meg.js is a parsing expression parser for HTML and Markdown. It is based on the insanely great peg.js by David Majda (@dmajda).

Note: While the transformatiom from Markdown to HTML is lossless, parsing HTML to Markdown is not. meg.js is designed to be used to help save data inside contenteditable elements in a (hopefully) more universal format.

You have to install node.js if you haven't already.

After you have setup node run npm install. When all packages have been installed you can run:

• grunt or grunt release to build the project (minified)
• grunt test to run the test suite (via Karma).
• grunt dev will build the project unminified.
• grunt tdd if you want to write new tests and work on the project.

For reference here is the grammar used to transform HTML to Markdown:

Char <- /[^<]/
LowerCase <- /[a-z]/
TagName <- LowerCase+
TextNode <- Char*
StartTag <- '<' TagName '>'
ClosingTag <- '</' TagName '>'
Element <- StartTag Content ClosingTag
Content <- Element / TextNode

while Content is the only starting expression. Char and LowerCase are terminal symbols.

var parser = meg(),
    result = parser.fromHTML('<div>This is an <em>example</em>!</div>');

    // result is "This is an *example*!"

You can also pass additional mutation rules to meg.js:

var parser = meg({ exp: /^sup$/, start: '^', end: '' }),
    result = parser.fromHTML('I am n<sup>2</sup>.');

    // result is "i am n²."

Warning: This feature is experimental, because it is not tested! For more informations see the tests.