BeautifulSoup: get css classes from html

BeautifulSoup itself doesn't parse CSS style declarations at all, but you can extract such sections then parse them with a dedicated CSS parser.

Depending on your needs, there are several CSS parsers available for python; I'd pick cssutils (requires python 2.5 or up (including python 3)), it is the most complete in it's support, and supports inline styles too.

Other options are css-py and tinycss.

To grab and parse such all style sections (example with cssutils):

import cssutils
sheets = []
for styletag in tree.findAll('style', type='text/css')
    if not styletag.string: # probably an external sheet
        continue
    sheets.append(cssutils.parseStyle(styletag.string))

With cssutil you can then combine these, resolve imports, and even have it fetch external stylesheets.

A BeautifulSoup & cssutils combo will do the trick nicely:

    from bs4 import BeautifulSoup as BSoup
    import cssutils
    selectors = {}
    with open(htmlfile) as webpage:
        html = webpage.read()
        soup = BSoup(html, 'html.parser')
    for styles in soup.select('style'):
        css = cssutils.parseString(styles.encode_contents())
        for rule in css:
            if rule.type == rule.STYLE_RULE:
                style = rule.selectorText
                selectors[style] = {}
                for item in rule.style:
                    propertyname = item.name
                    value = item.value
                    selectors[style][propertyname] = value

BeautifulSoup parses all "style" tags in the html (head & body), .encode_contents() converts the BeautifulSoup objects into a byte format that cssutils can read, and then cssutils parses the individual CSS styles all the way down to the property/value level via rule.selectorText & rule.style.

Note: The "rule.STYLE_RULE" filters out only styles. The cssutils documentation details options for filtering media rules, comments and imports.

It'd be cleaner if you broke this down into functions, but you get the gist...