class Parser(object):
'''
Parser class.
>>> p = Parser()
>>> print type(p)
<class 'parser.Parser'>
# test exceptions
>>> p.tokenize()
Traceback (most recent call last):
raise ParserError, 'cannot tokenize empty sentence'
ParserError: cannot tokenize empty sentence
>>>
>>> p.getParseList()
Traceback (most recent call last):
raise ParserError, 'cannot parse empty sentence'
ParserError: cannot parse empty sentence
>>>
>>> p.setSentence('I move quickly East')
'''
def __init__(self, sentence=''):
grammar = Grammar()
self.grammar = grammar.grammar
self.parser = None
if sentence:
self.setSentence(sentence)
else:
self.sentence = sentence
def tokenize(self):
from copy import copy
if not self.sentence:
raise ParserError, 'cannot tokenize empty sentence'
_sentence = tokenizer.Token(TEXT=self.sentence)
tokenizer.WhitespaceTokenizer().tokenize(_sentence)
self._sentence = _sentence
def setParser(self):
self.parser = Parser(self.grammar)
def getParser(self):
if not self.parser:
self.setParser()
return self.parser
# XXX getParseList is broken right now...
def getParseList(self):
if not self.sentence:
raise ParserError, 'cannot parse empty sentence'
if not self.parser:
self.setParser()
return self.parser.get_parser_list(self._sentence)
def setSentence(self, sentence=''):
self.sentence = sentence
self.tokenize()
def setParser(self):
self.parser = Parser(self.grammar)
CFGProduction(V, ['saw']), CFGProduction(P, ['in']),
CFGProduction(P, ['with']), CFGProduction(N, ['park']),
CFGProduction(N, ['dog']), CFGProduction(N, ['telescope'])
)
grammar = CFG(S, productions)
# Tokenize a simple sentence
sentence = 'I saw a man in the park'
sent = Token(TEXT=sentence)
print sent
print type(sent)
WhitespaceTokenizer().tokenize(sent)
print sent
# Build a parser
parser = RecursiveDescentParser(grammar)
print dir(parser)
print parser.__dict__
for p in parser.get_parse_list(sent):
print p
print '\n\n\n'
from parser import Parser
p = Parser()
p.setSentence('I move slowly south')
print p
print p.sentence
print p._sentence
