#!/usr/bin/env python
"""
Copyright 2007, Sean B. Palmer, inamidst.com
Licensed under the Eiffel Forum License 2.
"""

import re

# @@ duck-typing for the win!

# Kinds of serialisation:
# GRDDL with various profiles
# N-Triples
# RDF/XML (done)
# RDF/XML Pretty
# Turtle Pretty
# N3 Pretty
# 
# +bz2

r_hex4_32 = re.compile(ur'([\x00-\x08\x0b\x0C\x0E-\x1F\x7F-\uFFFF]+)')
r_hex4_33 = re.compile(ur'([\x00-\x08\x0b\x0C\x0E-\x1F\x3E\x7F-\uFFFF]+)')

# Cf. http://bugs.python.org/issue1477
try: r_hex6 = re.compile(u'([\U00010000-\U0010FFFF]+)')
except: r_hex6 = None

def hex4(m): 
   return u''.join('\\u%04X' % ord(c) for c in m.group(1))

def hex6(m): 
   return u''.join('\\U%08X' % ord(c) for c in m.group(1))

def escaped32(s): 
   # http://www.w3.org/TR/rdf-testcases/#ntrip_strings
   s = s.replace('\\', '\\\\')
   s = s.replace('\t', '\\t')
   s = s.replace('\n', '\\n')
   s = s.replace('\r', '\\r')
   s = s.replace('"', '\\"')
   s = r_hex4_32.sub(hex4, s)
   if r_hex6 is not None: 
      return r_hex6.sub(hex6, s)
   
   if __import__('sys').maxunicode <= 0xffff: 
      warning = ("Warning: You're using a Narrow Python build", 
         "This means that N-Triples output might not be fully compliant", 
         "Use a python compiled with --enable-unicode=ucs4 to fix this")
      for line in warning: 
         print >> __import__('sys').stderr, line
      return s
   raise ValueError('r_hex6 did not compile properly')

def escaped33(s): 
   # http://lists.w3.org/Archives/Public/www-rdf-comments/2007OctDec/0008
   s = s.replace('\\', '\\\\')
   s = s.replace('\t', '\\t')
   s = s.replace('\n', '\\n')
   s = s.replace('\r', '\\r')
   s = s.replace('"', '\\"')
   s = r_hex4_33.sub(hex4, s)
   if r_hex6 is not None: 
      return r_hex6.sub(hex6, s)
   
   if __import__('sys').maxunicode <= 0xffff: 
      warning = ("Warning: You're using a Narrow Python build", 
         "This means that N-Triples output might not be fully compliant", 
         "Use a python compiled with --enable-unicode=ucs4 to fix this")
      for line in warning: 
         print >> __import__('sys').stderr, line
      return s
   raise ValueError('r_hex6 did not compile properly')

def ntriples(graph): 
   for triple in graph: 
      if hasattr(triple.subject, 'value'): 
         uri = triple.subject.value
         subject = '<' + escaped33(uri.encode('utf-8')) + '>'
      elif hasattr(triple.subject, 'label'): 
         label = triple.subject.label
         subject = '_:' + label.encode('utf-8')

      if hasattr(triple.predicate, 'value'): 
         uri = triple.predicate.value
         predicate = '<' + escaped33(uri.encode('utf-8')) + '>'
      else: raise ValueError('P: %s' % triple.predicate)

      if hasattr(triple.object, 'value'): 
         uri = triple.object.value
         object = '<' + escaped33(uri.encode('utf-8')) + '>'
      elif hasattr(triple.object, 'label'): 
         label = triple.object.label
         object = '_:' + label.encode('utf-8')
      elif hasattr(triple.object, 'lexical'): 
         lexical = triple.object.lexical
         object = '"' + escaped32(lexical) + '"'

      yield subject + ' ' + predicate + ' ' + object + ' .'

def rdfxml(graph): 
   yield '<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">'
   def split(uri): 
      if uri.startswith(u'http://www.w3.org/1999/02/22-rdf-syntax-ns#'): 
         return uri[:43].encode('utf-8'), uri[43:].encode('utf-8')
      return uri[:-1].encode('utf-8'), uri[-1].encode('utf-8')
   for triple in graph: 
      if hasattr(triple.subject, 'value'): 
         uri = triple.subject.value
         yield '<rdf:Description rdf:about="%s">' % uri.encode('utf-8')
      elif hasattr(triple.subject, 'label'): 
         label = triple.subject.label
         yield '<rdf:Description rdf:nodeID="%s">' % label
      else: raise ValueError('Unable to serialise %s' % triple.subject)

      if hasattr(triple.predicate, 'value'): 
         nsname, tagname = split(triple.predicate.value)
         yield '<%s xmlns="%s"' % (tagname, nsname)
      else: raise ValueError('Unable to serialise %s' % triple.predicate)

      if hasattr(triple.object, 'language'): 
         if triple.object.language: # might be None
            yield 'xml:lang="%s"' % triple.object.language
      elif hasattr(triple.object, 'datatype'): 
         if triple.object != XMLLiteral: 
            yield 'rdf:datatype="%s"' % triple.object.datatype.encode('utf-8')
         else: yield 'rdf:parseType="Literal"'

      if hasattr(triple.object, 'value'): 
         uri = triple.object.value
         yield 'rdf:resource="%s"/>' % uri.encode('utf-8')
      elif hasattr(triple.object, 'label'): 
         label = triple.object.label
         yield 'rdf:nodeID="%s"/>' % label
      elif hasattr(triple.object, 'lexical'): 
         lexical = triple.object.lexical
         end = '</%s>' % tagname
         for line in ('>' + lexical.encode('utf-8') + end).split('\n'): 
            yield line
      else: raise ValueError('Unable to serialise %s' % triple.object)

      yield '</rdf:Description>'
   yield '</rdf:RDF>'

# EOF