Source Code for Module nltk.olac

 1  # Natural Language Toolkit: Support for OLAC Metadata 
 2  # 
 3  # Copyright (C) 2001-2008 NLTK Project 
 4  # Author: Steven Bird <[email protected]> 
 5  # URL: <http://nltk.org> 
 6  # For license information, see LICENSE.TXT 
 7   
 8   
 9  from StringIO import StringIO 
10   
11 -def read_olac(xml): 
12      """ 
13      Read an OLAC XML record and return a list of attributes. 
14   
15      @param xml: XML string for conversion     
16      @type xml: C{string} 
17      @rtype: C{list} of C{tuple} 
18      """ 
19      from lxml import etree 
20   
21      root = etree.parse(StringIO(xml)).getroot() 
22      return [(element.tag, element.attrib, element.text) for element in root.getchildren()] 
23   
24 -def pprint_olac(xml): 
25      for tag, attrib, text in read_olac(xml): 
26          print "%-12s" % tag + ':', 
27          if text: 
28              print text, 
29          if attrib: 
30              print "(%s=%s)" % (attrib['type'], attrib['code']), 
31          print 
32   
33 -def demo(): 
34      from lxml import etree 
35      import nltk.data 
36   
37      file = nltk.data.find('corpora/treebank/olac.xml') 
38      xml = open(file).read() 
39      pprint_olac(xml) 
40   
41  if __name__ == '__main__': 
42      demo() 
43   
44  __all__ = ['read_olac', 'pprint_olac'] 
45