tools/xincludator.py

   1 #!/usr/bin/python
   2
   3 from sys import argv, stdout, stderr
   4 import codecs, locale
   5 import os
   6 import xml.dom.minidom
   7
   8 stdout = codecs.getwriter('utf-8')(stdout)
   9
  10 NS_XI = 'http://www.w3.org/2001/XInclude'
  11
  12 def xincludate(dom, base, dropns = []):
  13     remove_attrs = []
  14     for i in xrange(dom.documentElement.attributes.length):
  15         attr = dom.documentElement.attributes.item(i)
  16         if attr.prefix == 'xmlns':
  17             if attr.localName in dropns:
  18                 remove_attrs.append(attr)
  19             else:
  20                 dropns.append(attr.localName)
  21     for attr in remove_attrs:
  22         dom.documentElement.removeAttributeNode(attr)
  23     for include in dom.getElementsByTagNameNS(NS_XI, 'include'):
  24         href = include.getAttribute('href')
  25         # FIXME: assumes Unixy paths
  26         filename = os.path.join(os.path.dirname(base), href)
  27         subdom = xml.dom.minidom.parse(filename)
  28         xincludate(subdom, filename, dropns)
  29         if './' in href:
  30             subdom.documentElement.setAttribute('xml:base', href)
  31         include.parentNode.replaceChild(subdom.documentElement, include)
  32
  33 if __name__ == '__main__':
  34     argv = argv[1:]
  35     dom = xml.dom.minidom.parse(argv[0])
  36     xincludate(dom, argv[0])
  37     xml = dom.toxml()
  38     stdout.write(xml)
  39     stdout.write('\n')