Simple XML file scrubber in Python

scrub_it.py

from xml.sax.saxutils import XMLGenerator
import xml

def mask_field(field):
    return 'X'*len(field) 

class ScrubHandler(XMLGenerator, object):
    
    def __init__(self, scrub_map={}, out=None):
        super(ScrubHandler, self).__init__(out=out)
        self.scrub_map = scrub_map
        self.buffer = ""

    def startElement(self, name, attr):
        super(ScrubHandler, self).startElement(name, attr)

    def characters(self, data):
        self.buffer += data
 
    def endElement(self, name):
        if name in self.scrub_map:
            self.buffer = self.scrub_map[name](self.buffer)

        super(ScrubHandler, self).characters(self.buffer)
        self.buffer = ""
        super(ScrubHandler, self).endElement(name)

def scrub_xml_file(input_filename, output_filename, scrub_map={}):
    
    output_file = open(output_filename, 'w')
    handler = ScrubHandler(out=output_file, scrub_map=scrub_map)
    parser = xml.sax.make_parser()
    parser.setContentHandler(handler)
    parser.parse(input_filename)
    output_file.close()

if __name__ == "__main__":
    scrub_elements = {'SSN': mask_field, 'LastName': mask_field }
    sample_xml = "<Doc><SSN>123456789</SSN><LastName>CORBIN</LastName><FirstName>Brian</FirstName></Doc>"
    handler = ScrubHandler(scrub_map=scrub_elements)
    xml.sax.parseString(sample_xml, handler)