scifont 0.1.3

Force publication-quality, editable fonts in Matplotlib figures.

scifont

Professional, Reproducible, and Editable Figures for Scientific Publishing.

scifont is a lightweight Python library designed to solve the two most persistent headaches in scientific visualization with Matplotlib:

1. Missing Fonts in Docker/Linux: "Arial" or "Times New Roman" are often missing in CI/CD environments or Linux servers, causing plots to fallback to default fonts.
2. Non-Editable Text: Default Matplotlib settings often export text as paths (shapes), making it impossible to correct typos or adjust formatting in Adobe Illustrator or Inkscape.

🚀 Key Features

• Smart Font Selection: Automatically prioritizes system fonts (Arial, Helvetica, Times New Roman) when available. Falls back to bundled open-source fonts (Arimo, Tinos) only when system fonts are missing. This ensures optimal performance and user preference while maintaining compatibility.
• Bundled Open-Source Fonts: We ship with Arimo and Tinos (Apache 2.0 Licensed). These are metric-compatible alternatives to Arial and Times New Roman. They are legally safe to distribute and use anywhere.
• Metric-Compatible: Arimo has the exact same character widths and spacing as Arial. Tinos matches Times New Roman. Your layout will remain identical to the proprietary versions.
• Editable Vector Graphics: Automatically configures Matplotlib to export PDFs as Type 42 objects and SVGs as <text> tags. Text remains editable in downstream vector software.
• Journal Presets: One-line configuration for Nature, Science, Cell, and IEEE standards.

📦 Installation

pip install scifont

(Or if installing from source)

git clone https://github.com/studentiz/scifont.git
cd scifont
pip install .

⚡ Quick Start

Simply import scifont and select your target journal style at the start of your script.

import matplotlib.pyplot as plt
import scifont

# 1. Apply style (e.g., 'nature' for Life Sciences, 'ieee' for Engineering)
scifont.use('nature')

# 2. Plot as usual
plt.figure(figsize=(4, 3))
plt.plot([1, 2, 3], [4, 5, 6], label='Data')

# The font will be Arial/Helvetica if available, otherwise Arimo (Arial-compatible)
plt.title("Editable Text in Nature Style")
plt.xlabel("Time (s)")
plt.ylabel("Velocity (m/s)")
plt.legend()

# 3. Save
# The output PDF/SVG will have editable text!
plt.tight_layout()
plt.savefig("figure.pdf")
plt.savefig("figure.svg")

📖 Supported Journal Styles

scifont adjusts font families, sizes, and line widths according to specific journal guidelines (2025 standards).

Style Argument	Target Font Family	Primary Font (if available)	Fallback Font	Base Font Size	Target Journals
'nature'	Sans-serif	Arial/Helvetica	Arimo (matches Arial)	7 pt	Nature, Nature Comms, Scientific Reports
'cell'	Sans-serif	Arial/Helvetica	Arimo (matches Arial)	8 pt	Cell, Molecular Cell, Neuron
'science'	Sans-serif	Arial/Helvetica	Arimo (matches Arial)	9 pt	Science, Science Advances
'ieee'	Serif	Times New Roman/Times	Tinos (matches Times New Roman)	8 pt	IEEE Transactions, Phys. Rev. Lett.

Note: You can override any setting (e.g., font size) using standard plt.rcParams after calling scifont.use().

🛡️ Smart Font Selection & Legal Compliance

Font Selection Strategy: scifont intelligently selects fonts based on availability:

1. System Fonts First: If Arial, Helvetica, or Times New Roman are available on your system, they will be used automatically. This provides the best performance and respects user preferences.
2. Automatic Fallback: If system fonts are missing (common in Docker/Linux environments), scifont automatically falls back to bundled open-source fonts (Arimo/Tinos).

Why Bundle Open-Source Fonts? Many researchers are unaware that Arial and Times New Roman are proprietary fonts owned by Monotype.

1. Distribution: It is illegal to bundle arial.ttf inside a Python package and distribute it without a costly license.
2. Reproducibility: Reliance on system fonts alone leads to inconsistencies. A plot rendered on macOS (with Arial) looks different on Ubuntu (without Arial).

Our Solution: We bundle Arimo and Tinos, developed by Steve Matteson for Google, as fallback fonts.

• License: Apache License 2.0. You can use, bundle, and distribute them freely.
• Compatibility: They are designed to be Metric-Compatible with Arial and Times New Roman. This means if you replace Arial with Arimo, the text takes up the exact same space. No line breaks will change; no labels will be cut off.

🔧 Editable Text (Vector Graphics)

By default, scifont.use() applies the following hidden Matplotlib settings to ensure compatibility with Adobe Illustrator, CorelDRAW, and Inkscape:

plt.rcParams['pdf.fonttype'] = 42  # Embbed TrueType fonts (editable)
plt.rcParams['ps.fonttype'] = 42
plt.rcParams['svg.fonttype'] = 'none' # Do not convert text to path

License

The scifont code is distributed under the MIT License. The bundled fonts (Arimo and Tinos) are distributed under the Apache License, Version 2.0.