def parsing_examples():
rules = list()
rules.append("GAMMA -> R [1.0]")
rules.append("R -> R '+' R [0.4]")
rules.append("R -> '0' [0.3]")
rules.append("R -> '1' [0.3]")
grammar_rules = grammarutils.get_pcfg(rules)
grammar = nltk.PCFG.fromstring(grammar_rules)
sentence = '0 + 1'
tokens = sentence.split(' ')
# earley_parser = nltk.EarleyChartParser(grammar, trace=1)
# e_chart = earley_parser.chart_parse(tokens)
symbols = ['0', '1', '+']
symbol_index = dict()
for s in symbols:
symbol_index[s] = symbols.index(s)
grammar_rules = grammarutils.get_pcfg(rules,
index=True,
mapping=symbol_index)
grammar = nltk.PCFG.fromstring(grammar_rules)
classifier_output = [[0.8, 0.8, 0.1, 0.1, 0.1], [0.1, 0.1, 0.1, 0.8, 0.8],
[0.1, 0.1, 0.8, 0.1, 0.1]]
classifier_output = np.array(classifier_output).T
print classifier_output.shape
gen_earley_parser = generalizedearley.GeneralizedEarley(grammar)
best_string, prob = gen_earley_parser.parse(classifier_output)
print best_string, prob
def __init__(self, rule_path, symbols):
self._grammar = read_infuced_grammar(rule_path, symbols)
self._symbols = symbols
self._bottom_up_prob_matrix = []
self._generalized_earley_parser = generalizedearley.GeneralizedEarley(
self._grammar)
self._current_action_seq = []
self._exec_status = False
def test_generalized_earley(grammar, classifier_output):
gen_earley_parser = generalizedearley.GeneralizedEarley(grammar)
best_string, prob = gen_earley_parser.parse(classifier_output)
print 'best_string with prob {:.3f}:'.format(prob), best_string
print gen_earley_parser.compute_labels()
print np.argmax(classifier_output, axis=1)