simple conversion from etree parsed xml to objects

xmlhelp.py

''' xml utilities '''
__author__ = '[email protected]'

from itertools import groupby
from operator import attrgetter

def split_ns(txt):
   ''' returns [ns, tag] for '{ns}tag'

   >>> split_ns('{http://www.w3.org/2001/XMLSchema-instance}nil')
   ['http://www.w3.org/2001/XMLSchema-instance', 'nil']
   >>> split_ns('foobar')
   ['', 'foobar']
   '''
   s = txt[txt.startswith('{'):].split('}', 1)
   return s if len(s) > 1 else [''] + s

class protected(dict):
   ''' raises exception if you try to update an existing key

   >>> x = protected(zip('abcd', '1234'))
   >>> x['a'] = 3
   Traceback (most recent call last):
   KeyError: ('already exists', 'a')
   >>> x['e'] = 5
   ''' 
   def __setitem__(self, k, v):
      if k in self: raise KeyError('already exists', k)
      dict.__setitem__(self, k, v)

def objectify(elm, prefix='', base=None):
   ''' converts a parsed etree element into an object.
   
   if prefix is specified (str), it will prefix all the tags it inserts into
   the base.  usefuly for preventing namespace collisions.

   when base is specified, it will create a heirarchy of bases to represent the
   parse.  base needs to implement MutableMapping (e.g. dict, OrderedDict,
   etc.).  if base is unspecified, it **WILL MODIFY THE ETREE ELEMENT OBJECTS**
   it traverses by adding attributes to them.  WARNING: if it wasn't clear,
   this is potentially quite destructive on the etree elements!
   see dictify to for an example.
   '''
   #print '--', elm.tag, elm
   rval = elm.__dict__ if base is None else base()
  
   # text
   text = elm.text.strip() if elm.text else elm.text
   tail = elm.tail.strip() if elm.tail else elm.tail
   if not (len(elm) or elm.attrib or tail): return text
   if text: rval[prefix + 'text'] = text
   if tail: rval[prefix + 'tail'] = tail

   # attributes
   rval.update((prefix + split_ns(k)[1], v) for k,v in elm.attrib.items())
   # test for nil
   if rval.get(prefix + 'nil') == 'true': return None

   # children
   c_groups = [ (split_ns(k)[1], [objectify(i, prefix, base) for i in g]) for k,g in groupby(list(elm), attrgetter('tag')) ]
   for t, c in c_groups:
      if len(c) > 1 and len(c_groups) == 1 and not (elm.attrib or elm.text):
         # many children of one type and nothing else
         return c
      rval[prefix + t] = c if len(c) > 1 else c[0]

   return elm if base is None else rval

def dictify(elm, prefix='', base=protected):
   ''' returns a dict representation of the xml '''
   return objectify(elm, prefix, base)