A data cleaning cerberus validator (remove unknown fields)

gistfile1.py

import copy
import collections
from cerberus import Validator

class CleaningValidator(Validator):
    def validate_clean(self, document, *args, **kwargs):
        self.validate(document, *args, **kwargs)
        if self.errors and self.allow_unknown == False:
            self._clean()
        return len(self.errors) == 0

    def _clean(self, document=None, errors=None, context=(None, {})):
        if not document:
            document = self.document
        if not errors:
            errors = self._errors

        key, parent  = context
        clean_errors = copy.copy(errors)
        
        for k, v in clean_errors.items():
            if v == 'unknown field':
                del(document[k])
                del(errors[k])

            if isinstance(v, collections.Mapping):
                self._clean(document[k], errors[k], (k, errors))

        if key and parent:
            if not parent.get(key):
                del(parent[key])