README.md

   1 # HTMLMinifier
   2
   3 [![NPM version](https://img.shields.io/npm/v/html-minifier.svg)](https://www.npmjs.com/package/html-minifier)
   4 [![Build Status](https://img.shields.io/travis/kangax/html-minifier.svg)](https://travis-ci.org/kangax/html-minifier)
   5 [![Dependency Status](https://img.shields.io/david/kangax/html-minifier.svg)](https://david-dm.org/kangax/html-minifier)
   6
   7 [HTMLMinifier](https://kangax.github.io/html-minifier/) is a highly **configurable**, **well-tested**, JavaScript-based HTML minifier.
   8
   9 See [corresponding blog post](http://perfectionkills.com/experimenting-with-html-minifier/) for all the gory details of [how it works](http://perfectionkills.com/experimenting-with-html-minifier/#how_it_works), [description of each option](http://perfectionkills.com/experimenting-with-html-minifier/#options), [testing results](http://perfectionkills.com/experimenting-with-html-minifier/#field_testing) and [conclusions](http://perfectionkills.com/experimenting-with-html-minifier/#cost_and_benefits).
  10
  11 [Test suite is available online](https://kangax.github.io/html-minifier/tests/).
  12
  13 Also see corresponding [Ruby wrapper](https://github.com/stereobooster/html_minifier), and for Node.js, [Grunt plugin](https://github.com/gruntjs/grunt-contrib-htmlmin), [Gulp module](https://github.com/jonschlinkert/gulp-htmlmin), [Koa middleware wrapper](https://github.com/koajs/html-minifier) and [Express middleware wrapper](https://github.com/melonmanchan/express-minify-html).
  14
  15 For lint-like capabilities take a look at [HTMLLint](https://github.com/kangax/html-lint).
  16
  17 ## Minification comparison
  18
  19 How does HTMLMinifier compare to other solutions — [HTML Minifier from Will Peavy](http://www.willpeavy.com/minifier/) (1st result in [Google search for "html minifier"](https://www.google.com/#q=html+minifier)) as well as [htmlcompressor.com](http://htmlcompressor.com) and [minimize](https://github.com/Swaagie/minimize)?
  20
  21 | Site                                                                         | Original size *(KB)* | HTMLMinifier | minimize | Will Peavy | htmlcompressor.com |
  22 | ---------------------------------------------------------------------------- |:--------------------:| ------------:| --------:| ----------:| ------------------:|
  23 | [Google](https://www.google.com/)                                            | 48                   | **44**       | 48       | 49         | 48                 |
  24 | [HTMLMinifier](https://github.com/kangax/html-minifier)                      | 154                  | **117**      | 128      | 133        | 128                |
  25 | [Twitter](https://twitter.com/)                                              | 203                  | **162**      | 195      | 219        | 195                |
  26 | [Stack Overflow](https://stackoverflow.com/)                                 | 254                  | **196**      | 208      | 216        | 205                |
  27 | [Bootstrap CSS](https://getbootstrap.com/docs/3.3/css/)                      | 271                  | **260**      | 269      | 229        | 269                |
  28 | [BBC](https://www.bbc.co.uk/)                                                | 288                  | **230**      | 280      | 281        | 272                |
  29 | [Amazon](https://www.amazon.co.uk/)                                          | 508                  | **439**      | 495      | 501        | n/a                |
  30 | [Wikipedia](https://en.wikipedia.org/wiki/President_of_the_United_States)    | 533                  | **434**      | 517      | 536        | 517                |
  31 | [New York Times](https://mathiasbynens.be/_tmp/nyt.html)                     | 699                  | **619**      | 693      | 683        | n/a                |
  32 | [NBC](https://www.nbc.com/)                                                  | 700                  | **657**      | 698      | 699        | n/a                |
  33 | [Eloquent Javascript](https://eloquentjavascript.net/1st_edition/print.html) | 870                  | **815**      | 840      | 864        | n/a                |
  34 | [ES6 table](https://kangax.github.io/compat-table/es6/)                      | 5308                 | **4529**     | 5025     | n/a        | n/a                |
  35 | [ES draft](https://tc39.github.io/ecma262/)                                  | 6082                 | **5456**     | 5624     | n/a        | n/a                |
  36
  37 ## Options Quick Reference
  38
  39 Most of the options are disabled by default.
  40
  41 | Option                         | Description     | Default |
  42 |--------------------------------|-----------------|---------|
  43 | `caseSensitive`                | Treat attributes in case sensitive manner (useful for custom HTML tags) | `false` |
  44 | `collapseBooleanAttributes`    | [Omit attribute values from boolean attributes](http://perfectionkills.com/experimenting-with-html-minifier/#collapse_boolean_attributes) | `false` |
  45 | `collapseInlineTagWhitespace`  | Don't leave any spaces between `display:inline;` elements when collapsing. Must be used in conjunction with `collapseWhitespace=true` | `false` |
  46 | `collapseWhitespace`           | [Collapse white space that contributes to text nodes in a document tree](http://perfectionkills.com/experimenting-with-html-minifier/#collapse_whitespace) | `false` |
  47 | `conservativeCollapse`         | Always collapse to 1 space (never remove it entirely). Must be used in conjunction with `collapseWhitespace=true` | `false` |
  48 | `customAttrAssign`             | Arrays of regex'es that allow to support custom attribute assign expressions (e.g. `'<div flex?="{{mode != cover}}"></div>'`) | `[ ]` |
  49 | `customAttrCollapse`           | Regex that specifies custom attribute to strip newlines from (e.g. `/ng-class/`) | |
  50 | `customAttrSurround`           | Arrays of regex'es that allow to support custom attribute surround expressions (e.g. `<input {{#if value}}checked="checked"{{/if}}>`) | `[ ]` |
  51 | `customEventAttributes`        | Arrays of regex'es that allow to support custom event attributes for `minifyJS` (e.g. `ng-click`) | `[ /^on[a-z]{3,}$/ ]` |
  52 | `decodeEntities`               | Use direct Unicode characters whenever possible | `false` |
  53 | `html5`                        | Parse input according to HTML5 specifications | `true` |
  54 | `ignoreCustomComments`         | Array of regex'es that allow to ignore certain comments, when matched | `[ /^!/ ]` |
  55 | `ignoreCustomFragments`        | Array of regex'es that allow to ignore certain fragments, when matched (e.g. `<?php ... ?>`, `{{ ... }}`, etc.)  | `[ /<%[\s\S]*?%>/, /<\?[\s\S]*?\?>/ ]` |
  56 | `includeAutoGeneratedTags`     | Insert tags generated by HTML parser | `true` |
  57 | `keepClosingSlash`             | Keep the trailing slash on singleton elements | `false` |
  58 | `maxLineLength`                | Specify a maximum line length. Compressed output will be split by newlines at valid HTML split-points |
  59 | `minifyCSS`                    | Minify CSS in style elements and style attributes (uses [clean-css](https://github.com/jakubpawlowicz/clean-css)) | `false` (could be `true`, `Object`, `Function(text, type)`) |
  60 | `minifyJS`                     | Minify JavaScript in script elements and event attributes (uses [UglifyJS](https://github.com/mishoo/UglifyJS2)) | `false` (could be `true`, `Object`, `Function(text, inline)`) |
  61 | `minifyURLs`                   | Minify URLs in various attributes (uses [relateurl](https://github.com/stevenvachon/relateurl)) | `false` (could be `String`, `Object`, `Function(text)`) |
  62 | `preserveLineBreaks`           | Always collapse to 1 line break (never remove it entirely) when whitespace between tags include a line break. Must be used in conjunction with `collapseWhitespace=true` | `false` |
  63 | `preventAttributesEscaping`    | Prevents the escaping of the values of attributes | `false` |
  64 | `processConditionalComments`   | Process contents of conditional comments through minifier | `false` |
  65 | `processScripts`               | Array of strings corresponding to types of script elements to process through minifier (e.g. `text/ng-template`, `text/x-handlebars-template`, etc.) | `[ ]` |
  66 | `quoteCharacter`               | Type of quote to use for attribute values (' or ") | |
  67 | `removeAttributeQuotes`        | [Remove quotes around attributes when possible](http://perfectionkills.com/experimenting-with-html-minifier/#remove_attribute_quotes) | `false` |
  68 | `removeComments`               | [Strip HTML comments](http://perfectionkills.com/experimenting-with-html-minifier/#remove_comments) | `false` |
  69 | `removeEmptyAttributes`        | [Remove all attributes with whitespace-only values](http://perfectionkills.com/experimenting-with-html-minifier/#remove_empty_or_blank_attributes) | `false` (could be `true`, `Function(attrName, tag)`) |
  70 | `removeEmptyElements`          | [Remove all elements with empty contents](http://perfectionkills.com/experimenting-with-html-minifier/#remove_empty_elements) | `false` |
  71 | `removeOptionalTags`           | [Remove optional tags](http://perfectionkills.com/experimenting-with-html-minifier/#remove_optional_tags) | `false` |
  72 | `removeRedundantAttributes`    | [Remove attributes when value matches default.](http://perfectionkills.com/experimenting-with-html-minifier/#remove_redundant_attributes) | `false` |
  73 | `removeScriptTypeAttributes`   | Remove `type="text/javascript"` from `script` tags. Other `type` attribute values are left intact | `false` |
  74 | `removeStyleLinkTypeAttributes`| Remove `type="text/css"` from `style` and `link` tags. Other `type` attribute values are left intact | `false` |
  75 | `removeTagWhitespace`          | Remove space between attributes whenever possible. **Note that this will result in invalid HTML!** | `false` |
  76 | `sortAttributes`               | [Sort attributes by frequency](#sorting-attributes--style-classes) | `false` |
  77 | `sortClassName`                | [Sort style classes by frequency](#sorting-attributes--style-classes) | `false` |
  78 | `trimCustomFragments`          | Trim white space around `ignoreCustomFragments`. | `false` |
  79 | `useShortDoctype`              | [Replaces the `doctype` with the short (HTML5) doctype](http://perfectionkills.com/experimenting-with-html-minifier/#use_short_doctype) | `false` |
  80
  81 ### Sorting attributes / style classes
  82
  83 Minifier options like `sortAttributes` and `sortClassName` won't impact the plain-text size of the output. However, they form long repetitive chains of characters that should improve compression ratio of gzip used in HTTP compression.
  84
  85 ## Special cases
  86
  87 ### Ignoring chunks of markup
  88
  89 If you have chunks of markup you would like preserved, you can wrap them `<!-- htmlmin:ignore -->`.
  90
  91 ### Preserving SVG tags
  92
  93 SVG tags are automatically recognized, and when they are minified, both case-sensitivity and closing-slashes are preserved, regardless of the minification settings used for the rest of the file.
  94
  95 ### Working with invalid markup
  96
  97 HTMLMinifier **can't work with invalid or partial chunks of markup**. This is because it parses markup into a tree structure, then modifies it (removing anything that was specified for removal, ignoring anything that was specified to be ignored, etc.), then it creates a markup out of that tree and returns it.
  98
  99 Input markup (e.g. `<p id="">foo`)
 100
 101 ↓
 102
 103 Internal representation of markup in a form of tree (e.g. `{ tag: "p", attr: "id", children: ["foo"] }`)
 104
 105 ↓
 106
 107 Transformation of internal representation (e.g. removal of `id` attribute)
 108
 109 ↓
 110
 111 Output of resulting markup (e.g. `<p>foo</p>`)
 112
 113 HTMLMinifier can't know that original markup was only half of the tree; it does its best to try to parse it as a full tree and it loses information about tree being malformed or partial in the beginning. As a result, it can't create a partial/malformed tree at the time of the output.
 114
 115 ## Installation Instructions
 116
 117 From NPM for use as a command line app:
 118
 119 ```shell
 120 npm install html-minifier -g
 121 ```
 122
 123 From NPM for programmatic use:
 124
 125 ```shell
 126 npm install html-minifier
 127 ```
 128
 129 From Git:
 130
 131 ```shell
 132 git clone git://github.com/kangax/html-minifier.git
 133 cd html-minifier
 134 npm link .
 135 ```
 136
 137 ## Usage
 138
 139 Note that almost all options are disabled by default. For command line usage please see `html-minifier --help` for a list of available options. Experiment and find what works best for you and your project.
 140
 141 * **Sample command line:** ``html-minifier --collapse-whitespace --remove-comments --remove-optional-tags --remove-redundant-attributes --remove-script-type-attributes --remove-tag-whitespace --use-short-doctype --minify-css true --minify-js true``
 142
 143 ### Node.js
 144
 145 ```js
 146 var minify = require('html-minifier').minify;
 147 var result = minify('<p title="blah" id="moo">foo</p>', {
 148   removeAttributeQuotes: true
 149 });
 150 result; // '<p title=blah id=moo>foo</p>'
 151 ```
 152
 153 ## Running benchmarks
 154
 155 Benchmarks for minified HTML:
 156
 157 ```shell
 158 node benchmark.js
 159 ```