Pretty printing XML in Python

pretty_print_xml.py

from StringIO import StringIO
import xml.etree.cElementTree as ET

def pretty_xml(xml_str, indent="  "):
    """
    A very simple, hopefully not simplistic, XML pretty printer.
    Concept courtesy Mark Williams.
    """
    if not hasattr(xml_str, "read"): # ElementTree uses file-like objects
        fn = StringIO(xml_str)  # cStringIO doesn't support UTF-8
    else:
        fn = xml_str
    cursor = 0
    out_list = []
    for event, elem in ET.iterparse(fn, events=('start', 'end')):
        if event == 'start':
            attrs = ' '.join([k+'="'+v+'"' for k, v in elem.items()])
            cur_tag = ('<{tag} {attrs}>'.format(tag=elem.tag, attrs=attrs)
                       if attrs else '<{tag}>'.format(tag=elem.tag))
            if elem.text is None:
                had_txt = False
                txt = '\n'
            else:
                had_txt = True
                txt = elem.text
            out_list.extend([indent*cursor, cur_tag, txt])
            cursor += 1
        else:
            cursor -= 1
            cur_ind = cursor*indent if not had_txt else ''
            out_list.extend([cur_ind, '</{0}>'.format(elem.tag), '\n'])
            had_txt = False
    return ''.join(out_list)

"""
The function above is about twice as fast as the following, with comparable output
formats. The one below might be slightly better, as far as looks and accuracy are concerned.
"""
import re
from xml.dom.minidom import parseString

_xml_re = re.compile('>\n\s+([^<>\s].*?)\n\s+</', re.DOTALL)

def pretty_xml_old(xml_str, indent="  "):
    xml_re = _xml_re
    # avoid re-prettifying large amounts of xml that is fine
    if xml_str.count("\n") < 20:
        pxml = parseString(xml_str).toprettyxml(indent)
        return xml_re.sub('>\g<1></', pxml)
    else:
        return xml_str