1 import sys

     2 

     3 from xml.sax import make_parser, handler

     4 

     5 class FancyCounter(handler.ContentHandler):

     6 

     7     def __init__(self):

     8         self._elems = 0

     9         self._attrs = 0

    10         self._elem_types = {}

    11         self._attr_types = {}

    12 

    13     def startElement(self, name, attrs):

    14         self._elems = self._elems + 1

    15         self._attrs = self._attrs + len(attrs)

    16         self._elem_types[name] = self._elem_types.get(name, 0) + 1

    17 

    18         for name in attrs.keys():

    19             self._attr_types[name] = self._attr_types.get(name, 0) + 1

    20 

    21     def endDocument(self):

    22         print "There were", self._elems, "elements."

    23         print "There were", self._attrs, "attributes."

    24 

    25         print "---ELEMENT TYPES"

    26         for pair in  self._elem_types.items():

    27             print "%20s %d" % pair

    28 

    29         print "---ATTRIBUTE TYPES"

    30         for pair in  self._attr_types.items():

    31             print "%20s %d" % pair

    32 

    33 

    34 parser = make_parser()

    35 parser.setContentHandler(FancyCounter())

    36 parser.parse(sys.argv[1])