturin
maturin develop --release    # build + install into the venv

maturin develop builds the extension and installs it editable, so Python changes are picked up immediately and only Rust changes need a rebuild. Drop --release for faster compiles while developing (slower runtime).

Optional extras

Extra Installs For
pyplotrs[docs] mkdocs-material, mkdocs-jupyter, mkdocstrings, black Building this documentation site
pyplotrs[test] pytest Running the test suite
pyplotrs[bench] matplotlib, numpy Running the head-to-head benchmarks
pyplotrs[dev] ipykernel, nbclient, matplotlib, numpy Running the notebooks and re-executing them 
pip install -e ".[docs]"
mkdocs serve

None of these are needed to use pyplotrs — the library itself has no runtime dependencies.

What gets bundled

pyplotrs embeds four fonts into the compiled extension (so rendering always works without relying on system fonts):

  • Liberation Sans — the permissive, Arial-metric-compatible fallback for body text, labels, ticks and legends.
  • Fira Math (175 KB) — the sans math font, and what a text face cannot draw in a $...$ span: radicals, big operators and stretchy delimiters, all of which have to grow with their content. It carries the OpenType MATH table that positions every atom.
  • DejaVu Sans, subset to the math symbol blocks (95 KB) — sans shapes for symbols neither the body family nor Fira Math carries (⊥, ⊢, ⋃, ∖).
  • STIX Two Math — the whole span under set_mathtext_fontset("stix"), and otherwise the last resort: the Script and Fraktur alphabets and double-struck digits, which no sans math font here has.

Letters, Greek, digits and the common operators come from the body face itself, so $10^{-3}$ matches the plain ticks beside it and $E = mc^2$ is one typeface throughout.

Liberation Sans, Fira Math and STIX Two Math are under the SIL Open Font License 1.1; the DejaVu Sans subset is under the Bitstream Vera license. For body text pyplotrs prefers the host's Arial, then Helvetica, before falling back to Liberation Sans — and embeds whichever it picks into every saved figure. See styling & themes for how to control this.