#!/usr/bin/python
"""<http://infomesh.net/2002/eep/>"""

import re, string, random

S, P, O = 0, 1, 2

class Article: 
   def __init__(self, s): 
      s, self.r = str(s), str(s)
      if s[0] == '"': self.v, self.t = s[1:-1], 'Literal'
      elif s[0] == '?': self.v, self.t = s[1:], 'Univar'
      elif s[0] == '_': self.v, self.t = s[2:], 'Exivar'
      elif s[0] == '<': self.v, self.t = s[1:-1], 'URI'
      else: self.v, self.t, self.r = s, 'URI', '<%s>' % s
   def __repr__(self): return self.r
   def set(self, s): self.__init__(s)

def parse(s, t=[]): 
   if len(s) == 0: raise 'Document has no content'
   rc, a = re.compile(r'(\#[^\n]*)'), r'(<.*?>|_:\S+|\?\S+|"(?:\\"|[^"])*")[ \t]'
   rw, rt, y = re.compile(r'[ \t]+'), r'[ \t]*%s%s%s*.[ \t]*' % (a, a, a), t[:]
   for line in s.replace('\r\n', '\n').replace('\r', '\n').split('\n'): 
      if re.compile(rt).match(line):
          x = re.compile(rt).findall(line)[0]
          y.append([Article(x[0]), Article(x[1]), Article(x[2])])
      elif rc.match(line) or rw.match(line) or (len(line) == 0): continue
      else: raise 'Line is invalid', line
   return y

def query(q, triples, r=[]): 
   result = r[:]
   for t in triples: 
      if (((q[0].t == 'Univar') or (t[0].r == q[0].r)) 
          and ((q[1].t == 'Univar') or (t[1].r == q[1].r)) 
          and ((q[2].t == 'Univar') or (t[2].r == q[2].r))): result.append(t)
   return result

def nquery(s, x): 
   return query(parse(s)[0], x)

def u(uri): return '<'+uri+'>'

def b(label=''): 
   if label == '': # 208,827,064,576 combinations
      for x in range(8): label += string.lowercase[random.randrange(0, 25)]
      return '_:'+label
   else: return '_:'+label

def l(s): return '"'+s+'"'

def serialize(triples): 
   result = ''
   for triple in triples: 
      result += '%s %s %s .\n' % (triple[S], triple[P], triple[O])
   return result.rstrip()

def nt2xmlrdf(ntriples): 
   return serializeXML(parse(ntriples))

def serializeXML(triples): 
   xrdf = '<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">\n'
   for t in triples:
      for x in (S, P, O):
         if '#' in t[x].v: 
            term, txmlns = t[x].v, t[x].v.split('#')[0]+'#'
            tname = t[x].v.split('#')[1]
         else: term, txmlns, tname = t[x].v, t[x].v[:-1], t[x].v[-1]
         if x == S and t[x].t == 'URI': 
            xrdf += '<rdf:Description rdf:about="'+term+'"/>\n'
         elif x == S and t[x].t == 'Exivar': xrdf += '<rdf:Description>\n'
         elif x == S and t[x].t == 'Literal': # @@ encoding
            xrdf += '<rdf:Description rdf:about="data:,'+term+'">\n'
         elif x == P and t[x].t == 'URI':
            xrdf += '   <'+tname+' xmlns="'+txmlns+'"'
            blargh = tname[:]
         elif x == P and t[x].t == 'Exivar':
            sys.stderr.write('Kaboom!') # i.e. not implementable
            xrdf += '   <'+term+' xmlns="anon:_" \n' # A new URI Scheme...
         elif x == P and t[x].t == 'Literal': raise 'Literal as predicate'
         elif x == O and t[x].t == 'URI': 
            xrdf += '\n     rdf:resource="'+term+'"/>\n'
         elif x == O and t[x].t == 'Exivar':
            xrdf += '\n     rdf:resource="anon:_'+term+'"/>\n'
         elif x == O and t[x].t == 'Literal': # This is a hack
            if blargh: xrdf += '>'+term+'</'+blargh+'>\n'
            else: xrdf += '\n  rdf:resource="data:,'+term+'"/>\n'
      xrdf += '</rdf:Description>\n'
   return xrdf + '</rdf:RDF>'

class Namespace: 
    def __init__(self, ns='', sep=''): 
       self.ns, self.sep = ns, sep
    def __getattr__(self, name): 
       return self.ns+self.sep+name

def test(): 
   print serialize(parse('<p> _:q ?r .\n "x"\t"y" <#z> . '))

if __name__=="__main__": 
   print __doc__