analyze-css:CSS 选择器的复杂度和性能分析器
analyze-css
CSS selectors complexity and performance analyzer. analyze-css is built as a set of rules bound to events fired by CSS parser. Each rule can generate metrics and add "offenders" with more detailed information (see Usage section for an example).
Install
analyze-css comes as a "binary" for command-line and as CommonJS module. Run the following to install them globally:
npm install --global analyze-css
Usage
Command line tool
You can use analyze-css "binary" to analyze local CSS files or remote CSS assets:
analyze-css --file examples/elecena.css analyze-css --url http://jigsaw.w3.org/css-validator/style/base.css
You can provide CSS via stdin as well (notice the dash: -
):
echo ".foo {margin: 0 \!important}" | analyze-css -
This will emit JSON formatted results on stdout
. Use --pretty
(or -p
shortcut) option to make the output readable for human beings.
Basic HTTP authentication can be provided through the options --auth-user
and --auth-pass
.
CommonJS module
var analyzer = require('analyze-css'); new analyzer('.foo {margin: 0 !important}', function(err, results) { console.error(err); console.log(results); // example? see below });
// options can be provided var opts = { 'noOffenders': true }; new analyzer('.foo {margin: 0 !important}', opts, function(err, results) { console.error(err); console.log(results); // example? see below });```
grunt task
Created by @DeuxHuitHuit
npm i grunt-contrib-analyze-css
It uses configurable threshold and compares the analyze-css result with it.
Results
{ "generator": "analyze-css v0.8.0", "metrics": { "base64Length": 11332, "redundantBodySelectors": 0, "redundantChildNodesSelectors": 1, "colors": 106, "comments": 1, "commentsLength": 68, "complexSelectors": 37, "complexSelectorsByAttribute": 3, "duplicatedSelectors": 7, "duplicatedProperties": 24, "emptyRules": 0, "expressions": 0, "oldIEFixes": 51, "imports": 0, "importants": 3, "mediaQueries": 0, "multiClassesSelectors": 74, "parsingErrors": 0, "oldPropertyPrefixes": 79, "propertyResets": 0, "qualifiedSelectors": 28, "specificityIdAvg": 0.04, "specificityIdTotal": 25, "specificityClassAvg": 1.27, "specificityClassTotal": 904, "specificityTagAvg": 0.79, "specificityTagTotal": 562, "selectors": 712, "selectorLengthAvg": 1.5722460658082975, "selectorsByAttribute": 92, "selectorsByClass": 600, "selectorsById": 25, "selectorsByPseudo": 167, "selectorsByTag": 533, "universalSelectors": 5, "length": 55173, "rules": 433, "declarations": 1288 }, "offenders": { "importants": [ ".foo {margin: 0 !important}" ] } }
Metrics
- base64Length: total length of base64-encoded data in CSS source (will warn about base64-encoded data bigger than 4 kB)
- redundantBodySelectors: number of redundant body selectors (e.g.
body .foo
,section body h2
, but notbody > h1
) - redundantChildNodesSelectors: number of redundant child nodes selectors (e.g.
ul li
,table tr
) - colors: number of unique colors used in CSS
- comments: number of comments in CSS source
- commentsLength: length of comments content in CSS source
- complexSelectors: number of complex selectors (consisting of more than three expressions, e.g.
header ul li .foo
) - complexSelectorsByAttribute: number of selectors with complex matching by attribute (e.g.
[class$="foo"]
) - duplicatedSelectors: number of CSS selectors defined more than once in CSS source
- duplicatedProperties: number of CSS property definitions duplicated within a selector
- emptyRules: number of rules with no properties (e.g.
.foo { }
) - expressions: number of rules with CSS expressions (e.g.
expression( document.body.clientWidth > 600 ? "600px" : "auto" )
) - oldIEFixes: number of fixes for old versions of Internet Explorer (e.g.
* html .foo {}
and.foo { *zoom: 1 }
, read more) - imports number of
@import
rules - importants: number of properties with value forced by
!important
- mediaQueries: number of media queries (e.g.
@media screen and (min-width: 1370px)
) - multiClassesSelectors: reports selectors with multiple classes (e.g.
span.foo.bar
) - parsingErrors: number of CSS parsing errors
- oldPropertyPrefixes: number of properties with no longer needed vendor prefix, powered by data provided by autoprefixer (e.g.
--moz-border-radius
) - propertyResets: number of accidental property resets
- qualifiedSelectors: number of qualified selectors (e.g.
header#nav
,.foo#bar
,h1.title
) - specificityIdAvg: average specificity for ID
- specificityIdTotal: total specificity for ID
- specificityClassAvg: average specificity for class, pseudo-class or attribute
- specificityClassTotal: total specificity for class, pseudo-class or attribute
- specificityTagAvg: average specificity for element
- specificityTagTotal: total specificity for element
- selectors: number of selectors (e.g.
.foo, .bar { color: red }
is counted as two selectors -.foo
and.bar
) - selectorLengthAvg: average length of selector (e.g. for
.foo .bar, #test div > span { color: red }
will be set as 2.5) - selectorsByAttribute: number of selectors by attribute (e.g.
.foo[value=bar]
) - selectorsByClass: number of selectors by class
- selectorsById: number of selectors by ID
- selectorsByPseudo: number of pseudo-selectors (e,g.
:hover
) - selectorsByTag: number of selectors by tag name
- universalSelectors: number of selectors trying to match every element (e.g.
.foo > *
) - length: length of CSS source (in bytes)
- rules: number of rules (e.g.
.foo, .bar { color: red }
is counted as one rule) - declarations: number of declarations (e.g.
.foo, .bar { color: red }
is counted as one declaration -color: red
)
Read more
- Writing Efficient CSS (by Mozilla)
- Optimize browser rendering (by Google)
- Profiling CSS for fun and profit. Optimization notes.
- CSS specificity
- CSS Selector Performance has changed! (For the better) (by Nicole Sullivan)
- CSS Performance (by Paul Irish)
- GitHub's CSS Performance (by Joh Rohan)
Dev hints
Running tests and linting the code:
npm test && npm run-script lint
Turning on debug mode (i.e. verbose logging to stderr via debug module):
DEBUG=analyze-css* analyze-css ...