HTML cleansing (feincms3.cleanse)

HTML cleansing is by no means only useful for user generated content. Managers also copy-paste content from word processing programs, the rich text editor’s output isn’t always (almost never) in the shape we want it to be, and a strict allowlist based HTML sanitizer is the best answer I have.

class feincms3.cleanse.CleansedRichTextField(*args, **kwargs)[source]

This is a subclass of django-ckeditor’s RichTextField. The recommended configuration is as follows:

    "default": {
        "toolbar": "Custom",
        "format_tags": "h1;h2;h3;p;pre",
        "toolbar_Custom": [[
            "Format", "RemoveFormat", "-",
            "Bold", "Italic", "Subscript", "Superscript", "-",
            "NumberedList", "BulletedList", "-",
            "Anchor", "Link", "Unlink", "-",
            "HorizontalRule", "SpecialChar", "-",

# Settings for feincms3.plugins.richtext.RichText
CKEDITOR_CONFIGS["richtext-plugin"] = CKEDITOR_CONFIGS["default"]

The corresponding HTML_SANITIZERS configuration for html-sanitizer would look as follows:

    "default": {
        "tags": {
            "a", "h1", "h2", "h3", "strong", "em", "p",
            "ul", "ol", "li", "br", "sub", "sup", "hr",
        "attributes": {
            "a": ("href", "name", "target", "title", "id", "rel"),
        "empty": {"hr", "a", "br"},
        "separate": {"a", "p", "li"},

        # Additional default settings not listed here.

At the time of writing those are the defaults of html-sanitizer, so you don’t have to do anything.

If you want or require a different cleansing function, simply override the default with CleansedRichTextField(cleanse=your_function). The cleansing function receives the HTML as its first and only argument and returns the cleansed HTML.

clean(value, instance)[source]

Convert the value’s type and run validation. Validation errors from to_python() and validate() are propagated. Return the correct value if no error is raised.


Pass ugly HTML, get nice HTML back.