tueplots

TUEplots is an extension for matplotlib designed to facilitate the creation of scientific plots for publications (papers, presentations, posters).

Unlike many styling libraries, TUEplots focuses on avoiding common publication issues such as:

• Incorrect figure sizes (e.g., not matching publication text-width).
• Inappropriate font sizes.
• Inconsistencies across different figures.

It is designed to be non-opinionated regarding aesthetics (colors, markers, etc.) and follows a stateless design to ensure compatibility with other matplotlib extensions.

The tueplots.constants.color package provides access to predefined color palettes and RGB color definitions used for scientific plotting. This package is organized into two main submodules:

• tueplots.constants.color.palettes: Contains various color palettes suitable for scientific publications.
• tueplots.constants.color.rgb: Contains specific RGB color definitions.

You can use these constants to ensure consistent and publication-quality coloring across your matplotlib plots.

TUEplots operates without internal state. It works by generating and passing around dictionaries containing key-value pairs that match standard matplotlib parameters (rcParams).

Because it does not modify global state directly, you can manage settings using standard matplotlib patterns:

1. Global changes: Use matplotlib.pyplot.rcParams.update() to apply settings across your entire session.
2. Context-specific changes: Use matplotlib.pyplot.rc_context() to apply settings only to a specific block of code.

This approach ensures that TUEplots remains compatible with other matplotlib-based tools and extensions.

The tueplots.constants package provides centralized access to predefined constants used for scientific plotting, specifically focusing on markers and colors. This ensures consistency across different plots in a publication. The package is organized into submodules:

• tueplots.constants.markers: Contains predefined marker styles.
• tueplots.constants.color: Contains predefined color palettes and color constants.

Most functions in tueplots (such as markers) return dictionaries that follow the matplotlib rcParams format. You can apply these configurations in two ways:

1. As a context manager: Use plt.rc_context() to apply settings only to a specific block of code. This is recommended to avoid polluting global state.
2. Globally: Use plt.rcParams.update() to change the configuration for the entire session.

import matplotlib.pyplot as plt
from tueplots import markers

# Get the configuration dictionary
config = markers.inverted()

# Option 1: Use as a context manager (local scope)
with plt.rc_context(config):
    plt.plot([1, 2, 3], marker='o')
    plt.show()

# Option 2: Update global configuration (global scope)
plt.rcParams.update(config)

To set up a development environment with all dependencies required for continuous integration (CI), install the package using the [ci] extra.

pip install .[ci]

If you prefer not to use a complete bundle, you can manually configure specific aspects of the ICML 2022 style using individual modules:

• Figure Sizes: Use figsizes.icml2022_full() for full-width figures or figsizes.icml2022_half(nrows=2, ...) for half-width figures.
• Font Sizes: Use fontsizes.icml2022() to set standard sizes for labels, legends, and ticks.
• Font Families: Use fonts.icml2022() for standard serif/sans-serif settings, or fonts.icml2022_tex(family="sans-serif") to enable LaTeX-based font configurations.

from tueplots import figsizes, fontsizes, fonts

# Configure figure size
plt.rcParams.update(figsizes.icml2022_full())

# Configure font sizes
plt.rcParams.update(fontsizes.icml2022())

# Configure font families
plt.rcParams.update(fonts.icml2022(family="serif"))

The project uses tox to automate various quality checks. You can run all checks at once, or target specific environments for testing or linting.

# Run all checks
tox

# Run only the tests
tox -e test

# Run only the linter
tox -e format-and-lint

To automatically ensure code compliance with black and isort, you can install and use pre-commit. This should be done once during setup. After installation, your code will be automatically checked for compatibility every time you commit.

pip install pre-commit
pre-commit install

Contributions to tueplots should focus on the style files used in scientific papers. This includes:

• New versions of existing styles: e.g., adding icml2025 if icml2023 already exists.
• New venues: Adding style files for conferences or journals not currently supported (e.g., CVPR).

For any other type of contribution, you should create an issue first to discuss it with the maintainers.

When implementing a new style, use the conference/journal's author instructions to find figure sizes and font sizes.

Important Guidelines:

• Font Size: Use the Figure captions as your reference for font size, not the main body text.
• Font Selection: Consult the LaTeX style file source code to identify the correct font implementation. Common patterns include:
  • If \includepackage{times} is used: Use the same configuration as ICLR-2023.
  • If \includepackage{ptm} is used: Mirror this by including the same package (even if it is obsolete) to match the template closely.
  • If no font-related package is included: It likely uses Computer Modern. Copy the configuration from JMLR-2001.

Note: Be skeptical of templates claiming to use 'Times New Roman' without a LaTeX style file, as native Times New Roman implementation in LaTeX typically requires XeLaTeX.