Skip to main content

remove un-used css from a file

Project description

CSS Clean!

PyPI version

This is a library for cleaning up css based on styles that are actually used in html files. Although there are libraries that use npm I am not a fan so I wanted something in Python.

Approach

We take a conservative approach to filtering, meaning:

  • for any style block, one matching class or identifier is sufficient for inclusion
  • multiple levels selectors are not combined (e.g., using "button" is sufficient to add .button. red.
  • we always keep keyframes and imports. Additional parsing could be added to support checking, but they are fairly rare (imho).
  • The library is new and I may have missed a case! Please open an issue if you find something.

Of course this could be improved upon but we would need more careful token parsing! Please open an issue if you are interested.

Usage

Install

First, clone and install.

$ git clone https://github.com/vsoch/cssclean
$ cd cssclean
$ pip install -e .

or install from pip.

$ pip install cssclean

And then try running an example!

Clean Directory

Below we ask cssclean to automatically detect files (css and html) under the same test directory:

$ cssclean clean --css cssclean/tests/testdata/1 --html cssclean/tests/testdata/1
Sheet style.css has 124 rules.
Sheet style.css filtered down to 43 rules.
1 files written:
  style.css => /home/vanessa/Desktop/Code/unused-css/cssclean/tests/testdata/1/style.clean.css

As you can see, by default, it will write the updated css files alongside the old files but with .clean added. Note that you can provide individual files, one at a time too, for both of --css and --html.

$ cssclean clean --css cssclean/tests/testdata/1 --html cssclean/tests/testdata/1/home.html --html cssclean/tests/testdata/1/table.html

Dry Run

You can also do a dry run to print the filtered style to the screen:

$ cssclean clean --css cssclean/tests/testdata/1 --html cssclean/tests/testdata/1 --dry-run
...
.card.yellow {
  background-color: var(--color-yellow);
}
tr {
  background-color: white;
}

SCSS

We have support for cleaning of scss files, meaning that you will generate a final cleaned css file from a potential directory of inputs. As an example, we can use the hierarchy under test data, where the main style file is style.scss that includes all the others:

$ cssclean clean --css cssclean/tests/testdata/3/style.scss --html cssclean/tests/testdata/3/index.html
Sheet style.css has 638 rules.
Sheet style.css filtered down to 298 rules.
1 files written:
  style.css => cssclean/tests/testdata/3/style.clean.css

All imports are required to exist since we are actually rendering the sass into css in the library.

Minify

And you can ask for the same preview, but minified:

$ cssclean clean --css cssclean/tests/testdata/1 --html cssclean/tests/testdata/1 --dry-run --minify

Or a non-preview minified:

$ cssclean clean --css cssclean/tests/testdata/1 --html cssclean/tests/testdata/1 --outdir ./out --minify
Sheet style.css has 124 rules.
Sheet style.css filtered down to 43 rules.
1 files written:
  style.css => ./out/style.clean.min.css

Note that "min" is added to the saved filename, and also note that this is experimental - I'll need to do some more tests and reading to figure out what exactly the process is.

Output Directory

You can also ask for a custom output directory:

$ mkdir out
$ cssclean clean --css cssclean/tests/testdata/1 --html cssclean/tests/testdata/1 --outdir ./out
Sheet style.css has 124 rules.
Sheet style.css filtered down to 43 rules.
1 files written:
  style.css => ./out/style.clean.css

Example

As an example, I tested against the usrse site. We can render the site into _site:

$ bundle exec jekyll build

and then run the clean in place. Here is the original style file:

$ ls -l style.css 
-rw-rw-r-- 1 dinosaur dinosaur 263390 Jun  4 20:57 style.css

We can then run the clean in place:

$ cssclean clean --css assets/css/style.css --html _site/ --in-place
Sheet style.css has 4080 rules.
Sheet style.css filtered down to 1380 rules.
1 files written:
  style.css => assets/css/style.css

And what is the new size?

$ ls -l style.css 
-rw-rw-r-- 1 dinosaur dinosaur 201347 Jun  5 15:11 style.css

So there is a huge reduction in size when we first remove unused stuff (-62043 bytes)! To summarize:

  • original: ~263KB
  • cleaned: ~201KB

And importantly, the site looks the same!

docs/img/usrse-site.png

So personally, I don't have issue with the size differences between minified and cleaned - my preference is not to minify so I can easily read it, but I can also imagine a workflow where you minify just for the production site and the development version is just cleaned. It's really up to you!

GitHub Action

You can use this tool as a GitHub action! Here's how.

  - name: Test CSS Clean
    uses: vsoch/cssclean/action/clean@main
    with:
      css: cssclean/tests/testdata/1/style.css
      html: cssclean/tests/testdata/1/table.html cssclean/tests/testdata/1/home.html
      in_place: true

See the action folder, action.yml file for more parameters and descriptions.

Contributors

We use the all-contributors tool to generate a contributors graphic below.


Vanessasaurus

💻

Thanks!

This library wouldn't be possible without these fantastic other libraries!

TODO

  • add tests for different cases
  • create GitHub action to detect (fail if unused)
  • automated release
  • some kind of summary command that shows used/unused styles, or the count
  • can we use asp to determine used rules?

License

This code is licensed under the MPL 2.0 LICENSE.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

cssclean-0.0.12.tar.gz (17.3 kB view details)

Uploaded Source

File details

Details for the file cssclean-0.0.12.tar.gz.

File metadata

  • Download URL: cssclean-0.0.12.tar.gz
  • Upload date:
  • Size: 17.3 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.1 CPython/3.9.12

File hashes

Hashes for cssclean-0.0.12.tar.gz
Algorithm Hash digest
SHA256 6dc427ac559ecb216cd7b5894b2f4471ce649a66dd796783d570483653f582ef
MD5 c4931116d198e76c468f0edd0c2b8ad1
BLAKE2b-256 43c1063d34a99bcc52afc2fd91d2f5e6c0e4ce2d69bc47e1777c296e0f3684d8

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page