I recently wrote some code to translate XML into a python data structure, although I did have to handle attributes. I used xml.dom.minidom rather than ElementTree, for a similar reason. I haven't actually tested this on Python 2.4.4, but I think it will work. I didn't write a reverse XML generator, though you can probably use the 'lispy_string' function I included to do this.
I also included some shortcuts specific to the application I was writing (explained in the docstring), but you might find those shortcuts useful too, from the sounds of it. Essentially, an xml tree technically translates into a dictionary of lists of dictionaries of lists of dictionaries of lists, etc. I omit creating the intermediary lists unless they are necessary, so you can reference elements by dictname[element1][element2] rather than dictname[element1][0][element2][0] and so on.
Attribute handling is a little kludgy, I strongly recommend reading the code before doing anything with attributes.
import sys
from xml.dom import minidom
def dappend(dictionary, key, item):
"""Append item to dictionary at key. Only create a list if there is more than one item for the given key.
dictionary[key]=item if key doesn't exist.
dictionary[key].append(item) if key exists."""
if key in dictionary.keys():
if not isinstance(dictionary[key], list):
lst=[]
lst.append(dictionary[key])
lst.append(item)
dictionary[key]=lst
else:
dictionary[key].append(item)
else:
dictionary.setdefault(key, item)
def node_attributes(node):
"""Return an attribute dictionary """
if node.hasAttributes():
return dict([(str(attr), str(node.attributes[attr].value)) for attr in node.attributes.keys()])
else:
return None
def attr_str(node):
return "%s-attrs" % str(node.nodeName)
def hasAttributes(node):
if node.nodeType == node.ELEMENT_NODE:
if node.hasAttributes():
return True
return False
def with_attributes(node, values):
if hasAttributes(node):
if isinstance(values, dict):
dappend(values, '#attributes', node_attributes(node))
return { str(node.nodeName): values }
elif isinstance(values, str):
return { str(node.nodeName): values,
attr_str(node): node_attributes(node)}
else:
return { str(node.nodeName): values }
def xmldom2dict(node):
"""Given an xml dom node tree,
return a python dictionary corresponding to the tree structure of the XML.
This parser does not make lists unless they are needed. For example:
'<list><item>1</item><item>2</item></list>' becomes:
{ 'list' : { 'item' : ['1', '2'] } }
BUT
'<list><item>1</item></list>' would be:
{ 'list' : { 'item' : '1' } }
This is a shortcut for a particular problem and probably not a good long-term design.
"""
if not node.hasChildNodes():
if node.nodeType == node.TEXT_NODE:
if node.data.strip() != '':
return str(node.data.strip())
else:
return None
else:
return with_attributes(node, None)
else:
#recursively create the list of child nodes
childlist=[xmldom2dict(child) for child in node.childNodes if (xmldom2dict(child) != None and child.nodeType != child.COMMENT_NODE)]
if len(childlist)==1:
return with_attributes(node, childlist[0])
else:
#if False not in [isinstance(child, dict) for child in childlist]:
new_dict={}
for child in childlist:
if isinstance(child, dict):
for k in child:
dappend(new_dict, k, child[k])
elif isinstance(child, str):
dappend(new_dict, '#text', child)
else:
print "ERROR"
return with_attributes(node, new_dict)
def load(fname):
return xmldom2dict(minidom.parse(fname))
def lispy_string(node, lst=None, level=0):
if lst==None:
lst=[]
if not isinstance(node, dict) and not isinstance(node, list):
lst.append(' "%s"' % node)
elif isinstance(node, dict):
for key in node.keys():
lst.append("\n%s(%s" % (spaces(level), key))
lispy_print(node[key], lst, level+2)
lst.append(")")
elif isinstance(node, list):
lst.append(" [")
for item in node:
lispy_print(item, lst, level)
lst.append("]")
return lst
if __name__=='__main__':
data = minidom.parse(sys.argv[1])
d=xmldom2dict(data)
print d
