Python script for minifying CSS?

Question

I'm looking for a simple Python script that can minify CSS as part of a web-site deployment process. (Python is the only scripting language supported on the server and full-blown parsers like CSS Utils are overkill for this project).

Basically I'd like jsmin.py for CSS. A single script with no dependencies.

Any ideas?

Answer 1

I don't know of any ready made python css minifiers, but like you said css utils has the option. After checking and verifying that the license allows for it, you could go through the source code and snip out the portions that do the minifying yourself. Then stick this in a single script and voila! There you go.

As a head start, the csscombine function in .../trunk/src/cssutils/script.py seems to do the work of minifying somewhere around line 361 (I checked out revision 1499). Note the boolean function argument called "minify".

Answer 2

This seemed like a good task for me to get into python, which has been pending for a while. I hereby present my first ever python script:

import sys, re

css = open( sys.argv[1] , 'r' ).read()

# remove comments - this will break a lot of hacks :-P
css = re.sub( r'\s*/\*\s*\*/', "$$HACK1$$", css ) # preserve IE<6 comment hack
css = re.sub( r'/\*[\s\S]*?\*/', "", css )
css = css.replace( "$$HACK1$$", '/**/' ) # preserve IE<6 comment hack

# url() doesn't need quotes
css = re.sub( r'url\((["\'])([^)]*)\1\)', r'url(\2)', css )

# spaces may be safely collapsed as generated content will collapse them anyway
css = re.sub( r'\s+', ' ', css )

# shorten collapsable colors: #aabbcc to #abc
css = re.sub( r'#([0-9a-f])\1([0-9a-f])\2([0-9a-f])\3(\s|;)', r'#\1\2\3\4', css )

# fragment values can loose zeros
css = re.sub( r':\s*0(\.\d+([cm]m|e[mx]|in|p[ctx]))\s*;', r':\1;', css )

for rule in re.findall( r'([^{]+){([^}]*)}', css ):

    # we don't need spaces around operators
    selectors = [re.sub( r'(?<=[\[\(>+=])\s+|\s+(?=[=~^$*|>+\]\)])', r'', selector.strip() ) for selector in rule[0].split( ',' )]

    # order is important, but we still want to discard repetitions
    properties = {}
    porder = []
    for prop in re.findall( '(.*?):(.*?)(;|$)', rule[1] ):
        key = prop[0].strip().lower()
        if key not in porder: porder.append( key )
        properties[ key ] = prop[1].strip()

    # output rule if it contains any declarations
    if properties:
        print "%s{%s}" % ( ','.join( selectors ), ''.join(['%s:%s;' % (key, properties[key]) for key in porder])[:-1] )

I believe this to work, and output it tests fine on recent Safari, Opera, and Firefox. It will break CSS hacks other than the underscore & /**/ hacks! Do not use a minifier if you have a lot of hacks going on (or put them in a separate file).

Any tips on my python appreciated. Please be gentle though, it's my first time. :-)

Answer 3

There is a port of YUI's CSS compressor available for python.

Here is its project page on PyPi: http://pypi.python.org/pypi/cssmin/0.1.1