Reber grammar generation and validation

grammar.py

import nltk
from nltk.parse.generate import generate


# Define Reber grammar
# http://christianherta.de/lehre/dataScience/machineLearning/neuralNetworks/pics/embeddedReberGrammar.png
grammarStr = '''
            START -> 'B' 'T' REBER 'T' 'E' | 'B' 'P' REBER 'P' 'E'
            REBER -> 'B' E1
            E1 -> 'T' E2 | 'P' E3
            E2 -> 'S' E2 | 'X' E4
            E3 -> 'T' E3 | 'V' E5
            E4 -> 'X' E3 | 'S' E6
            E5 -> 'P' E4 | 'V' E6
            E6 -> 'E'
            '''

grammar = nltk.CFG.fromstring(grammarStr)
sentences = set([''.join(sentence) for sentence in generate(grammar, depth=30)])
print(len(sentences))

def valid(string):
  r = nltk.ChartParser(grammar).parse(string)
  return len(list(r)) > 0

print("'BTBPVVETE' is valid?", valid('BTBPVVETE'))
print("'BTBPVVETE' is valid?", 'BTBPVVETE' in sentences)
print("'BTBPVVETEE' is valid?", valid('BTBPVVETEE'))
print("'BTBPVVETEE' is valid?", 'BTBPVVETEE' in sentences)