mathjax-node-page

This Node.js module builds on mathjax-node and provides processing of larger content fragments

Installation

Use

npm install mathjax-node-page

to install mathjax-node-page and its dependencies.

Usage

mathjax-node-page exports mjpage which expects four parameters:

mjpage(input, mjpageConfig, mjnodeConfig, callback)

Where input is a string with HTML, pageConfig specifies page-wide options, and mjnodeConfig expects mathjax-node configuration options.

The defaults for pageConfig are

{
    format: ["MathML", "TeX", "AsciiMath"], // determines type of pre-processors to run
    output: '', // global override for output option; 'svg', 'html' or 'mml'
    tex: {}, // configuration options for tex pre-processor, cf. lib/tex.js
    ascii: {}, // configuration options for ascii pre-processor, cf. lib/ascii.js
    singleDollars: false, // allow single-dollar delimiter for inline TeX
    fragment: false, // return body.innerHTML instead of full document
    cssInline: true,  // determines whether inline css should be added
    jsdom: {... }, // jsdom-related options
    displayMessages: false, // determines whether Message.Set() calls are logged
    displayErrors: true, // determines whether error messages are shown on the console
    undefinedCharError: false, // determines whether unknown characters are saved in the error array
    extensions: '', // a convenience option to add MathJax extensions
    fontURL: 'https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.0/fonts/HTML-CSS', // for webfont urls in the CSS for HTML output
    MathJax: {} // options MathJax configuration, see https://docs.mathjax.org
}

and where mjnodeConfig represents mathjax-node configuration options, the defaults are.

{
  ex: 6, // ex-size in pixels
  width: 100, // width of math container (in ex) for linebreaking and tags
  useFontCache: true, // use <defs> and <use> in svg output?
  useGlobalCache: false, // use common <defs> for all equations?
  state: mjstate, // track global state
  linebreaks: false, // do linebreaking?
  equationNumbers: "none", // or "AMS" or "all"
  math: "", // the math to typeset
  html: false, // generate HTML output?
  css: false, // generate CSS for HTML output?
  mml: false, // generate mml output?
  svg: false, // generate svg output?
  speakText: true, // add spoken annotations to output?
  timeout: 10 * 1000, // 10 second timeout before restarting MathJax
}

Advanced usage

mathjax-node customization

mathjax-node-page exports init function that allows you to pass in a custom mathjax-node (for example, mathjax-node-svg2png).

const mjnode = require('mathjax-node-svg2png');
mjpage.init(mjnode);

If your custom mathjax-node provides new output options, you can add them by calling addOutput. As a second parameter, you can pass custom output handler, which is a function that modifies a DOM element with the conversion result. The default output handler behavior is to write contents to wrapper.innerHTML.

mjpage.addOutput('png', (wrapper, data) => {
	wrapper.innerHTML = `<img src="${data}">`;
});
// ...now you can use standard mathjax-node-page API

Reset to default mathjax-node behavior by calling init with empty parameters. Ensure that all your current mathjax-node-page tasks have been completed before calling it.

mjpage.init();  // reset back to default mathjax-node

Events

mjpage runs jobs which inherit EventEmitter and provide the following event hooks. Add the corresponding event handlers to manipulate the input/output and DOM before/after conversion.

All the event handlers are destroyed when job ends to prevent memory leaks.

Formula conversion events

• beforeConversion -> handler(parsedFormula): runs before individual formula conversion started, but after initial DOM processing. All the formulas are wrapped in <script type="..."> tags, where @type is one of the following:

const scripts = document.querySelectorAll(`
    script[type="math/TeX"],
    script[type="math/inline-TeX"],
    script[type="math/AsciiMath"],
    script[type="math/MathML"],
    script[type="math/MathML-block"]`
);

• afterConersion -> handler(parsedFormula): runs after individual formula conversion completed and DOM was changed. Formula DOM node is a <span class="mjpage..."> wrapper whose contents are the conversion result.

All formula conversion events pass ParsedFormula instance to the event handler.

{
    id, // index of formula on the page
    jobID, // mjpage job ID; formulas belonging to the same page run have the same jobID 
    node, // DOM node with the formula (contents change before and after conversion)
    sourceFormula, // the source formula
    sourceFormat, // the source formula format (e.g. "inline-TeX")
    outputFormula, // the converted formula result from mathjax-node typeset function; use outputFormula[outputFormat] to get the resulting formula string 
    outputFormat // the resulting formula format (e.g. "svg")
}

Page conversion events

• beforeSerialiation -> handler(document, css): runs when converted page DOM was prepared immediately before serialization. Use to manipulate resulting page DOM. The event handler receives document node (jsdom) and page css.

mjpage function callback receives result after the DOM serialization.

Example

mjpage(input, {
    format: ["TeX"]
}, {
    svg: true
}, function(output) {
    // output is your final result
})
.on('afterConversion', function(parsedFormula) {
    // manipulate parsed result and DOM at your will
    // see description of parsedFormula object above
});

CLI

mathjax-node-page installs a CLI tool. Run mjpage to print usage instructions.

Example

const mjpage = require('../lib/main.js').mjpage;
const fs = require('fs');
const input = fs.readFileSync('input.html');

mjpage(input, {format: ["TeX"]}, {svg: true}, function(output) {
    console.log(output); // resulting HTML string
});

Tips

• mathjax-node-page is intended to work with HTML pages and it uses DOM parser while extracting formulas from the document. If your page is not HTML, make sure you escape < (<) and > (>) before usage. This test contains an example.