NLTK语句的基因组表示

Question

给定NLTK语法，我如何使用整数数组来表示句子？

我正在使用NLTK从特定的语法生成一些句子。我想生成一个整数数组来表示生成的句子的基因组(表型)。

有了整数的表示，我将在遗传算法中进化基因组，执行一些变异以获得更好的句子。

例如,

from nltk import CFG
from nltk.parse.generate import generate, demo_grammar

g = CFG.fromstring(demo_grammar)
sentence = next(generate(g, n=1))

print(sentence) # ex: ['the', 'man', 'saw', 'the', 'park']

convert_to_genotype(sentence) # returns [253, 69, 221, 97, 190, 254, 67, 137, 95, 72, 54, 232, 11, 136] for example.

如何创建convert_to_genotype函数？

谢谢

Answer 1

经过一些研究，我创建了一个为给定基因组创建表型的实现。这就是我正在寻找的，用来自语法的规则创建的进化我的个体。

import nltk
from nltk import CFG

GRAMMAR = CFG.fromstring("""
string -> letter | letter string
letter -> vowel | consonant | char
char   -> ' '|'!'|'?'|','|'.'

vowel       -> lower_vowel | upper_vowel
lower_vowel -> 'a'|'e'|'o'|'i'|'u'
upper_vowel -> 'A'|'E'|'I'|'O'|'U'

consonant       -> lower_consonant | upper_consonant
lower_consonant -> 'b'|'c'|'d'|'f'|'g'|'h'|'j'|'k'|'l'|'m'|'n'|'p'|'q'|'r'|'s'|'t'|'v'|'w'|'x'|'y'|'z'
upper_consonant -> 'B'|'C'|'D'|'F'|'G'|'H'|'J'|'K'|'L'|'M'|'N'|'P'|'Q'|'R'|'S'|'T'|'V'|'W'|'X'|'Y'|'Z'
""")

def genome_to_grammar(array):
  sb = []
  stack = [GRAMMAR.start()]
  index = 0
  wraps = 0

  while stack:
    symbol = stack.pop()
    if isinstance(symbol, str):
      sb.append(symbol)
    else:
      rules = [i for i in GRAMMAR.productions() if i.lhs().symbol() == symbol.symbol()]
      rule_index = 0
      if len(rules) > 1:
        rule_index = array[index] % len(rules)
        index += 1
        if index >= len(array):
          index = 0
          wraps += 1
          if wraps > 10:
            return None
      rule = rules[rule_index]
      for production in reversed(rule.rhs()):
        stack.append(production)

  return ''.join(sb)

genome = [253, 69, 221, 97, 190, 254, 67, 137, 95, 72, 54, 232, 11, 136]
print(genome_to_grammar(genome))