Quite often there is a need to clean up HTML from some source, be it user input or data gathered by scraping, which needs to be cleaned up. With the SoupStrainer class in collective.soupstrainer this is made easy. It uses beautifulsoup4 to parse and clean up HTML. The constructor of the class takes four arguments.

exclusions

This is a list of tuples with two items each. The first item is a list of tag names, the second item is a list of attributes. If the list of attributes is empty, then each tag in the first list is completely removed from the passed in HTML. If the list of tags is empty, then each attribute listed is completely removed. If there are both tags and attributes listed, then the attributes are only removed from matching tags.

style_whitelist

This is a white list of CSS styles allowed in 'style' attributes. All other styles are removed.

class_blacklist

This is a black list for CSS classes. Each matching class is removed from 'class' attributes.

parser

This is the parser used by beautifulsoup4, when the strainer is called with a string. It must be an installed parser for beautifulsoup4, defaults to html.parser

An instance of the SoupStrainer class can be called directly with one argument. The argument can either be a string, in which case it will internally be parsed by beautifulsoup4 and the result will be unicode (or string in python 3), or it can be a parsed HTML tree created by beautifulsoup4, in which case it will be modified in place and be returned again.