my-mpl-style

Two small helpers that give matplotlib figures a publication-ready look:

• plot_rc() applies a consistent rcParams preset — serif fonts, Computer Modern math, a MATLAB-like colour cycle, optional grid and tick direction, plus 300 dpi / tight-bbox save defaults.
• ax_setting3D() turns a 3D axes into a MATLAB-like orthographic box with explicit tick positions and projected tick marks.

Installation

pip install my-mpl-style

Requires Python 3.10+, matplotlib 3.10+ and NumPy 1.24+.

Quick start

import matplotlib.pyplot as plt
from my_mpl_style import plot_rc

plot_rc(grid=True)                 # serif + CM math, 300 dpi, tight bbox

fig, ax = plt.subplots()
ax.plot([0, 1, 2], [0, 1, 4], label="data")
ax.set(xlabel="$x$", ylabel="$y$")
ax.legend(frameon=False)

fig.savefig("figure.png")          # uses savefig.dpi=300, bbox="tight"

Every default can be overridden per call:

plot_rc(
    save_dpi=600,
    save_bbox="standard",
    save_pad_inches=0.1,
    font_family="sans-serif",
    font_serif=["DejaVu Sans"],
)

3D axes

import matplotlib.pyplot as plt
import numpy as np
from my_mpl_style import ax_setting3D

a = np.linspace(-1, 1, 200)
x, y = np.meshgrid(a, a)
z = np.sin(np.pi * x) * np.exp(y)

fig = plt.figure()
ax = fig.add_subplot(111, projection="3d")
ax.plot_surface(x, y, z, linewidth=0, antialiased=True)
ax_setting3D(
    ax,
    (-1, 1), (-1, 1), (z.min(), z.max()),
    xticks=[-1, -0.5, 0, 0.5, 1],
    yticks=[-1, -0.5, 0, 0.5, 1],
    zticks=[-2, 0, 2],
    box_aspect=(1, 1, 1),
)
plt.show()

Ticks are passed to ax_setting3D() rather than set separately with ax.set(xticks=...), so the projected tick marks and z labels are computed from the final limits and ticks.

API

plot_rc(...)

plot_rc(
    grid=False,
    tick_direction="out",
    tick_right=False,
    tick_top=False,
    save_dpi=300,
    save_bbox="tight",
    save_pad_inches=0.05,
    font_family="serif",
    font_serif=("Arial", "Helvetica", "DejaVu Sans"),
)

Updates the global matplotlib.rcParams, so it affects every figure created afterwards.

rcParams written by the preset:

ax_setting3D(...)

ax_setting3D(
    ax,
    xlim, ylim, zlim,
    xticks, yticks, zticks,
    aspect="equal",
    box_aspect=None,
    xticks_pad=-5,
    yticks_pad=-5,
    zticks_pad=0,
    box_on=False,
    ztick_length=0.02,
    xtick_length=0.02,
    ytick_length=0.02,
)

Makes the panes transparent, thins the grid lines, sets the tick positions and limits, scales every tick to the longest projected axis, and places the z tick labels beyond their tick endpoints. X and y tick directions and label positions are preserved.

Development

python -m venv .venv
.venv\Scripts\activate          # Windows
pip install -e ".[dev,examples]"

pytest
python examples/demo.py          # regenerate docs/images/demo.png
python -m build
twine check dist/*

examples/demo.py prepends src/ to sys.path, so it always renders with the checkout's code even when an older release is installed in site-packages.

Changelog

See CHANGELOG.md.

Example

2D lines and a 3D surface — examples/demo.py

Intersecting planes — examples/Intersecting_planes.py

Three orthogonal colour slices inside a single 3D box, each drawn with contourf and a shared colour bar.

Isosurfaces — examples/isosurface.py

Marching-cubes isosurfaces of the phase-field snapshots in data/, shaded with a LightSource. This example needs scikit-image and the data/*.npz files from the repository.

Triangular surfaces — examples/tri_surface.py

Triangular meshes loaded from data/cheese.npz and data/peanut.npz: a colour-mapped surface on the left and a shaded wireframe on the right.

License

MIT © 2026 chengtau