Cleaning is not configurable enough

[mjl]

The set of allowed tags and merge tags is hardcoded and cannot be changed.

Ideally, one would like to be able to set those once as default or per cleanse call.

[sbaechler]

The class based approach is a good start. It still needs a hook for custom cleaner functions.

The # remove elements containing only whitespace or linebreaks function must be configurable as well. Because now it also cleans empty <td></td> tags which invalidates tables. And sometimes <p></p> is ok.

[l-monnier]

You can actually edit the list of allowed tags by changing the value of the class attribute.
For example, if you wish to add some extra tags for a particular instance::

additional_allowed_tags = {
    'h1':(),
    'h4':(),
    'h5':(),
    'h6':(),
    'i':(),
    'div':(),
}
cleanse = Cleanse()
cleanse.allowed_tags.update(additional_allowed_tags)

Note that since it is a class attribute, you can also edit the value of the class (so all future instances will have the new value)::

additional_allowed_tags = {
    'h1':(),
    'h4':(),
    'h5':(),
    'h6':(),
    'i':(),
    'div':(),
}
Cleanse.allowed_tags.update(additional_allowed_tags)