plotwright

A small matplotlib companion for consistent, presentable figures: a composable color/linestyle/ marker theme system, and a FigureSaver for exporting whole figures or single cropped subplots at a fixed, reproducible size.

• Github repository: https://github.com/AdamHallengreen/plotwright/

Installation

pip install plotwright

or, with uv:

uv add plotwright

Quickstart

import matplotlib.pyplot as plt
from plotwright import FigureSaver, apply_theme

apply_theme("vibrant", base="paper")  # both are defaults; shown here for clarity

fig, ax = plt.subplots()
ax.plot([0, 1, 2], [0, 1, 4], label="series 1")
ax.legend()

FigureSaver("figures").save_figure(fig, "example.png")

apply_theme sets matplotlib rcParams for the rest of the session -- scope it to one figure with plt.rc_context() if you need more than one theme active side by side. FigureSaver handles writing figures to disk at a consistent size.

Palettes, bases, linestyles & markers

apply_theme(palette=, *, base=, linestyle=, marker=, ...) layers a base theme with a color cycle, and optionally a linestyle/marker cycle alongside it. Every name below is also available programmatically, so you're never guessing:

from plotwright import list_bases, list_palettes, list_linestyle_sets, list_marker_sets

list_bases()            # ('paper', 'presentation')
list_palettes()         # ('fishy', 'ggplot2', 'ibm', 'okabe_ito', 'petroff10', ...)
list_linestyle_sets()   # ('basic',)
list_marker_sets()      # ('basic',)

Palettes (palette=): vibrant (default), fishy, ggplot2, ibm, okabe_ito, petroff10, seaborn_deep, tab10, tol_bright, tol_muted. Linestyle sets and marker sets (linestyle=/marker=): basic. Preview a palette's actual colors without applying it via palette_colors("vibrant").

Pass your own colors instead of a named palette with colors=[...] (a list of hex strings), and layer extra rcParams on top with extra_rc= (a dict, or a path to another .mplstyle file). See notebooks/demo.ipynb for a runnable tour of every option, including a side-by-side swatch of all palettes, sequential=, and latex=True.

Saving figures

saver = FigureSaver("figures", dpi=300)

saver.save_figure(fig, "whole_figure.png")                        # fixed size, no tight bbox
saver.save_subplot(fig, ax, "one_panel.png")                       # cropped to that axes' tight bbox
saver.save_subplots_aligned(fig, [ax1, ax2], ["a.png", "b.png"])   # cropped to a common width

Use mark_rasterized(*artists) to rasterize dense scatter/heatmap artists before saving to a vector format (pdf/eps/ps) -- keeps text and axes vector while the heavy artist becomes a bitmap.

Contributing

See CONTRIBUTING.md.

License

MIT -- see LICENSE.