uCSS is made for crawling (large) websites to find unused CSS selectors.
Features of uCSS include:
- Find unused CSS selectors, given a HTML code base.
- Find duplicate CSS selectors.
- Count matches for each rule.
- Follows links (crawl), within the given domain.
- Give information about server responses, redirects, etc. while crawling, allowing you to find e.g. dead links.
By setting up a config file, uCSS can also:
- Do login, and visit pages both as a logged in and logged out user.
- Exclude specific pages and/or subfolders.
- Visit specific pages instead of, or in addition to, crawling.
- White list CSS rules to be ignored (e.g. those toggled by JavaScript).
- Customizable output.
uCSS is written for Node. It can be used both as a library and as a command line tool.
Want to contribute? Please see at the bottom.
Please note: uCSS is currently in beta. There will be bugs, docs may be outdated, and functionality may change.
uCSS does NOT:
- Capture rules that are switched on using JavaScript after page load.
- Look for style sheet URLs, internal style sheets, or inline styles in HTML code.
- Remove unused CSS. If you need to automatically strip away unused CSS from your favourite CSS library, you may find UnCSS helpful.
These features may, or may not, be added in the future.
npm install ucss -g
If you use Grunt, check out grunt-ucss.
For basic usage, you can use the command line options:
$ ucss --help
Usage: ucss [OPTION]...
Options:
--help This help text.
--html, -h HTML to load (local file or URL).
--css, -c CSS to load (local file or URL).
--config, -g Config file to use.
--full, -f Show full report, with details for each rule.
--silent, -s Only output list of rules. Nice if you need to pipe the output somewhere.
--duplicates, -d Show duplicates.
Either a config file, or HTML and CSS files are required. If no arguments are
specified, uCSS will look for a config_ucss.js file in the current directory.
So, to check a web page you could write
$ ucss -h http://example.com/foo.html -c http://example.com/foo.css
Note that if you use a selector both inside and outside a media query, it will
be counted as a duplicate.
To output a full report, with all found selectors and an overview of duplicates and ignored ones, you can do:
$ ucss -f -h foo.html -c foo.css
As you can see in the examples above, files can be stored locally as well as on the web.
To use a config file ("config_ucss.js") that you have created in the current folder, simply run
$ ucss
or specify another file name using the -g option. For more info on the config file, see below.
For advanced usage, please see the sections about config files.
// css can be an array of strings, file paths, or URLs
var css = [".foo {} .bar {} .baz {}"];
// html can be an array of strings, file paths, or URLs
var html = ["<html><head></head><body class='foo'></body></html>"];
var context = {
whitelist: [".baz"], // CSS selectors to ignore
auth: null, // For login (please se example elsewhere)
timeout: 400 // Request timeout (defaults to 400ms)
};
var logger = null; // Function for logging HTTP requests
// Do the magic
ucss.analyze(html, css, context, logger, function(result) {
// Do something to the result object
console.log(result);
);
There are several things you can do with a config file, that you cannot do with command line arguments.
uCSS follows links by default. But there may be specific files, or whole subfolders, that you don't want to check. These can be listed inside pages.exclude. If you want to exclude a whole subfolder, use a wildcard ("*") at the end of the url (please see example config file).
In other cases, you may want to visit just a single file, or there are files that the crawler cannot reach (e.g. because they are not linked to). Those can be added to pages.include. Also, if you want to visit certain pages under an excluded subfolder, you can add those to pages.include. Note that pages.include does not support wildcards.
In addition to managing what pages to visit (and not to visit), you can check pages both as a regular visitor and as a logged in user. This is done by specifying a function that performs a log in, and then returns a session cookie for uCSS to use for identifying itself to the server. You can also set headers to send to the server.
Furthermore, you can create a white list of selectors to be ignored. This is useful if you e.g. have classes toggled by JavaScript, or if you have special styling for various error situations that is tricky to trigger.
As well as checking several html files, uCSS can also combine CSS from several files. You can specify a list of CSS files in your config file.
If you name your config file "config_ucss.js", it will automatically be picked up by ucss. You can also name your config file something else, and use the -g option to point to it.
Again, please see the example config file. If you want to write a custom login function, please see below.
Login requires you to set up a config file. In the config file, you can specify your own login function:
module.exports = {
...,
auth: {
"username": "foo",
"password": "bar",
"loginUrl": "http://example.com/login/",
"loginFunc": function(url, username, password, callback) {
// Do login, get cookie
var cookie = "sessionid:1234"
callback(cookie);
}
},
...
}
If you use Django, you can use the supplied Django login helper (see example config file).
While crawling, uCSS will output all URLs it visits, with response status code. If logged in, it will say so. It will also output a hash of the response body, which may be useful if you want to make sure that uCSS was successfully logged in (compare the hash of the logged in and the logged out visit).
When the crawling is done, you will get a list of all unused selectors. If you've asked for a list of duplicates, that will be printed as well.
Lastly, a summary will be printed. This contains the total amount of CSS selectors found, and how many used, unused and duplicates there were.
Using the --full option, you will get a more detailed and colorful output, with all rules being listed with number of matches, in addition to number of duplicates in CSS and details about ignored and whitelisted rules.
If the output doesn't suit your needs, and you know some JavaScript, you may customize it.
You can configure uCSS to do logging and handle the result differently from what's default. Do this by adding an "output" property in the config, which can contain two functions, named "logger" and "result".
module.exports = {
...,
"output": {
"logger": function(response, originalUrl, loggedIn) {
// Do some logging here, e.g. using console.log.
},
"result": function(result) {
// Do something with the result object, e.g. print every rule
// found, together with positions in CSS file:
for (var s in result.selectors) {
// Only unused rules:
if (result.selectors[s].matches_html === 0) {
// Print position(s), given it's only one CSS file:
var pos_css = result.selectors[s].pos_css;
var key = Object.keys(pos_css)[0];
console.log(s + ": " + pos_css[key]);
}
}
}
},
...
}
The "logger" function is called every time there is a response to a HTTP request. It takes three parameters: "res" is a response object, as returned by request. "originalUrl" is a string that points to the HTML instance being visited. "loggedIn" is a boolean that is true if uCSS has sent an authentication cookie in the request header.
This function is normally used for logging, but you can make it do whatever you want. Just note that it is triggered by an event, so uCSS will not wait for it to return.
Setting "logger" to null will silence logging.
The "result" function is called when uCSS is done. It recieves an object with three properties: "used", "duplicates" and "ignored". "used" shows all CSS rules that has been matched in the HTML (including how many times). "duplicates" shows all duplicate CSS rules, including how many times they've been found. "ignored" shows all ignored rules.
This function can also whatever you want, e.g. write the result to a file.
Some pages are not accessible when crawling:
- Pages that are only accessible by posting a form will not be checked. You may add them to pages.include if they are reachable without posting data.
- All parameters in links are normally stripped away when crawling, which may have side effects for the rendering of some pages. If you want an URL to be visited with specific parameters, you have to include it in pages.include.
- When crawling, pages that are not linked to in other pages will not be checked. You may add them to pages.include.
- Style sheets are only fetched as a non-logged in user.
All at-rules are ignored, except @media: All the content inside media queries is read as if there were no media query.
Great! Feel free to pick one of the issues, or submit a bug/feature you would want to work on.
Please be tidy in your commits. Also, try to touch as small parts of the code as possible. This makes it easier to review and manage pull requests. Make sure your code is covered by tests, and write new ones if needed.
If you plan to do big changes or refactoring, please notify me first, so that we can discuss this in advance.