scifont 0.1.6

Force publication-quality, editable fonts in Matplotlib figures.

scifont

中文 | English

Publication-ready Matplotlib figures with editable text and journal-compliant fonts.

The Problem

As a researcher creating figures for publication, you've likely encountered these issues:

1. Font inconsistencies across environments: Your figures look perfect on your local machine (Windows, macOS, or Linux), but when you run the same code on a server, in a Docker container, or in CI/CD pipelines, the fonts are missing. Arial and Times New Roman aren't available everywhere, causing your figures to break or look different.

2. Uneditable text in vector graphics software: You export your figure as PDF or SVG for final editing in Adobe Illustrator or Inkscape, only to discover that the text is converted to uneditable paths. Fixing a typo or adjusting labels means redrawing the entire figure.

3. Manual journal style configuration: Each journal has specific font and size requirements. You spend hours tweaking Matplotlib's rcParams to match Nature's guidelines, then repeat the entire process when submitting to Science or IEEE.

The Solution

scifont solves all three problems with one line of code:

import scifont
scifont.use('nature')  # That's it!

What it does:

• ✅ Automatically uses system fonts (Arial, Helvetica, Times New Roman) when available, seamlessly falls back to metric-compatible open-source fonts (Arimo/Tinos) when system fonts are missing
• ✅ Configures Matplotlib to export editable text in PDF/SVG files, so you can fix typos in Adobe Illustrator without redrawing
• ✅ Provides one-line configuration for Nature, Science, Cell, and IEEE style guidelines—no more manual rcParams tweaking

🚀 Key Features

• Smart Font Selection: Prioritizes system fonts (Arial, Helvetica, Times New Roman) when available. Automatically falls back to bundled open-source fonts (Arimo, Tinos) when system fonts are missing.

• Metric-Compatible Fonts: We bundle Arimo and Tinos (Apache 2.0 Licensed) as alternatives to Arial and Times New Roman. They have identical character widths and spacing, so your layout stays exactly the same.

• Editable Vector Graphics: Automatically configures Matplotlib to export PDFs as Type 42 objects and SVGs as <text> tags. Text remains editable in Adobe Illustrator, Inkscape, and CorelDRAW.

• Journal Presets: One-line configuration for Nature, Science, Cell, and IEEE standards.

📦 Installation

pip install --upgrade scifont

Or install from source:

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

⚡ Quick Start

import matplotlib.pyplot as plt
import scifont

# Apply journal style
scifont.use('nature')

# Plot as usual
plt.figure(figsize=(4, 3))
plt.plot([1, 2, 3], [4, 5, 6], label='Data')
plt.title("Editable Text in Nature Style")
plt.xlabel("Time (s)")
plt.ylabel("Velocity (m/s)")
plt.legend()

# Save - text will be editable in Adobe Illustrator!
plt.tight_layout()
plt.savefig("figure.pdf")
plt.savefig("figure.svg")

📖 Supported Journal Styles

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

Note: You can override any setting using plt.rcParams after calling scifont.use().

🎯 What scifont Does (and Doesn't Do)

What scifont solves:

• ✅ Font consistency across all platforms (Windows, macOS, Linux) and environments (local machines, servers, Docker containers)
• ✅ Editable text in PDF/SVG files—fix typos in Adobe Illustrator without redrawing
• ✅ Journal-compliant font and size settings for major journals
• ✅ True cross-platform compatibility—same code works everywhere

What scifont doesn't do:

• ❌ It's not a design tool—you still need to adjust layout, colors, and aesthetics
• ❌ It won't automatically make your figures publication-ready—you'll still need to fine-tune spacing, legends, and annotations
• ❌ It's a configuration helper, not a replacement for manual work

Think of it as a solid foundation. One scifont.use('nature') call replaces dozens of manual rcParams tweaks, but you're still in full control to customize everything afterward.

🛡️ Why Bundle Open-Source Fonts?

Arial and Times New Roman are proprietary fonts owned by Monotype. It's illegal to bundle them in a Python package without a costly license. Also, relying on system fonts alone causes inconsistencies—a plot rendered on Windows or macOS (with Arial) looks different on Linux servers or Docker containers (without Arial).

Our solution: We bundle Arimo and Tinos (developed by Steve Matteson for Google) as fallback fonts. They're:

• Apache License 2.0: Free to use, bundle, and distribute
• Metric-compatible: Identical character widths and spacing to Arial and Times New Roman. No layout changes when switching fonts.

🔧 Technical Details

By default, scifont.use() applies these Matplotlib settings:

plt.rcParams['pdf.fonttype'] = 42  # Embed TrueType fonts (editable)
plt.rcParams['ps.fonttype'] = 42
plt.rcParams['svg.fonttype'] = 'none'  # Don't convert text to paths

This ensures compatibility with Adobe Illustrator, CorelDRAW, and Inkscape.

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.