gsplot creates publication-quality scientific figures with a concise API on
top of Matplotlib. It adds paper-aware layouts, deterministic plotting and
styling helpers, validated JSON defaults, and lightweight build metadata while
returning ordinary Matplotlib Figure, Axes, and Artist objects.
The package is still evolving. Check the documentation and the issue tracker before depending on behavior that is not covered by the public API.
gsplot supports Python 3.10 and newer:
python -m pip install gsplotimport gsplot as gs
fig, axes = gs.subplots("AB")
gs.line(axes["A"], [0, 1, 2], [0, 1, 4], label="quadratic")
gs.scatter(axes["B"], [0, 1, 2], [0, 1, 4], label="samples", s=15)
gs.label(axes, "x", "value", square=True, index="in")
gs.legend(axes)
gs.save(fig, "quickstart", show=False)This creates a two-panel Matplotlib figure, saves quickstart.png and
quickstart.pdf, and remains compatible with regular Matplotlib operations.
The canonical helpers always receive their Figure or Axes target explicitly.
save writes PNG and PDF transactionally at 600 DPI with a tight crop and
displays the Figure after successful writes by default; pass show=False for
batch or headless output. Use crop=False when output dimensions must match
the Figure design canvas exactly. The advanced savefig helper retains its
conservative output controls.
For a complete scientific example, see the publication example.
Configuration is optional and explicit. Load a schema-2 JSON file and pass the immutable value to supported functions:
import gsplot as gs
config = gs.load_config("path/to/gsplot.json")
fig, axes = gs.subplots(config=config)When a value is specified more than once, the precedence is:
- an argument passed directly to the function;
- the supplied immutable
Configvalue; - the function's default value.
Canonical code never searches the working directory or home directory for a configuration file. See the configuration guide for the supported schema, precedence, and backend notes.
The repository uses Poetry and targets Python 3.10 or newer. With a compatible Python interpreter:
python -m pip install "poetry==2.4.1"
poetry install
MPLBACKEND=Agg poetry run pytest -q
MPLBACKEND=Agg poetry run sphinx-build -W -b html docs docs/_build/htmlThe scripts under examples/ are executable documentation. Run one from its own
directory, for example:
cd examples/layouts && python mosaic.pyexamples/compatibility/legacy_v0.py intentionally demonstrates the deprecated
0.x surface. All other plotting examples use the canonical explicit-target API.
See the developer setup guide for formatting, type checking, packaging, and Docker instructions.
For private vulnerability reports and the supported-version policy, see SECURITY.md.
This repository builds on the original code developed by Giordano Mattoni.
- Original author: Giordano Mattoni
- Maintainer: Soichiro Yamane
This project is distributed under the MIT License.
