Python wrapper and CLI utility to render LaTeX markup and equations as SVG using dvisvgm and scour (or optionally svgo).

Based on the original work by Tino Wagner, this version has enhanced features.

The design goals are:

• Auto vertical alignment without any further styling.
  • by adding a style="vertical-align:" to the SVG attributes
• Auto scaling on font size/zoom change.
  • by using em units for width, height and style
• Unique IDs so multiple SVGs on one page won’t corrupt each other.
  • by generating random 3-character ID prefixes within the SVG
• Minified SVG for direct inclusion.
  • by using scour (or optionally svgo, a Node app)
• For e-book readers and dictionaries:
  • Good legibility.
  • Direct inclusion as <svg>, not necessarily <img>.
  • Possibility for easy CSS styling, i.e. using fill.
  • Compatibility with pyglossary and the ebook-reader-dict project.
  • Possibility of LaTeX preamble changes/additions, to correct LaTeX code in automated processes like converting a Wiktionary dump to an e-reader dictionary.
  • See examples/screenshots.md for a real-life example.

Note: This tool is intended to produce inline SVGs, so the XML prolog normally found in standard SVG files gets stripped per default.

Should you need the SVG for other purposes that require the XML prolog, just add

<?xml version="1.0" encoding="UTF-8"?>

as the first line in the file.

As long as the PyPI name issue is unresolved, just install directly from this repository:

pip3 install git+https://github.com/Moonbase59/latex2svg.git#egg=latex2svg

A specific version can be installed by adding that version’s tag to the command:

pip3 install git+https://github.com/Moonbase59/latex2svg.git@v0.4.1#egg=latex2svg

Should your pip or setuptools be an older version, or you just prefer to have the source code on your machine, you can use

git clone https://github.com/Moonbase59/latex2svg.git
cd latex2svg
pip3 install . -U

to install it (including the CLI utility latex2svg).

from latex2svg import latex2svg
out = latex2svg(r'$\sin(x) = \sum_{n=0}^{\infty} \dots$')
print(out['valign'])  # baseline offset in em
print(out['svg'])  # rendered SVG

$ latex2svg --help
usage: latex2svg [-h] [--version] [--preamble PREAMBLE]
                 [--optimizer {scour,svgo,none}] [--scale SCALE]

Render LaTeX code from stdin as SVG to stdout. Writes metadata (baseline
offset, width, height in em units) into the SVG attributes.

optional arguments:
  -h, --help            show this help message and exit
  --version             show program's version number and exit
  --preamble PREAMBLE   LaTeX preamble code to read from file
  --optimizer {scour,svgo,none}
                        SVG optimizer to use (default: scour)
  --scale SCALE         SVG output scaling (default: 1.000000)

$ echo '$\sin(x) = \sum_{n=0}^{\infty} \dots$' | latex2svg > sample.svg

Result:

In case you want to change it using --preamble, here is the built-in default preamble:

\usepackage[utf8x]{inputenc}
\usepackage{amsmath}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amstext}
\usepackage{newtxtext}
\usepackage[libertine]{newtxmath}
% prevent errors from old font commands
\DeclareOldFontCommand{\rm}{\normalfont\rmfamily}{\mathrm}
\DeclareOldFontCommand{\sf}{\normalfont\sffamily}{\mathsf}
\DeclareOldFontCommand{\tt}{\normalfont\ttfamily}{\mathtt}
\DeclareOldFontCommand{\bf}{\normalfont\bfseries}{\mathbf}
\DeclareOldFontCommand{\it}{\normalfont\itshape}{\mathit}
\DeclareOldFontCommand{\sl}{\normalfont\slshape}{\@nomath\sl}
\DeclareOldFontCommand{\sc}{\normalfont\scshape}{\@nomath\sc}
% prevent errors from undefined shortcuts
\newcommand{\N}{\mathbb{N}}
\newcommand{\R}{\mathbb{R}}
\newcommand{\Z}{\mathbb{Z}}

Python module coding example:

import latex2svg
params = latex2svg.default_params
params['preamble'] = r"""
\usepackage[utf8x]{inputenc}
\usepackage{amsmath}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amstext}
"""
out = latex2svg.latex2svg(r'$\sin(x) = \sum_{n=0}^{\infty} \dots$', params)
with open('sample2.svg', 'w') as f:
    f.write(out['svg'])

Using SVGs as inline SVG should work perfectly well in all cases.

Should you need to include SVGs as an external HTML <img>, you must include latex2svg’s computed valign value in the <img> tag as a style="vertical-align: …" attribute to achieve correct alignment. The SVG won’t be stylable using a CSS fill in this case.

Hint: You can always find the value later by examining the SVG’s root style attribute.

<img src="density.svg" style="vertical-align:-.600321em">

This is not a limitation of latex2svg.

Python module coding example:

from latex2svg import latex2svg
out = latex2svg(r'$\sin(x) = \sum_{n=0}^{\infty} \dots$')
with open('sample3.svg', 'w') as f:
    f.write(out['svg'])
print('<img src="sample3.svg" style="vertical-align:%.6fem">' % out['valign'])

A complete <img> definition (to avoid rendering side effects) would look like:

print('<img src="sample3.svg" style="width:%.6fem;height:%.6fem;vertical-align:%.6fem">'
    % (out['width'], out['height'], out['valign']))

and result in:

<img src="sample3.svg" style="width:6.891840em;height:1.054628em;vertical-align:-0.313464em">

• Python 3
• A working LaTeX installation, like Tex Live or MacTeX
• dvisvgm
• scour (preferred and default), —or—
• svgo

0.2.0

• first public version

0.2.1

• some small bug fixes

0.3.0, 0.3.1 — 2022-02-14

• Now uses scour as default SVG optimizer (instead of svgo). This produces even smaller SVGs, and it’s "all Python" (pip3 install scour).
• The SVG optimizer to be used can now be set via the --optimizer command line option, or by setting params['optimizer'] if using this as a Python module. Possible values are: scour (the default), svgo or none.

0.4.0 — 2022-02-16

• Added --scale command line option (params['scale']) for some rare cases where additional scaling is needed. Float, defaults to 1.000000.
• Bugfix: Set valign to zero if dvisvgm returns no depth.

0.4.1 — 2022-02-18

• Attempt at making this an installable Python Package. Unfortunately, PyPI refuses the name latex2svg, although I couldn’t find a package with this name on PyPI. So for now you’ll have to git clone and manually install it.

This project is licensed under the MIT license. See LICENSE for details.

Copyright © 2022 Matthias C. Hormann